
From sh@defuze.org  Wed Apr  1 00:25:19 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BC9223A6B58 for <hybi@core3.amsl.com>; Wed,  1 Apr 2009 00:25:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.665
X-Spam-Level: 
X-Spam-Status: No, score=-1.665 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, J_CHICKENPOX_73=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5Gg9v7QGQVK7 for <hybi@core3.amsl.com>; Wed,  1 Apr 2009 00:25:13 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id F18213A6888 for <hybi@ietf.org>; Wed,  1 Apr 2009 00:25:12 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n317QCCY006045 for <hybi@ietf.org>; Wed, 1 Apr 2009 02:26:12 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n317QBgG006038; Wed, 1 Apr 2009 09:26:11 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Wed, 1 Apr 2009 09:26:11 +0200 (CEST)
Message-ID: <49364.193.253.216.132.1238570771.squirrel@mail1.webfaction.com>
In-Reply-To: <Pine.LNX.4.62.0904010536340.25058@hixie.dreamhostps.com>
References: <Pine.LNX.4.62.0903302124580.25058@hixie.dreamhostps.com> <BBB4A96C-6101-43C5-AF62-2E242EE64B31@gbiv.com> <Pine.LNX.4.62.0903310201130.25058@hixie.dreamhostps.com> <7B624611-7368-4B5F-B043-C0D96C94E359@gbiv.com> <Pine.LNX.4.62.0903310308470.25058@hixie.dreamhostps.com> <73E288BB-8ED4-4712-A289-6CBF20078274@gbiv.com> <Pine.LNX.4.62.0904010536340.25058@hixie.dreamhostps.com>
Date: Wed, 1 Apr 2009 09:26:11 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: hybi@ietf.org
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Subject: Re: [hybi] WebSockets
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2009 07:25:19 -0000

Hello,

New to this list but highly interested in what could come of it.

> Yes, the Web Socket protocol is its own protocol. Is that not clear in the
> Web Socket Protocol Internet Draft?

Honestly it was a bit confusing at first but your next paragraph cleared
things up in regards to how it relates to HTTP.

>
> The ability to bootstrap using an HTTP connection, Upgrade:ing to
> WebSockets, is optional, and is only intended for advanced users that need
> to reuse the same port for both purposes.
>

The HTTP-bootstrap sounds sensible if only because HTTP probably federates
well developers from all kind of background.

In any case I agree with Lisa [1] that there's a need for better
documentation of what long-poll refers to and how various techniques
differ. In other words, the most interesting section of the Web Sockets
draft to me was its introduction which, if I read the most recent version,
is empty, meaning I couldn't really infer what it was trying to solve.


- Sylvain

[1] http://www.ietf.org/mail-archive/web/hybi/current/msg00029.html
-- 
Sylvain Hellegouarch
http://www.defuze.org

From gregw@webtide.com  Thu Apr  2 23:18:18 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 550653A6881 for <hybi@core3.amsl.com>; Thu,  2 Apr 2009 23:18:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0CQjhB40NUFc for <hybi@core3.amsl.com>; Thu,  2 Apr 2009 23:18:17 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.185]) by core3.amsl.com (Postfix) with ESMTP id 03F0728C11A for <hybi@ietf.org>; Thu,  2 Apr 2009 23:18:16 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so832580tid.25 for <hybi@ietf.org>; Thu, 02 Apr 2009 23:19:18 -0700 (PDT)
Received: by 10.110.7.18 with SMTP id 18mr1265738tig.8.1238739558529; Thu, 02 Apr 2009 23:19:18 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 2sm343023tif.26.2009.04.02.23.19.15 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 02 Apr 2009 23:19:17 -0700 (PDT)
Message-ID: <49D5AA5F.8030503@webtide.com>
Date: Fri, 03 Apr 2009 17:19:11 +1100
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was:  Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 06:18:18 -0000

Ian Hickson wrote:
> On Tue, 31 Mar 2009, Greg Wilkins wrote:
>> The websocket proposal is both good and bad.
>> The Unicode datagram API is probably very good
>> for javascript programming.   But the zero terminated
>> wire protocol is horrible.
> 
> Could you elaborate on this? I'm curious to find out why you think it is 
> horrible.

Because it means that I need to scan data received just to know if I
have all the data I need to proceed.

If I knew a content-length or a chunk length, then I would know
immediately that a read return that I have enough data to process.

With null termination, I have to scan the data looking for nulls (because
I could receive multiple datagrams).   But if I'm scanning the data then
I'd like to be converting the UTF-8 at the same time, but as I might
not have all the data, I could end up with 4 bytes of a 5 byte utf-8
character and so my utf-conversion has to be interrupted and I have
to save the chars that I've already converted and wait for another
packet to come in.

The worst part is that most datagrams will arrive in 1 packet and
most UTF-8 chars are 1 byte, so most of the time this will not be a
problem!   But I still need to write code that will work with data
split between packets or packets with multiple datagrams etc.  So
I expect there will be lots of buggy implementations out there that
will look like they work, but will fail on a nasty network.


>> We see no reason why the wire protocol should deviate so far from normal 
>> HTTP style framing.
> 
> What would you consider normal HTTP style framing here? What would HTTP 
> framing give you that isn't available with the framing currently proposed 
> with Web Sockets?

HTTP framing is:
  + Some CRLF terminated headers containing meta-data
  + A CRLF
  + The content that is delimited either by:
     * a known content length
     * chunks, each with a known content length
     * EOF

We already have the code written and tested to handle this.  This code
is already robust in the face of nasty networks that send us 1 byte per
packet etc. etc.

It supports more or less arbitrary extensible meta data along with
the content, so all the normal mime type mechanisms can be brought
to bear.


> It seems that having a single byte at the start and a single byte at the 
> end is all the framing one needs. The byte at the start gives the kind of 
> frame, and the byte at the end is a byte that isn't valid in UTF-8 so it 
> gets around the need to have a length declaration (such declaration are 
> often the source of buffer overrun bugs).

It is certainly sufficient, but it is not efficient or convenient to
handle.     More over it is new and different without any reason to
be new and different.   If there is some deficiency in HTTP framing?
Is there some reason that you don't want to transport non-utf8 data?
Eg why can't I push an image or a sound bite?


>> More over, normal HTTP framing would allow binary data to be sent to 
>> non-javascript clients!
> 
> Binary data is supported by the WebSocket protocol, it's just not exposed 
> yet. (One of the design principles we have been following in the HTML5 
> space is "baby steps", starting with small APIs and only adding more 
> features once the first version is widely implemented.)
> 
> Specifically, to send binary data using the protocol you would send a 0x80 
> byte, the length (as a packed integer), and then the binary data.

Again, this feels like re-inventing the wheel.   HTTP already supports
binary data with known and unknown content lengths (chunking).

Things like packed integer lengths can come back to haunt you. With
Jetty I have had several users have problems because we initially
had length being an int and we had to convert to using a long.
OK, I can't think of the use-case for pushing such large content,
but then I don't see any reason to pick a transport that has that
limit built in.

HTTP is tried and tested and can transport anything.
I don't want something new, I just want it to go both ways!


cheers












From gregw@webtide.com  Thu Apr  2 23:25:35 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 355843A6881 for <hybi@core3.amsl.com>; Thu,  2 Apr 2009 23:25:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id joPcZLAizcWh for <hybi@core3.amsl.com>; Thu,  2 Apr 2009 23:25:33 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.190]) by core3.amsl.com (Postfix) with ESMTP id 8A1C43A68C7 for <hybi@ietf.org>; Thu,  2 Apr 2009 23:25:33 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so835695tid.25 for <hybi@ietf.org>; Thu, 02 Apr 2009 23:26:34 -0700 (PDT)
Received: by 10.110.57.5 with SMTP id f5mr1264110tia.19.1238739994319; Thu, 02 Apr 2009 23:26:34 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm4502tib.28.2009.04.02.23.26.30 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 02 Apr 2009 23:26:32 -0700 (PDT)
Message-ID: <49D5AC13.4010200@webtide.com>
Date: Fri, 03 Apr 2009 17:26:27 +1100
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Lisa Dusseault <lisa.dusseault@gmail.com>
References: <0AAF0E03-4E58-419D-B9D0-5D775359784C@mnot.net>	<Pine.LNX.4.62.0903302336510.25058@hixie.dreamhostps.com>	<ec1a21d0903301752g3c38df30y34f043c0739e2aba@mail.gmail.com>	<Pine.LNX.4.62.0903310112140.25058@hixie.dreamhostps.com>	<1015C4DD-1A17-45C0-BB88-ED8A96BEAB9D@apple.com>	<ec1a21d0903301901l30d9dcd3uc377254ffb758132@mail.gmail.com>	<Pine.LNX.4.62.0903310212390.25058@hixie.dreamhostps.com> <ca722a9e0903311423i592d0e01ubda9d5617d63a127@mail.gmail.com>
In-Reply-To: <ca722a9e0903311423i592d0e01ubda9d5617d63a127@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Don't cross the streams...
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 06:25:35 -0000

Lisa Dusseault wrote:
> I would like to see the current techniques described.  I think there's
> enough variation in approaches, possibilities for advertising
> capabilities, and similar niggles, that even if we can roughly
> describe the idea behind "long-poll" in just a few sentences, we can
> still make it work better with some documentation.

Lisa,

some links that may be useful to whomever is writing up such
a description:

    http://en.wikipedia.org/wiki/Comet_(programming)
    http://cometdaily.com/2007/11/16/more-on-long-polling/
    http://cometdaily.com/2007/11/06/comet-is-always-better-than-polling/


I'll write up a summary of long-polling as used by cometd/Bayeux.

cheers

From ian@hixie.ch  Fri Apr  3 01:59:37 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C703F3A6868 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 01:59:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.123
X-Spam-Level: 
X-Spam-Status: No, score=-3.123 tagged_above=-999 required=5 tests=[AWL=-0.524, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id sP5z1+H9rETo for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 01:59:36 -0700 (PDT)
Received: from looneymail-a2.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 907513A69D0 for <hybi@ietf.org>; Fri,  3 Apr 2009 01:59:36 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a2.g.dreamhost.com (Postfix) with ESMTP id 77CB416D402; Fri,  3 Apr 2009 02:00:09 -0700 (PDT)
Date: Fri, 3 Apr 2009 09:00:09 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49D5AA5F.8030503@webtide.com>
Message-ID: <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was:  Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 08:59:37 -0000

On Fri, 3 Apr 2009, Greg Wilkins wrote:
> > >
> > > The websocket proposal is both good and bad. The Unicode datagram 
> > > API is probably very good for javascript programming.  But the zero 
> > > terminated wire protocol is horrible.
> 
> Because it means that I need to scan data received just to know if I 
> have all the data I need to proceed.
> 
> If I knew a content-length or a chunk length, then I would know 
> immediately that a read return that I have enough data to process.

Yes, that's the disadvantage of using sentinel markers. The reason for 
this design decision is to reduce the likelihood of buffer overruns on the 
server side. Many people writing Web software are unfortunately not the 
most competent software engineers, and this has to be taken into account 
when writing protocols that they are expected to implement. You are not 
representative of the majority, to say the least. :-)


> With null termination, I have to scan the data looking for nulls 
> (because I could receive multiple datagrams).  But if I'm scanning the 
> data then I'd like to be converting the UTF-8 at the same time, but as I 
> might not have all the data, I could end up with 4 bytes of a 5 byte 
> utf-8 character and so my utf-conversion has to be interrupted and I 
> have to save the chars that I've already converted and wait for another 
> packet to come in.

An interruptible UTF-8 decoder doesn't seem like a huge problem.


> So I expect there will be lots of buggy implementations out there that 
> will look like they work, but will fail on a nasty network.

I doubt most implementations would try to UTF-8 decode at the same time as 
looking for the sentinel byte. That's pretty advanced. I expect most would 
just scan for the 0xFF and then ship the whole string in one go to an off- 
the-shelf UTF-8 decoder.


> > > We see no reason why the wire protocol should deviate so far from 
> > > normal HTTP style framing.
> > 
> > What would you consider normal HTTP style framing here? What would 
> > HTTP framing give you that isn't available with the framing currently 
> > proposed with Web Sockets?
> 
> HTTP framing is:
>  + Some CRLF terminated headers containing meta-data
>  + A CRLF
>  + The content that is delimited either by:
>     * a known content length
>     * chunks, each with a known content length
>     * EOF

How is the headers part of this not the same as the 0xFF sentinel byte? 
You have to scan for a CR byte just like you would for Web Socket data, 
no? Indeed it seems the Web Socket format is simpler, since there is no 
complication with dealing with lone CR bytes with no matching LF, etc.


> We already have the code written and tested to handle this.

The Web Socket protocol is designed on the assumption that people will be 
writing new code, regardless of how similar we make it to existing 
protocols. This is similar to how many people have written their own CGI 
handling or form handling despite the existence of many libraries.

(I should note that HTTP is by far not the only protocol that could be 
adapted for this purpose, and indeed is far from the most appropriate one 
of the ones I looked at. IPv6 or Protocol Buffers, for example, would 
probably be more appropriate protocols to model for the purpose of 
arbitrary data transfer. Indeed, the protocol's one-byte frame type and 
the handling of binary data are both based on Protocol Buffers.)


> It supports more or less arbitrary extensible meta data along with the 
> content, so all the normal mime type mechanisms can be brought to bear.

I don't see the relevance of this; the idea is not to send MIME data, but 
to send either text or binary data, for subsequent handling by the client 
and server.


> > It seems that having a single byte at the start and a single byte at 
> > the end is all the framing one needs. The byte at the start gives the 
> > kind of frame, and the byte at the end is a byte that isn't valid in 
> > UTF-8 so it gets around the need to have a length declaration (such 
> > declaration are often the source of buffer overrun bugs).
> 
> It is certainly sufficient, but it is not efficient or convenient to 
> handle. More over it is new and different without any reason to be new 
> and different.

As noted above, it's not particularly new. The idea of indicating frame 
types with a single byte isn't new, nor are sentinels in strings (C has 
used null-terminated strings for decades), and nor are variable-sized 
integers for encoding lengths. In fact, pretty much every aspect of the 
Web Socket protocol is based on long-established practice.


> If there is some deficiency in HTTP framing?

HTTP framing is woefully inefficient for this purpose. I don't really know 
why we would even look at HTTP for this. If we were to use an existing 
text-based protocol's framing for this, I would imagine XMPP or IRC would 
be better choices. However, I see no reason why we would want to use 
text-based framing here.


> Is there some reason that you don't want to transport non-utf8 data? Eg 
> why can't I push an image or a sound bite?

The protocol does support binary data transfer, as noted earlier. You use 
frame type 0x80, followed by a variable-sized integer, followed by the 
binary data. This is not yet exposed in the API, and servers do not have 
to support it, but the protocol is designed to handle this and assuming 
the Web Socket protocol is adopted, future versions will almost certainly 
expose it. The protocol is also intended to be extensible to handle 
multiplexed streams and other more elaborate data types.


> > Specifically, to send binary data using the protocol you would send a 
> > 0x80 byte, the length (as a packed integer), and then the binary data.
> 
> Again, this feels like re-inventing the wheel.  HTTP already supports 
> binary data with known and unknown content lengths (chunking).

Many protocols support binary data with known and unknown content lengths, 
HTTP is certainly not alone here. :-)

In the case of the Web Socket protocol as designed today, there is trivial 
support for known-length binary data, and the protocol is designed in such 
a way that unknown-length binary data, using a scheme similar to HTTP 
chunking (but encoded in a more efficient maner), could be relatively 
easily supported.


> Things like packed integer lengths can come back to haunt you. With 
> Jetty I have had several users have problems because we initially had 
> length being an int and we had to convert to using a long.

I used the wrong term when I said "packed integer"; I was referring to 
variable-sized integers, not boundary-aligned integers, my apologies. The 
length encoding in Web Socket supports arbitrarily sized binary blobs, 
there is no upper limit. (Processors are expected to abort the connection 
if their peer tries to send data with an unsupported length.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Fri Apr  3 09:20:36 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9C6C63A6D94 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:20:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.326
X-Spam-Level: 
X-Spam-Status: No, score=-1.326 tagged_above=-999 required=5 tests=[AWL=0.650,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tEetAf1dapKk for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:20:35 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.234]) by core3.amsl.com (Postfix) with ESMTP id B52A328C166 for <hybi@ietf.org>; Fri,  3 Apr 2009 09:20:35 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1064332rvb.49 for <hybi@ietf.org>; Fri, 03 Apr 2009 09:21:38 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=AFzKt0jWhN6HDCHqHFXP0kQ4rEvpm8teYDuCMPs/Jv8=; b=ZofK3p1is+kuDttlOmwnnudJdYSS7+AHxaKvTW/wn9ofl3pLrbWteFOU/lxlU6f6hL N8i4VnLAjYB7jIOWFwPjUuc+xMMn4ln3uexfOL4kFajp1rPthrqsRArch3JzbHxvoFie 77rSibOxHareJd2ZnQAMzzErAfvgdFswHwMKc=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=GD0npJoeM4VzxIeDmLxrSptc0oUpLmdLj+8qufF/4AaG2kRlDy9pN6SfvX9nn4jJPi b05b6nDMDLXjPX82ZkU3knaslrjJ/WYSeziHaPDuhwlDYtUB1EWiSETfaVO4pN6fIy0u W7GztkfshwXzZB73quklKGm33NumKAjJlZ1ZU=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.198.9 with SMTP id a9mr636959rvq.7.1238775698487; Fri, 03  Apr 2009 09:21:38 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>
Date: Fri, 3 Apr 2009 09:21:38 -0700
X-Google-Sender-Auth: 4ed698ab0ed9079f
Message-ID: <1cb725390904030921p3f51c230kc14455064cc317d8@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=001517503d8676fa0c0466a8f17e
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 16:22:10 -0000

--001517503d8676fa0c0466a8f17e
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

In general, the benefit of using an existing protocol is that you take
advantage of a lot of infrastructure and in particular of protocols built ON
TOP of that protocol. Maybe that doesn't matter for train control, but it
matters a lot for webchat, which as far as I'm concerned is the single
biggest application of a "server->client" protocol. To me, this strongly
suggests Jabber, but I'll make the case for HTTP because I'm more familiar
with it:

 * new specification is very short and simple (e.g., not much more than [1])

 * client-libraries and intermediaries are more easily adapted (e.g. Curl,
httplib, ...)

 * domain knowledge is easily applied ("it's just like writing a PHP script
except you use Javascript and listen to Firefox rather than Apache")

 * standards built on HTTP will work more easily (cookies, AtomPub, WebDAV,
Subversion, etc.)

 * in general, the information system called "the Web" will have fewer
protocols and will be simpler, which will help it evolve and survive in the
long run. For example, if we add message signing to HTTP then we get it "for
free" in both directions. If we add a reliability protocol to HTTP then we
get it "for free" in both directions. ("for free" from a specification point
of view...from an implementation point of view it is "for cheap" because you
can reuse C code between Firefox and Apache if you like, or IIS and IE,
etc.)

That all said, my current feeling is still that Jabber is the best choice
because it is already the dominant protocol for the asynchronous
bi-directional use cases we're talking about. But I really cannot see the
advantage in inventing a new protocol from scratch. The Web and Internet are
complicated enough as they are without a proliferation of framing protocols.

 Paul Prescod

[1] http://www.ietf.org/internet-drafts/draft-lentczner-rhttp-00.txt

--001517503d8676fa0c0466a8f17e
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div>In general, the benefit of using an existing protocol is that you take=
 advantage of a lot of infrastructure and in particular of protocols built =
ON TOP of that protocol. Maybe that doesn&#39;t matter for train control, b=
ut it matters a lot for webchat, which as far as I&#39;m concerned is the s=
ingle biggest application of a &quot;server-&gt;client&quot; protocol. To m=
e, this strongly suggests Jabber, but I&#39;ll make the case for HTTP becau=
se I&#39;m more familiar with it:</div>
<div><div><br></div><div>=C2=A0* new specification is very short and simple=
 (e.g., not much more than [1])</div><div><br></div><div>=C2=A0* client-lib=
raries and intermediaries are more easily adapted (e.g. Curl, httplib, ...)=
</div>
<div><br></div><div>=C2=A0* domain knowledge is easily applied (&quot;it&#3=
9;s just like writing a PHP script except you use Javascript and listen to =
Firefox rather than Apache&quot;)</div><div><br></div><div>=C2=A0* standard=
s built on HTTP will work more easily (cookies, AtomPub, WebDAV, Subversion=
, etc.)<br>
</div></div><div><br></div><div>=C2=A0* in general, the information system =
called &quot;the Web&quot; will have fewer protocols and will be simpler, w=
hich will help it evolve and survive in the long run. For example, if we ad=
d message signing to HTTP then we get it &quot;for free&quot; in both direc=
tions. If we add a reliability protocol to HTTP then we get it &quot;for fr=
ee&quot; in both directions. (&quot;for free&quot; from a specification poi=
nt of view...from an implementation point of view it is &quot;for cheap&quo=
t; because you can reuse C code between Firefox and Apache if you like, or =
IIS and IE, etc.)</div>
<div><br></div><div>That all said, my current feeling is still that Jabber =
is the best choice because it is already the dominant protocol for the asyn=
chronous bi-directional use cases we&#39;re talking about. But I really can=
not see the advantage in inventing a new protocol from scratch. The Web and=
 Internet are complicated enough as they are without a proliferation of fra=
ming protocols.</div>
<div><br></div><div>=C2=A0Paul Prescod<br></div><div><div class=3D"gmail_qu=
ote"><br></div><div class=3D"gmail_quote">[1]=C2=A0<a href=3D"http://www.ie=
tf.org/internet-drafts/draft-lentczner-rhttp-00.txt">http://www.ietf.org/in=
ternet-drafts/draft-lentczner-rhttp-00.txt</a></div>
<div class=3D"gmail_quote"><br></div></div>

--001517503d8676fa0c0466a8f17e--

From mjs@apple.com  Fri Apr  3 09:31:20 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 31F143A6A07 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:31:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id NDMGsV8g3WNU for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:31:19 -0700 (PDT)
Received: from mail-out4.apple.com (mail-out4.apple.com [17.254.13.23]) by core3.amsl.com (Postfix) with ESMTP id 041363A680D for <hybi@ietf.org>; Fri,  3 Apr 2009 09:31:18 -0700 (PDT)
Received: from relay13.apple.com (relay13.apple.com [17.128.113.29]) by mail-out4.apple.com (Postfix) with ESMTP id E6E2A5D61D88 for <hybi@ietf.org>; Fri,  3 Apr 2009 09:32:20 -0700 (PDT)
Received: from relay13.apple.com (unknown [127.0.0.1]) by relay13.apple.com (Symantec Brightmail Gateway) with ESMTP id C19F42808C for <hybi@ietf.org>; Fri,  3 Apr 2009 09:32:20 -0700 (PDT)
X-AuditID: 1180711d-a36eabb000000259-76-49d63a14ebff
Received: from gertie.apple.com (gertie.apple.com [17.151.62.15]) by relay13.apple.com (Apple SCV relay) with ESMTP id 8869428094 for <hybi@ietf.org>; Fri,  3 Apr 2009 09:32:20 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by gertie.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KHJ00HVW9XV1320@gertie.apple.com> for hybi@ietf.org; Fri, 03 Apr 2009 09:32:20 -0700 (PDT)
Message-id: <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Ian Hickson <ian@hixie.ch>
In-reply-to: <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>
Date: Fri, 03 Apr 2009 09:32:19 -0700
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was:  Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 16:31:20 -0000

On Apr 3, 2009, at 2:00 AM, Ian Hickson wrote:

> On Fri, 3 Apr 2009, Greg Wilkins wrote:
>>>>
>>>> The websocket proposal is both good and bad. The Unicode datagram
>>>> API is probably very good for javascript programming.  But the zero
>>>> terminated wire protocol is horrible.
>>
>> Because it means that I need to scan data received just to know if I
>> have all the data I need to proceed.
>>
>> If I knew a content-length or a chunk length, then I would know
>> immediately that a read return that I have enough data to process.
>
> Yes, that's the disadvantage of using sentinel markers. The reason for
> this design decision is to reduce the likelihood of buffer overruns  
> on the
> server side. Many people writing Web software are unfortunately not  
> the
> most competent software engineers, and this has to be taken into  
> account
> when writing protocols that they are expected to implement. You are  
> not
> representative of the majority, to say the least. :-)

Do you mean a buffer overrun in the sense of a security problem? I  
don't see how the protocol framing makes this more or less likely. If  
anything, servers have to be careful to avoid being tricked into  
sending the end marker in the middle of a message. Multipart MIME has  
this problem, and implementing it in the context of form submission  
it's been difficult to get the security details right. I think a  
declared length has better security properties in a network protocol  
than a distinguished end marker.

>
>> With null termination, I have to scan the data looking for nulls
>> (because I could receive multiple datagrams).  But if I'm scanning  
>> the
>> data then I'd like to be converting the UTF-8 at the same time, but  
>> as I
>> might not have all the data, I could end up with 4 bytes of a 5 byte
>> utf-8 character and so my utf-conversion has to be interrupted and I
>> have to save the chars that I've already converted and wait for  
>> another
>> packet to come in.
>
> An interruptible UTF-8 decoder doesn't seem like a huge problem.

It's doable but you have to preserve state. However, having tricky  
edge cases that normally don't come up is a design problem. It is,  
once again, the sort of thing that tends to lead to security holes.

>
>> So I expect there will be lots of buggy implementations out there  
>> that
>> will look like they work, but will fail on a nasty network.
>
> I doubt most implementations would try to UTF-8 decode at the same  
> time as
> looking for the sentinel byte. That's pretty advanced. I expect most  
> would
> just scan for the 0xFF and then ship the whole string in one go to  
> an off-
> the-shelf UTF-8 decoder.
>
>
>>>> We see no reason why the wire protocol should deviate so far from
>>>> normal HTTP style framing.
>>>
>>> What would you consider normal HTTP style framing here? What would
>>> HTTP framing give you that isn't available with the framing  
>>> currently
>>> proposed with Web Sockets?
>>
>> HTTP framing is:
>> + Some CRLF terminated headers containing meta-data
>> + A CRLF
>> + The content that is delimited either by:
>>    * a known content length
>>    * chunks, each with a known content length
>>    * EOF
>
> How is the headers part of this not the same as the 0xFF sentinel  
> byte?
> You have to scan for a CR byte just like you would for Web Socket  
> data,
> no? Indeed it seems the Web Socket format is simpler, since there is  
> no
> complication with dealing with lone CR bytes with no matching LF, etc.

It definitely seems to me that HTTP-style headers are unnecessary and  
needlessly complex here. They may also look confusingly like a  
persistent http connection, which this is not. And headers would be  
pretty high overhead for short text-based messages, which is likely to  
be a common use case here. However framing with a length or known- 
length chunks tends to lead to simpler processing for the recipient.

Regards,
Maciej


From mjs@apple.com  Fri Apr  3 09:35:30 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EFA8E3A6A07 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:35:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WJVDDTcOvXUh for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 09:35:30 -0700 (PDT)
Received: from mail-out4.apple.com (mail-out4.apple.com [17.254.13.23]) by core3.amsl.com (Postfix) with ESMTP id 4C6B53A67EC for <hybi@ietf.org>; Fri,  3 Apr 2009 09:35:30 -0700 (PDT)
Received: from relay14.apple.com (relay14.apple.com [17.128.113.52]) by mail-out4.apple.com (Postfix) with ESMTP id 4A5C65D62076 for <hybi@ietf.org>; Fri,  3 Apr 2009 09:36:33 -0700 (PDT)
Received: from relay14.apple.com (unknown [127.0.0.1]) by relay14.apple.com (Symantec Brightmail Gateway) with ESMTP id 398A72808B for <hybi@ietf.org>; Fri,  3 Apr 2009 09:36:33 -0700 (PDT)
X-AuditID: 11807134-a985dbb000000ff0-df-49d63b113b69
Received: from elliott.apple.com (elliott.apple.com [17.151.62.13]) by relay14.apple.com (Apple SCV relay) with ESMTP id 20C9328085 for <hybi@ietf.org>; Fri,  3 Apr 2009 09:36:33 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by elliott.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KHJ0042VA4WQS30@elliott.apple.com> for hybi@ietf.org; Fri, 03 Apr 2009 09:36:33 -0700 (PDT)
Message-id: <E3BBCCB2-5FE4-4B5F-ADC2-6B95797F5460@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Paul Prescod <paul@prescod.net>
In-reply-to: <1cb725390904030921p3f51c230kc14455064cc317d8@mail.gmail.com>
Date: Fri, 03 Apr 2009 09:36:32 -0700
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <1cb725390904030921p3f51c230kc14455064cc317d8@mail.gmail.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 16:35:31 -0000

On Apr 3, 2009, at 9:21 AM, Paul Prescod wrote:

> In general, the benefit of using an existing protocol is that you  
> take advantage of a lot of infrastructure and in particular of  
> protocols built ON TOP of that protocol. Maybe that doesn't matter  
> for train control, but it matters a lot for webchat, which as far as  
> I'm concerned is the single biggest application of a "server- 
> >client" protocol. To me, this strongly suggests Jabber, but I'll  
> make the case for HTTP because I'm more familiar with it:

I think basing this protocol on the surface syntax of HTTP is a bad  
idea. HTTP is a request/response protocol, but the new protocol is  
simply two-way messaging. Messages are not (at least so far as the  
protocol knows) requests, nor are they responses to other messages.  
HTTP has the baked-in idea that messages are either requests or  
responses, that every request gets a response, and that every  
interaction is initiated by the client side. Using a similar framing  
syntax with totally different semantics is unlikely to make it easy to  
adapt existing tools.

Regards,
Maciej


From prescod@gmail.com  Fri Apr  3 10:31:10 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 873E93A6C87 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 10:31:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.543
X-Spam-Level: 
X-Spam-Status: No, score=-1.543 tagged_above=-999 required=5 tests=[AWL=0.433,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cEiOs4cEOMXQ for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 10:31:09 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.229]) by core3.amsl.com (Postfix) with ESMTP id E346D3A6BC6 for <hybi@ietf.org>; Fri,  3 Apr 2009 10:31:08 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1088510rvb.49 for <hybi@ietf.org>; Fri, 03 Apr 2009 10:32:11 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=nGWTFWGjEru3j05L+P7PrHj1SehNMCkmD+1awY/7fhc=; b=DdRtkkb1LpnvjlK7uGUeSFjK5H4Bv/cgT3uxzcE2N/9QPioPV8FLTNGa3b+EcYpL1i 6s2MghE/gWmf6J+ve2obwa0n0Di3k2/+iZn9eLmfLZ1hIiJjDKD2hVa8AbzgxZo3fPga 7WoiVMTHJ6kIYsB4PkBKREmN2dxqSkHnDLgEc=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=DIdmeV4hJaN7LEsYVU8XPiWE8MJjrKX9VEPnlwBh7EDaIEKopktcyTmi0q1es3I0ix VtD0KofIDvldmjKZTLycwNeSNu3MRBdf8ukcs6P0YWn20CZmva27dE+/kaYnBdaOJhbw AbzqQZDUKdk9QXRs0h68NkOVS7bXQ7OOvjNHE=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.18.15 with SMTP id v15mr645374rvi.197.1238779931781; Fri,  03 Apr 2009 10:32:11 -0700 (PDT)
In-Reply-To: <E3BBCCB2-5FE4-4B5F-ADC2-6B95797F5460@apple.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <1cb725390904030921p3f51c230kc14455064cc317d8@mail.gmail.com> <E3BBCCB2-5FE4-4B5F-ADC2-6B95797F5460@apple.com>
Date: Fri, 3 Apr 2009 10:32:11 -0700
X-Google-Sender-Auth: c2827cb7f7e11f28
Message-ID: <1cb725390904031032g44c9352bvc7af7332637b1059@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Maciej Stachowiak <mjs@apple.com>
Content-Type: multipart/alternative; boundary=000e0cd1518ec9eaec0466a9ed16
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 17:31:10 -0000

--000e0cd1518ec9eaec0466a9ed16
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

As I said, Jabber is probably a better fit. But if it were to use HTTP I
would not suggest to use the "surface syntax". If it were to use HTTP then
it would just add the minimum required to reverse the communications
direction:
 * http://www.ietf.org/internet-drafts/draft-lentczner-rhttp-00.txt

Then you would use one socket for initiator to listener communications and
another for the reverse. Any HTTP client library that is smart about
persistent connections could be quickly adapted to send "server->client"
messages.

But in any case, given that the dominant use case is web chat and Jabber is
the ascendent web chat protocol, I do think that Jabber is a better choice.

On Fri, Apr 3, 2009 at 9:36 AM, Maciej Stachowiak <mjs@apple.com> wrote:

>
> On Apr 3, 2009, at 9:21 AM, Paul Prescod wrote:
>
>  In general, the benefit of using an existing protocol is that you take
>> advantage of a lot of infrastructure and in particular of protocols built ON
>> TOP of that protocol. Maybe that doesn't matter for train control, but it
>> matters a lot for webchat, which as far as I'm concerned is the single
>> biggest application of a "server->client" protocol. To me, this strongly
>> suggests Jabber, but I'll make the case for HTTP because I'm more familiar
>> with it:
>>
>
> I think basing this protocol on the surface syntax of HTTP is a bad idea.
> HTTP is a request/response protocol, but the new protocol is simply two-way
> messaging. Messages are not (at least so far as the protocol knows)
> requests, nor are they responses to other messages. HTTP has the baked-in
> idea that messages are either requests or responses, that every request gets
> a response, and that every interaction is initiated by the client side.
> Using a similar framing syntax with totally different semantics is unlikely
> to make it easy to adapt existing tools.
>
> Regards,
> Maciej
>
>

--000e0cd1518ec9eaec0466a9ed16
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

As I said, Jabber is probably a better fit. But if it were to use HTTP I wo=
uld not suggest to use the &quot;surface syntax&quot;. If it were to use HT=
TP then it would just add the minimum required to reverse the communication=
s direction:<div>
<br></div><div>=C2=A0*=C2=A0<a href=3D"http://www.ietf.org/internet-drafts/=
draft-lentczner-rhttp-00.txt">http://www.ietf.org/internet-drafts/draft-len=
tczner-rhttp-00.txt</a></div><div><br></div><div>Then you would use one soc=
ket for initiator to listener communications and another for the reverse. A=
ny HTTP client library that is smart about persistent connections could be =
quickly adapted to send &quot;server-&gt;client&quot; messages.</div>
<div><br></div><div>But in any case, given that the dominant use case is we=
b chat and Jabber is the ascendent web chat protocol, I do think that Jabbe=
r is a better choice.<br><br><div class=3D"gmail_quote">On Fri, Apr 3, 2009=
 at 9:36 AM, Maciej Stachowiak <span dir=3D"ltr">&lt;<a href=3D"mailto:mjs@=
apple.com">mjs@apple.com</a>&gt;</span> wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex;"><div class=3D"im"><br>
On Apr 3, 2009, at 9:21 AM, Paul Prescod wrote:<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
In general, the benefit of using an existing protocol is that you take adva=
ntage of a lot of infrastructure and in particular of protocols built ON TO=
P of that protocol. Maybe that doesn&#39;t matter for train control, but it=
 matters a lot for webchat, which as far as I&#39;m concerned is the single=
 biggest application of a &quot;server-&gt;client&quot; protocol. To me, th=
is strongly suggests Jabber, but I&#39;ll make the case for HTTP because I&=
#39;m more familiar with it:<br>

</blockquote>
<br></div>
I think basing this protocol on the surface syntax of HTTP is a bad idea. H=
TTP is a request/response protocol, but the new protocol is simply two-way =
messaging. Messages are not (at least so far as the protocol knows) request=
s, nor are they responses to other messages. HTTP has the baked-in idea tha=
t messages are either requests or responses, that every request gets a resp=
onse, and that every interaction is initiated by the client side. Using a s=
imilar framing syntax with totally different semantics is unlikely to make =
it easy to adapt existing tools.<br>

<br>
Regards,<br><font color=3D"#888888">
Maciej<br>
<br>
</font></blockquote></div><br></div>

--000e0cd1518ec9eaec0466a9ed16--

From ian@hixie.ch  Fri Apr  3 14:06:22 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 883A53A6B39 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 14:06:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.098
X-Spam-Level: 
X-Spam-Status: No, score=-3.098 tagged_above=-999 required=5 tests=[AWL=-0.499, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IkA31cn23kL3 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 14:06:21 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 484923A6866 for <hybi@ietf.org>; Fri,  3 Apr 2009 14:06:21 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 415E427B79; Fri,  3 Apr 2009 14:07:24 -0700 (PDT)
Date: Fri, 3 Apr 2009 21:07:23 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>, Maciej Stachowiak <mjs@apple.com>
In-Reply-To: <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>
Message-ID: <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 21:06:22 -0000

On Fri, 3 Apr 2009, Paul Prescod wrote:
>
> In general, the benefit of using an existing protocol is that you take 
> advantage of a lot of infrastructure and in particular of protocols 
> built ON TOP of that protocol. Maybe that doesn't matter for train 
> control, but it matters a lot for webchat, which as far as I'm concerned 
> is the single biggest application of a "server->client" protocol.

It's _an_ application, but I don't think it's the biggest. I would expect 
games to be bigger, for example, and would expect background update 
notificiations in systems like Google Calendar to be even more common.

But it's hard to say in advance.


> but I'll make the case for HTTP because I'm more familiar with it:
> 
>  * new specification is very short and simple (e.g., not much more than [1])

The Web Socket draft is pretty simple.

I'm not convinced that it's really a fair comparison, either. rHTTP as a 
protocol is remarkably complex, in that it includes by reference the whole 
of HTTP, which is a remarkably complex protocol.


>  * client-libraries and intermediaries are more easily adapted (e.g. 
> Curl, httplib, ...)

One advantage of Web Socket is that the protocol is so simple that there's 
no need to adapt an existing library, new code can be easily written to 
support the protocol.


>  * domain knowledge is easily applied ("it's just like writing a PHP script
> except you use Javascript and listen to Firefox rather than Apache")

Actually using rHTTP for true full-duplex connections would be much more 
complicated, since you'd need to somehow get two separate connections from 
the client to the server into the same CGI script. With Web Socket, there 
is only one underlying TCP connection.


>  * standards built on HTTP will work more easily (cookies, AtomPub, 
> WebDAV, Subversion, etc.)

I agree entirely that if you want to do Subversion, or AtomPub, that you 
should not try to layer it on top of WebSocket. That isn't the use case 
we're targetting, though. (I don't know anyone who wants their server to 
be able to speak WebDAV to their JavaScript running in a browser.)


>  * in general, the information system called "the Web" will have fewer 
> protocols and will be simpler, which will help it evolve and survive in 
> the long run. For example, if we add message signing to HTTP then we get 
> it "for free" in both directions. If we add a reliability protocol to 
> HTTP then we get it "for free" in both directions. ("for free" from a 
> specification point of view...from an implementation point of view it is 
> "for cheap" because you can reuse C code between Firefox and Apache if 
> you like, or IIS and IE, etc.)

I don't think that a small number of very complex technologies is a better 
situation than a large number of simple technologies.


> That all said, my current feeling is still that Jabber is the best 
> choice because it is already the dominant protocol for the asynchronous 
> bi-directional use cases we're talking about. But I really cannot see 
> the advantage in inventing a new protocol from scratch. The Web and 
> Internet are complicated enough as they are without a proliferation of 
> framing protocols.

Jabber doesn't really meet the requirements for which Web Socket is 
designed [1]. Specifically:

 - It's not clear that one can easily upgrade from HTTP to a bidirectional 
   Jabber connection. Jabber over HTTP uses two TCP connections, as I 
   understand it.

 - It's not at all clear that a fully-conforming server-side component for 
   Jabber can be written in a few dozen lines of scripting code.

 - Jabber uses a separate sidechannel for streaming video, if I 
   understand correctly; it's not extensible to support efficient transfer 
   of binary data over the same channel.

 - It's not clear that Jabber's handling of virtual hosting is really 
   compatible with the "same-origin" model.

 - It's not clear that Jabber's protocol is designed to be resistant to 
   the attacks we expect (such as connecting to an SMTP server, or an 
   existing Jabber server, or whatnot).

 - It's not clear that Jabber's protocol supports the opt-in security 
   model that is desired for an HTML API.

Also, it's not clear to me that streaming XML is a particularly good 
solution. Experience with Web authors suggests that XML's draconian error 
handling and XML's namespaces are more confusing than desireable.

[1] http://www.ietf.org/mail-archive/web/hybi/current/msg00007.html


On Fri, 3 Apr 2009, Paul Prescod wrote:
>
> As I said, Jabber is probably a better fit. But if it were to use HTTP I 
> would not suggest to use the "surface syntax". If it were to use HTTP 
> then it would just add the minimum required to reverse the 
> communications direction:
>
>  * http://www.ietf.org/internet-drafts/draft-lentczner-rhttp-00.txt

I think rHTTP makes sense for certain use cases, but it doesn't fit the 
requirements that led to Web Socket's development (in particular, as you 
say, you'd need two TCP connections, which is non-trivial to set up in 
most scripting environments shared with a Web server).


On Fri, 3 Apr 2009, Maciej Stachowiak wrote:
> > 
> > Yes, that's the disadvantage of using sentinel markers. The reason for 
> > this design decision is to reduce the likelihood of buffer overruns on 
> > the server side.
> 
> Do you mean a buffer overrun in the sense of a security problem? I don't 
> see how the protocol framing makes this more or less likely.

Not necessarily a security problem. The problem with lengths is that since 
the data is UTF-8, but the length would have to be bytes, you are relying 
on the programmer being able to measure the byte-length of a Unicode 
string, which internally might well be stored as UTF-16 or some other 
encoding, only to be serialised to UTF-8 on output. Sentinel markers are 
significantly easier to deal with from the point of view of an amateur 
script author, IMHO.


> If anything, servers have to be careful to avoid being tricked into 
> sending the end marker in the middle of a message. Multipart MIME has 
> this problem, and implementing it in the context of form submission it's 
> been difficult to get the security details right.

The sentinel marker in Web Socket is 0xFF, which is never valid UTF-8. So 
long as the author properly encodes all text to UTF-8 on output, there 
shouldn't be a problem. If they fail to do that, then their system will 
have bigger problems than just being able to have 0xFF bytes inserted at 
random.


> I think a declared length has better security properties in a network 
> protocol than a distinguished end marker.

The Web Socket protocol uses a declared length for the case of binary 
data, where measuring the data length on the server is likely easy.


> > An interruptible UTF-8 decoder doesn't seem like a huge problem.
> 
> It's doable but you have to preserve state. However, having tricky edge 
> cases that normally don't come up is a design problem. It is, once 
> again, the sort of thing that tends to lead to security holes.

I agree, but in practice I don't think most authors are going to decode 
UTF-8 in parallel with looking for the sentinel. They're just going to 
buffer bytes until they have the sentinel, then get their scripting 
language to just treat the bytes as UTF-8 data.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Fri Apr  3 15:58:36 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AF5BB3A6912 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 15:58:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.651
X-Spam-Level: 
X-Spam-Status: No, score=-1.651 tagged_above=-999 required=5 tests=[AWL=0.325,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id iQUX8WRj6Dhf for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 15:58:35 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.230]) by core3.amsl.com (Postfix) with ESMTP id EF8403A68BF for <hybi@ietf.org>; Fri,  3 Apr 2009 15:58:34 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1182519rvb.49 for <hybi@ietf.org>; Fri, 03 Apr 2009 15:59:38 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=g2rpToju7Sv0Ad5/TaYBCy0hg1+qZSTuFsyALwQlJ60=; b=R26oW6CoyOAQlUCVlXbUTA0HyYgOXVLLX6m7wPk4K4iQ6/4hdHl+zRr7ojng4EGLrJ bq3X2h25TOe/RWN3bhduK69vzFesMumhOuCmoNZl1FL6qQcpT2qE4xelyg323UclrsBJ CQl1pNvEtvgYBxwkuaJ2HfEzDa307xo9LSoSU=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=i+MENWXxnal9DwSJ0rou2ntxWNZPg5fdJCRUuBqDwFUVv4QXBWpHetgjf1WkR1/LfG HcABf7nBXA9FG67Y/Sed7OhX9Lo1+D5B2/V1MT7dBKEeRJ816xObrbr6ERWiLr3SiP1M 9Q2hpqU9U1bz9fB8fxQtGcylGHegx52j0q4x4=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.48.6 with SMTP id a6mr752786rvk.161.1238799577993; Fri, 03  Apr 2009 15:59:37 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
Date: Fri, 3 Apr 2009 15:59:37 -0700
X-Google-Sender-Auth: 29ae5e7a231f33ac
Message-ID: <1cb725390904031559n2c52a3dcm7de5e4ae8719f715@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=000e0cd24f24cb531a0466ae80de
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 22:58:36 -0000

--000e0cd24f24cb531a0466ae80de
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On Fri, Apr 3, 2009 at 2:07 PM, Ian Hickson <ian@hixie.ch> wrote:

> > ...
> It's _an_ application, but I don't think it's the biggest. I would expect
> games to be bigger, for example, and would expect background update
> notificiations in systems like Google Calendar to be even more common.


Most real-time social games will have chat embedded or on the page around
them.


> > but I'll make the case for HTTP because I'm more familiar with it:
> >
> >  * new specification is very short and simple (e.g., not much more than
> [1])
>
> The Web Socket draft is pretty simple.


Yes, because all of the hard stuff is deferred or punted to another layer.

For example, how do I send a GIF over a Web Socket? HTTP and Jabber have
already solved that problem. How do I correlate a response to a message with
the original message?


> I'm not convinced that it's really a fair comparison, either. rHTTP as a
> protocol is remarkably complex, in that it includes by reference the whole
> of HTTP, which is a remarkably complex protocol.
>

I did say that the NEW specification is very short and simple.

>  * client-libraries and intermediaries are more easily adapted (e.g.
> Curl, httplib, ...)

One advantage of Web Socket is that the protocol is so simple that there's
> no need to adapt an existing library, new code can be easily written to
> support the protocol.
>

It's only simple because it doesn't do much yet.


> >  * domain knowledge is easily applied ("it's just like writing a PHP
> script
> > except you use Javascript and listen to Firefox rather than Apache")
>
> Actually using rHTTP for true full-duplex connections would be much more
> complicated, since you'd need to somehow get two separate connections from
> the client to the server into the same CGI script.


I don't know what you mean.

These services cannot be served with CGI. CGI is built around
request-response. Just as PHP scripts do not in general implement HTTP, I
wouldn't expect the handlers of these things to implement the protocol.

Here's how I expect it would work: there is a service optimized for
long-running TCP connections. Maybe it's written in C or Erlang. There are
PHP scripts (or whatever) for info coming in from web clients. When PHP
wants to tell the universe that "something happens" it puts a message in a
queue. Erlang or C delivers it.

Web developers do not want to become protocol handler developers and they
should not. The protocol should not be designed with the presumption that
they will be protocol developers. In fact: in my more than 10 years of web
development experience, I don't recall ever implementing an HTTP client or
server. What I have done frequently though, is trace the text flowing on the
wire, which is why a text-based protocol is preferable to me than a binary
one.


> With Web Socket, there is only one underlying TCP connection.


So what? How does that make my app simpler? Are you expecting PHP
programmers to write long-running TCP server apps?


> >  * standards built on HTTP will work more easily (cookies, AtomPub,
> > WebDAV, Subversion, etc.)
>
> I agree entirely that if you want to do Subversion, or AtomPub, that you
> should not try to layer it on top of WebSocket. That isn't the use case
> we're targetting, though. (I don't know anyone who wants their server to
> be able to speak WebDAV to their JavaScript running in a browser.)
>

If a chat transcript or game board is modelled as a resource then AtomPub is
one way you could edit that resource. Nevertheless, I agree that the
extensions built on top of Jabber are more directly appropriate than the
ones built on top of HTTP, due to the histories of those two protocols.


> I don't think that a small number of very complex technologies is a better
> situation than a large number of simple technologies.
>

Well we can't really remove the complexity of HTTP from the web or from Web
browsers or Web Servers. It's there already.

- It's not clear that one can easily upgrade from HTTP to a bidirectional
>   Jabber connection. Jabber over HTTP uses two TCP connections, as I
>   understand it.


I don't understand why it matters whether it is one connection or two.


> - It's not at all clear that a fully-conforming server-side component for
>   Jabber can be written in a few dozen lines of scripting code.


Most web developers are uninterested in writing protocol-handlers, and
rightly so.


> - Jabber uses a separate sidechannel for streaming video, if I
>   understand correctly; it's not extensible to support efficient transfer
>   of binary data over the same channel.


So? Why is it so important to have a single channel?


> - It's not clear that Jabber's handling of virtual hosting is really
>   compatible with the "same-origin" model.


That's just a question of how the connection is set up, not the protocol
running on top of the connection.


> - It's not clear that Jabber's protocol is designed to be resistant to
>   the attacks we expect (such as connecting to an SMTP server, or an
>   existing Jabber server, or whatnot).


That's just a question of how the connection is set up, not the protocol
running on top of the connection.


> - It's not clear that Jabber's protocol supports the opt-in security
>   model that is desired for an HTML API.


Not sure what that means.


> Also, it's not clear to me that streaming XML is a particularly good
> solution. Experience with Web authors suggests that XML's draconian error
> handling and XML's namespaces are more confusing than desireable.


The web developer is just putting messages into a queue. It's the protocol
handler (like Apache, ejabberd, postfix) that talks the protocol.

Can you point to a single protocol deployed on the Internet that is commonly
implemented and re-implemented and then re-re-implemented in scripts, rather
than being packaged in servers in libraries?

 Paul Prescod

--000e0cd24f24cb531a0466ae80de
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Fri, Apr 3, 2009 at 2:07 PM, Ian Hickson <span dir=3D"ltr">&lt;ian@hixie=
.ch&gt;</span> wrote:<br><div class=3D"gmail_quote"><blockquote class=3D"gm=
ail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt =
0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">&gt; ...<br>
</div>It&#39;s _an_ application, but I don&#39;t think it&#39;s the biggest=
. I would expect<br>
games to be bigger, for example, and would expect background update<br>
notificiations in systems like Google Calendar to be even more common.</blo=
ckquote><div><br>Most real-time social games will have chat embedded or on =
the page around them.<br>=C2=A0</div><blockquote class=3D"gmail_quote" styl=
e=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; =
padding-left: 1ex;">
<div class=3D"im">&gt; but I&#39;ll make the case for HTTP because I&#39;m =
more familiar with it:<br>
&gt;<br>
</div><div class=3D"im">&gt; =C2=A0* new specification is very short and si=
mple (e.g., not much more than [1])<br>
<br>
</div>The Web Socket draft is pretty simple.</blockquote><div><br>Yes, beca=
use all of the hard stuff is deferred or punted to another layer.<br><br>Fo=
r example, how do I send a GIF over a Web Socket? HTTP and Jabber have alre=
ady solved that problem. How do I correlate a response to a message with th=
e original message?<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">I&#39=
;m not convinced that it&#39;s really a fair comparison, either. rHTTP as a=
<br>
protocol is remarkably complex, in that it includes by reference the whole<=
br>
of HTTP, which is a remarkably complex protocol.<br><div class=3D"im"></div=
></blockquote><div class=3D"im"><br>I did say that the NEW specification is=
 very short and simple. <br><br>&gt; =C2=A0* client-libraries and intermedi=
aries are more easily adapted (e.g.<br>

&gt; Curl, httplib, ...)<br>
<br>
</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb=
(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">One advanta=
ge of Web Socket is that the protocol is so simple that there&#39;s<br>
no need to adapt an existing library, new code can be easily written to<br>
support the protocol.<br><div class=3D"im"></div></blockquote><div><br>It&#=
39;s only simple because it doesn&#39;t do much yet.<br>=C2=A0</div><blockq=
uote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 20=
4); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">
&gt; =C2=A0* domain knowledge is easily applied (&quot;it&#39;s just like w=
riting a PHP script<br>
&gt; except you use Javascript and listen to Firefox rather than Apache&quo=
t;)<br>
<br>
</div>Actually using rHTTP for true full-duplex connections would be much m=
ore<br>
complicated, since you&#39;d need to somehow get two separate connections f=
rom<br>
the client to the server into the same CGI script. </blockquote><div><br>I =
don&#39;t know what you mean. <br><br>These services cannot be served with =
CGI. CGI is built around request-response. Just as PHP scripts do not in ge=
neral implement HTTP, I wouldn&#39;t expect the handlers of these things to=
 implement the protocol.<br>
<br>Here&#39;s how I expect it would work: there is a service optimized for=
 long-running TCP connections. Maybe it&#39;s written in C or Erlang. There=
 are PHP scripts (or whatever) for info coming in from web clients. When PH=
P wants to tell the universe that &quot;something happens&quot; it puts a m=
essage in a queue. Erlang or C delivers it. <br>
<br>Web developers do not want to become protocol handler developers and th=
ey should not. The protocol should not be designed with the presumption tha=
t they will be protocol developers. In fact: in my more than 10 years of we=
b development experience, I don&#39;t recall ever implementing an HTTP clie=
nt or server. What I have done frequently though, is trace the text flowing=
 on the wire, which is why a text-based protocol is preferable to me than a=
 binary one.<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">With =
Web Socket, there is only one underlying TCP connection.</blockquote><div><=
br>So what? How does that make my app simpler? Are you expecting PHP progra=
mmers to write long-running TCP server apps?<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">&gt; =
=C2=A0* standards built on HTTP will work more easily (cookies, AtomPub,<br=
><div class=3D"im">

&gt; WebDAV, Subversion, etc.)<br>
<br>
</div>I agree entirely that if you want to do Subversion, or AtomPub, that =
you<br>
should not try to layer it on top of WebSocket. That isn&#39;t the use case=
<br>
we&#39;re targetting, though. (I don&#39;t know anyone who wants their serv=
er to<br>
be able to speak WebDAV to their JavaScript running in a browser.)<br><div =
class=3D"im"></div></blockquote><div><br>If a chat transcript or game board=
 is modelled as a resource then AtomPub is one way you could edit that reso=
urce. Nevertheless, I agree that the extensions built on top of Jabber are =
more directly appropriate than the ones built on top of HTTP, due to the hi=
stories of those two protocols.<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">I don=
&#39;t think that a small number of very complex technologies is a better<b=
r>
situation than a large number of simple technologies.<br><div class=3D"im">=
</div></blockquote><div><br>Well we can&#39;t really remove the complexity =
of HTTP from the web or from Web browsers or Web Servers. It&#39;s there al=
ready.<br>
<br></div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid=
 rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div cl=
ass=3D"im">
 - It&#39;s not clear that one can easily upgrade from HTTP to a bidirectio=
nal<br></div>
 =C2=A0 Jabber connection. Jabber over HTTP uses two TCP connections, as I<=
br>
 =C2=A0 understand it.</blockquote><div><br>I don&#39;t understand why it m=
atters whether it is one connection or two.<br>=C2=A0</div><blockquote clas=
s=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margi=
n: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
 - It&#39;s not at all clear that a fully-conforming server-side component =
for<br>
 =C2=A0 Jabber can be written in a few dozen lines of scripting code.</bloc=
kquote><div><br>Most web developers are uninterested in writing protocol-ha=
ndlers, and rightly so.<br>=C2=A0</div><blockquote class=3D"gmail_quote" st=
yle=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex=
; padding-left: 1ex;">
- Jabber uses a separate sidechannel for streaming video, if I<br>
 =C2=A0 understand correctly; it&#39;s not extensible to support efficient =
transfer<br>
 =C2=A0 of binary data over the same channel.</blockquote><div><br>So? Why =
is it so important to have a single channel?<br>=C2=A0</div><blockquote cla=
ss=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); marg=
in: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
 - It&#39;s not clear that Jabber&#39;s handling of virtual hosting is real=
ly<br>
 =C2=A0 compatible with the &quot;same-origin&quot; model.</blockquote><div=
><br>That&#39;s just a question of how the connection is set up, not the pr=
otocol running on top of the connection.<br></div><div>=C2=A0</div><blockqu=
ote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204=
); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
 - It&#39;s not clear that Jabber&#39;s protocol is designed to be resistan=
t to<br>
 =C2=A0 the attacks we expect (such as connecting to an SMTP server, or an<=
br>
 =C2=A0 existing Jabber server, or whatnot).</blockquote><div><br>That&#39;=
s just a question of how the connection is set up, not the protocol running=
 on top of the connection.<br>=C2=A0</div><blockquote class=3D"gmail_quote"=
 style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.=
8ex; padding-left: 1ex;">
 - It&#39;s not clear that Jabber&#39;s protocol supports the opt-in securi=
ty<br>
 =C2=A0 model that is desired for an HTML API.</blockquote><div><br>Not sur=
e what that means.<br>=C2=A0</div><blockquote class=3D"gmail_quote" style=
=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; p=
adding-left: 1ex;">
Also, it&#39;s not clear to me that streaming XML is a particularly good<br=
>
solution. Experience with Web authors suggests that XML&#39;s draconian err=
or<br>
handling and XML&#39;s namespaces are more confusing than desireable.</bloc=
kquote><div><br>The web developer is just putting messages into a queue. It=
&#39;s the protocol handler (like Apache, ejabberd, postfix) that talks the=
 protocol.<br>
<br>Can you point to a single protocol deployed on the Internet that is com=
monly implemented and re-implemented and then re-re-implemented in scripts,=
 rather than being packaged in servers in libraries?<br><br>=C2=A0Paul Pres=
cod<br>
<br></div></div>

--000e0cd24f24cb531a0466ae80de--

From stpeter@stpeter.im  Fri Apr  3 16:01:15 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 139403A6912 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:01:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.284
X-Spam-Level: 
X-Spam-Status: No, score=-2.284 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, SARE_MILLIONSOF=0.315]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 81sU0sUPo5SH for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:01:13 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 8CEFE3A6810 for <hybi@ietf.org>; Fri,  3 Apr 2009 16:01:13 -0700 (PDT)
Received: from squire.local (ip-65-183-77-85.rev.frii.com [65.183.77.85]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id 0FD7BE8002; Fri,  3 Apr 2009 17:02:12 -0600 (MDT)
Message-ID: <49D69573.8020508@stpeter.im>
Date: Fri, 03 Apr 2009 17:02:11 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49D1AE22.1080409@webtide.com>	<Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>	<49D5AA5F.8030503@webtide.com>	<Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>	<DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms070206020200000806060902"
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 23:01:15 -0000

This is a cryptographically signed message in MIME format.

--------------ms070206020200000806060902
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/3/09 3:07 PM, Ian Hickson wrote:
> On Fri, 3 Apr 2009, Paul Prescod wrote:
>
>> That all said, my current feeling is still that Jabber is the best 
>> choice because it is already the dominant protocol for the asynchronous 
>> bi-directional use cases we're talking about. But I really cannot see 
>> the advantage in inventing a new protocol from scratch. The Web and 
>> Internet are complicated enough as they are without a proliferation of 
>> framing protocols.
> 
> Jabber doesn't really meet the requirements for which Web Socket is 
> designed [1]. Specifically:
> 
>  - It's not clear that one can easily upgrade from HTTP to a bidirectional 
>    Jabber connection. 

Why upgrade from HTTP to XMPP? Just build XMPP support into the browser,
and you can refer to certain resources as XMPP URIs/IRIs and other
resources as HTTP URLs. Or the client (browser) and server can negotiate
an XMPP channel between themselves for more advanced bidirectional
functionality.

> Jabber over HTTP uses two TCP connections, as I 
>    understand it.

That's the HTTP binding for XMPP (BOSH). The native binding is TCP, and
I think that's what browser vendors would support (rather than BOSH,
which is a stop-gap until that happens).

>  - It's not at all clear that a fully-conforming server-side component for 
>    Jabber can be written in a few dozen lines of scripting code.

What exactly is the nature of this server-side component? I'd be curious
to know what functionality you expect.

That said, IMHO it should be pretty straightforward to build the
server-side XMPP handling as a shim onto an existing HTTP server,
because there are lots of XMPP servers and libraries available in
multiple languages:

http://xmpp.org/software/servers.shtml

http://xmpp.org/software/libraries.shtml

>  - Jabber uses a separate sidechannel for streaming video, if I 
>    understand correctly; it's not extensible to support efficient transfer 
>    of binary data over the same channel.

XMPP is optimized for streaming XML, in particular small snippets of
XML. It's hard to optimize for lots of small XML snippets *and* for
streaming video in the same protocol. Besides, the IETF has already
defined perfectly good technologies for streaming audio and video, in
the form of RTP. Why re-invent that wheel?

>  - It's not clear that Jabber's handling of virtual hosting is really 
>    compatible with the "same-origin" model.

Could you expand on that?

>  - It's not clear that Jabber's protocol is designed to be resistant to 
>    the attacks we expect (such as connecting to an SMTP server, or an 
>    existing Jabber server, or whatnot).

Have you completed a threat analysis for the projected use cases? I
don't see anything about that in draft-hixie-thewebsocketprotocol-07 but
perhaps it is defined elsewhere.

>  - It's not clear that Jabber's protocol supports the opt-in security 
>    model that is desired for an HTML API.

Could you expand on that, too? It's not clear to me what you mean by
opt-in security. XMPP supports TLS for channel encryption and SASL for
authentication. Both of those technologies are well-understood and quite
"pluggable" (different cipher suites for TLS, different mechanisms for
SASL). Aside from channel encryption and authentication, what security
features do you think are desirable?

> Also, it's not clear to me that streaming XML is a particularly good 
> solution. 

Solution for what? So far, XMPP has worked quite well for IM, presence,
lightweight middleware, RPC, cloud computing, gaming, and lots of other
use cases. Which use cases do you have in mind?

> Experience with Web authors suggests that XML's draconian error 
> handling and XML's namespaces are more confusing than desireable.

Whose experience? Where is this documented? The XMPP developer community
has produced a large volume of software (see links above) and runs a
large network of servers with millions of IM users. So far we have not
experienced any significant interop problems.

> [1] http://www.ietf.org/mail-archive/web/hybi/current/msg00007.html

Quoting therefrom, I'll address what you see as the requirements (though
it's still not clear to me for what applications or use cases these
requirements have been set forth):

> - The ability for Web pages to send arbitrary text to the server
>   from a script in a Web page, in the form of ordered discrete
>   blocks.
>
> - The ability for Web servers to send arbitrary text to a Web page
>   that has established a connection, in the form of ordered discrete
>   blocks.

As far as I can see, something very much like this already exists in
XMPP. I suppose it depends on what you mean by "arbitrary text in the
form of discrete ordered blocks".

> - Text messages should support all of Unicode.

Done in XMPP.

> - The protocol should support running over ports 80 and 443,
>   ideally with the ability to share the port on the server
>   with an HTTP server.

Why? Using a different port for a different protocol seems like the
proper approach.

> - It must be possible to implement a fully-conforming
>   server-side component for this in a few dozen lines of
>   scripting code, in the case where sharing the connection
>   with an HTTP server isn't required.

See above links about the large number of XMPP servers, clients, and
libraries.

> - The must be extensible so that it can be extended to support
>   binary data and structured data in both directions in the
>   future, without breaking deployments from the first generation.

What kind of binary data, and in what volume? We have ways of chunking
binary data over XMPP (without using a side channel), but those methods
are not intended for things like streaming video but instead for smaller
bits of binary such as thumbnails and files.

> - It must not be possible for the mechanism to be used to
>   successfully connect to existing servers and send them
>   arbitrary commands. In particular, SMTP, HTTP, HTTPS, IMAP,
>   POP, and similar protocols must be safe from this.

That's covered in XMPP -- it is a separate protocol run on a separate port.

> - It must be possible to connect to arbitrary remote hosts,
>   but connections to these hosts must only be allowed if the
>   remote host explicitly opts in to allowing that Web page to
>   connect (ideally using the 'origin' convention).

In XMPP we use a "dynamic federation" model, whereby a client connected
to a given server can route messages to other servers or to clients
connected to other servers -- like email, but optimized for real-time
delivery. That dynamic federation model doesn't require the remote
server or client to explicitly opt in to communication from each origin
host or client (although that is one deployment scenario). But perhaps I
misunderstand your idea of the requirements here.

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms070206020200000806060902
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms070206020200000806060902--

From prescod@gmail.com  Fri Apr  3 16:01:47 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3BB813A69D5 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:01:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.716
X-Spam-Level: 
X-Spam-Status: No, score=-1.716 tagged_above=-999 required=5 tests=[AWL=0.260,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UIzHF-rJVo63 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:01:46 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.230]) by core3.amsl.com (Postfix) with ESMTP id A0DC83A69C9 for <hybi@ietf.org>; Fri,  3 Apr 2009 16:01:46 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id g37so465182rvb.5 for <hybi@ietf.org>; Fri, 03 Apr 2009 16:02:49 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=SkZmB7Ms03f8fBm7xgnWnyTSIVFbMqMj3W6gZ8tWPrU=; b=pMBiddCtgqx3cJCrq/mtkpdVhdFctFBKcUkOpIjkqqIZIBOrXe5oIJkZVEQtNkXWOa 9DqBzQZgq9g/qiWDVcOVVt3YVD4ufLxCJUgkSTcyFymfbuHt7UEGxbGf6+yUrKUV3QRQ QkAMihmYpFj79D04AVIV5q7of62He1/5SBbVo=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=gyGfp14HzrmJaw2o4uKOyPVqy8mIMxgUZ9WKcjU7+MJFec0A954nSleYB1cLa/hMV8 KTlNtmLf+EbiL81eseRRoSM21Y/9RoepFdeutLXB9zUJ2W9pWYsfL4s04N+X06xnooEX T/a92pjnR8pmx9hdd4g8gVifvDpbnenObJTaU=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.114.15 with SMTP id r15mr763215rvm.42.1238799769839; Fri,  03 Apr 2009 16:02:49 -0700 (PDT)
In-Reply-To: <1cb725390904031559n2c52a3dcm7de5e4ae8719f715@mail.gmail.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com> <1cb725390904031559n2c52a3dcm7de5e4ae8719f715@mail.gmail.com>
Date: Fri, 3 Apr 2009 16:02:49 -0700
X-Google-Sender-Auth: aa96a6a588263670
Message-ID: <1cb725390904031602u6bf839f9kd11aa7a7e2670178@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=000e0cd22d463aa8380466ae8c96
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 23:01:47 -0000

--000e0cd22d463aa8380466ae8c96
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

By the way, you mentioned calendars: calDAV is the web protocol for
manipulating them. It might be a good fit for reverse HTTP.

 Paul Prescod

--000e0cd22d463aa8380466ae8c96
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

By the way, you mentioned calendars: calDAV is the web protocol for manipul=
ating them. It might be a good fit for reverse HTTP.<br><br>=C2=A0Paul Pres=
cod<br><br>

--000e0cd22d463aa8380466ae8c96--

From gregw@webtide.com  Fri Apr  3 16:03:14 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1F42C3A691B for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:03:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 92d6uwK8pb1u for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:03:13 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id 83BA428C116 for <hybi@ietf.org>; Fri,  3 Apr 2009 16:02:55 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so1157595tid.25 for <hybi@ietf.org>; Fri, 03 Apr 2009 16:03:57 -0700 (PDT)
Received: by 10.110.53.14 with SMTP id b14mr2436369tia.47.1238799837544; Fri, 03 Apr 2009 16:03:57 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id a4sm1981867tib.31.2009.04.03.16.03.54 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 03 Apr 2009 16:03:56 -0700 (PDT)
Message-ID: <49D695D5.4050902@webtide.com>
Date: Sat, 04 Apr 2009 10:03:49 +1100
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 23:03:14 -0000

Ian Hickson wrote:
> On Fri, 3 Apr 2009, Paul Prescod wrote:
>> ... Maybe that doesn't matter for train 
>> control, but it matters a lot for webchat, which as far as I'm concerned 
>> is the single biggest application of a "server->client" protocol.
> 
> It's _an_ application, but I don't think it's the biggest. I would expect 
> games to be bigger, for example, and would expect background update 
> notificiations in systems like Google Calendar to be even more common.
> 
> But it's hard to say in advance.


Exactly!

If the IETF does come out with a bidirectional HTTP recommendation,
it will be a huge once in a decade opportunity to upgrade the
protocols supported by the network infrastructure.

Websockets sentinel UTF-8 datagrams may well be ideally suited
for a specific type of application, but can we really say that this
will be the only content type that we need to transport asynchronously
server to client for the next decade or two?

What if UTF-16 becomes the preferred encoding? or gzipped UTF-16, or
some as-of-yet-un-invented compression algorithm specially designed
for packets of UTF-16 data holding JSON becomes in common usage.

I'm already working with clients that wish to push images.


> The Web Socket protocol uses a declared length for the case of binary 
> data, where measuring the data length on the server is likely easy.

If web socket supports arbitrary binary data, then why is there
a need for the sentinel mechanism and special status for UTF-8 data?

Why can't there just be one mechanism that does binary content
with an assumption that it will be utf-8 unless otherwise
indicated.

It is already a pain with HTTP that there is both Content-Length
and chunking framing, but at least they are both variations of
a theme.

Note also with websocket binary transfer, I believe it
is a limitation that there cannot be meta data (headers)
associated with a datagram. Thus information like  mime types,
char-sets content-encodings and transfer encodings will need
to be handled with stateful mechanisms based on assumption,
connection and/or previous sent datagrams.

HTTP framing, while not the best framing possible does
contain a lot of derived wisdom and required flexibility.
While there are deficiencies in with HTTP, I'm not sure
that it is the remit of bidirectional HTTP to unilaterally
try to solve them.

Moreover, I don't think that a true solution can be
found by restricting the problem domain (ie utf-8 datagrams)
to one that is more easily solvable.


However, I do see a gap between rHTTP and Websocket.

While rHTTP does build on existing framing (as I advocate),
it is still a request/response paradigm.   What websocket does
support is multiple responses for a single request.

What I'm looking for is the WebSocket API in javascript,
with rHTTP data framing, but the websocket capability to
send multiple content responses to a single request, but
with the rHTTP capability of associating optional meta
data with arbitrary content.

I want the best of both worlds :)

cheers






































From gregw@webtide.com  Fri Apr  3 16:25:43 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CC7B23A688D for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:25:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.442
X-Spam-Level: 
X-Spam-Status: No, score=-2.442 tagged_above=-999 required=5 tests=[AWL=-0.157, BAYES_00=-2.599, SARE_MILLIONSOF=0.315]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kexeMQi8JtSE for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 16:25:43 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id A804E3A6869 for <hybi@ietf.org>; Fri,  3 Apr 2009 16:25:42 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so1161104tid.25 for <hybi@ietf.org>; Fri, 03 Apr 2009 16:26:45 -0700 (PDT)
Received: by 10.110.50.19 with SMTP id x19mr2480879tix.31.1238801204948; Fri, 03 Apr 2009 16:26:44 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm2721264tib.28.2009.04.03.16.26.41 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 03 Apr 2009 16:26:43 -0700 (PDT)
Message-ID: <49D69B2C.5070405@webtide.com>
Date: Sat, 04 Apr 2009 10:26:36 +1100
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Maciej Stachowiak <mjs@apple.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>
In-Reply-To: <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was:  Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Apr 2009 23:25:43 -0000

Maciej Stachowiak wrote:

 > It definitely seems to me that HTTP-style headers are unnecessary and
 > needlessly complex here. [...] And headers would be
 > pretty high overhead for short text-based messages, which is likely to
 > be a common use case here.

While I agree that when you look at HTTP-style headers they
are unnecessarily verbose, redundant and frequently ignored.

But is this actually a problem?  and if so, is it a problem
that should be solved by the bidirectional mechanism in
isolation?



There are many (albeit mostly anecdotal) reasons that I don't
think that HTTP-style framing is really a problem:

  + If datasize was really is an issue, then there is much
    more to be gained by replacing common content formats
    such as XML, JSON, HTML etc.

  + If datasize for really in an issue, then the flexibility
    of HTTP framing has already demonstrated an ability to
    adapt by supporting compression as it evolves (was nothing,
    then compress, now gzip, next ???).  Supporting
    extensible meta-data as HTTP headers do is key to supporting
    such compression.

  + My understanding is that from a network perspective
    there is very little benefit in sending less that an MTU
    of data. 1 byte or 1500 bytes will transit the network
    at the same speed/latency.   So much so, that TCP/IP will
    induce latency in an effort to fill up packets with more
    data

  + Headers are optional. If your content can be interpreted
    with assumed or contextual meta data, then those headers
    don't need to be sent.    If we designed a protocol that
    assumed chunking then a minimal framing could be <10 bytes

  + Headers are optional, yet there does not appear to be
    any significant push to get browsers or servers to reduce
    many of the under utilized headers in normal web browsing.
    How many millions of times per second is
      Accept-Language: en-us,en;q=0.5
    sent and ignored?   If we want to start byte stuffing
    in internet protocols, there there is a lot of low
    hanging fruit we can get before taking out a CRLF that
    would allow us to have optional meta data.



So I do agree that HTTP-style, humanly readable protocols are
inefficient and represent some overhead.   But this has been
an overhead worth paying for FTP, HTTP, SMTP, POP, IMAP, etc.

One day, this style of protocol will be replaced by
something better.  But is this the day? Is bidirectional
HTTP the use-case that necessitates the divergence from
decades of experience?


cheers









From gregw@webtide.com  Fri Apr  3 17:15:33 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B6B0C3A6883 for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 17:15:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.577
X-Spam-Level: 
X-Spam-Status: No, score=-2.577 tagged_above=-999 required=5 tests=[AWL=0.023,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id W8+XKE-NrJ0p for <hybi@core3.amsl.com>; Fri,  3 Apr 2009 17:15:33 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id C90243A687E for <hybi@ietf.org>; Fri,  3 Apr 2009 17:15:32 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so1168957tid.25 for <hybi@ietf.org>; Fri, 03 Apr 2009 17:16:35 -0700 (PDT)
Received: by 10.110.63.17 with SMTP id l17mr2519055tia.51.1238804195324; Fri, 03 Apr 2009 17:16:35 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d7sm1997745tib.22.2009.04.03.17.16.32 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 03 Apr 2009 17:16:34 -0700 (PDT)
Message-ID: <49D6A6DB.9040701@webtide.com>
Date: Sat, 04 Apr 2009 11:16:27 +1100
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Peter Saint-Andre <stpeter@stpeter.im>
References: <49D1AE22.1080409@webtide.com>	<Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>	<49D5AA5F.8030503@webtide.com>	<Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>	<DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>	<Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com> <49D69573.8020508@stpeter.im>
In-Reply-To: <49D69573.8020508@stpeter.im>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 04 Apr 2009 00:15:33 -0000

Peter Saint-Andre wrote:

> Why upgrade from HTTP to XMPP? Just build XMPP support into the browser,
> and you can refer to certain resources as XMPP URIs/IRIs and other
> resources as HTTP URLs. Or the client (browser) and server can negotiate
> an XMPP channel between themselves for more advanced bidirectional
> functionality.

Peter,

if opening arbitrary connections for arbitrary protocols was always an
option, then we would not be having this list.

I wish the XMPP's of the world luck in their efforts to establish
world domination (like HTTP has) and thus to convince all the
firewall/proxy's to support XMPP connections.  I'm not being
sarcastic here, as eventually some protocol will have a sufficiently
compelling story that it will become widely supported.

But isn't it the focus of this list to deal with the situation
where HTTP is the only available protocol.


What I'd like to see from the IETF are recommendations about
how to better use HTTP as it exists today for bidirectional
purposes: eg.

   + should long-polling clients be exempt from the 2 connection
     recommendation of RFC2616? if so why
   + should we be recommending that HTTP clients allow some
     application control of pipeling etc.
   + can we agree a standard header to identify long poll
     requests.

Such recommendations could start having an immediate impact.


Then with a longer term focus, we can consider things such
as websocket, rHTTP or 1xx responses with content  as
incremental improvements to HTTP.    But it will still
be HTTP until such time as others come up with something
better than we all adopt.


cheers






From sh@defuze.org  Sat Apr  4 01:27:24 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BB2073A6970 for <hybi@core3.amsl.com>; Sat,  4 Apr 2009 01:27:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.965
X-Spam-Level: 
X-Spam-Status: No, score=-1.965 tagged_above=-999 required=5 tests=[AWL=0.300,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xSe5cCX4Px62 for <hybi@core3.amsl.com>; Sat,  4 Apr 2009 01:27:23 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id C28DC3A680B for <hybi@ietf.org>; Sat,  4 Apr 2009 01:27:23 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n348SPUN012895 for <hybi@ietf.org>; Sat, 4 Apr 2009 03:28:26 -0500
Message-ID: <49D71A28.5050703@defuze.org>
Date: Sat, 04 Apr 2009 10:28:24 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090318)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49D1AE22.1080409@webtide.com>	<Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>	<49D5AA5F.8030503@webtide.com>	<Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>	<DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <49D69B2C.5070405@webtide.com>
In-Reply-To: <49D69B2C.5070405@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSockets feedback (Was:  Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 04 Apr 2009 08:27:24 -0000

>
> One day, this style of protocol will be replaced by
> something better.  But is this the day? Is bidirectional
> HTTP the use-case that necessitates the divergence from
> decades of experience?
Interestingly there is even a protocol that claims it'd be the perfect 
ground for higher level protocols: BEEP: http://beepcore.org/

- Sylvain

From prescod@gmail.com  Sun Apr  5 09:20:42 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 544F93A6B16 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 09:20:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.015
X-Spam-Level: 
X-Spam-Status: No, score=-1.015 tagged_above=-999 required=5 tests=[AWL=-0.528, BAYES_05=-1.11, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XRynJBy9aQyR for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 09:20:41 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.236]) by core3.amsl.com (Postfix) with ESMTP id 919113A6923 for <hybi@ietf.org>; Sun,  5 Apr 2009 09:20:41 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1585932rvb.49 for <hybi@ietf.org>; Sun, 05 Apr 2009 09:21:38 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=+pJYfKvrIEeV5wXXi5kGslBMVEasBaY7bgOmhM7vlZ0=; b=OicBEd7QLKzm5Q3byC0iFBihlEIDdVviAUABL8yMliqv6/pU61km7Atcx4k9JJDnmm ULdrdqjwi/y4lTjJ92oSlktKktFbgnjwJ5zbl27xZOrMM5d4YE0X7FRcZTTQD6CFtTHX /Ks+Tc56IdWvc0IW67a5UI4H0EZso1W2Ambd4=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=uRCeBdqXedj3LaPTeie4gJffywBu6gfUkLgn3rtARHoEnTGQXRVAZqpFS/5LUZmf6g hWmLCBWubb3q+jAukvTMxisLwP0pW1OF8D5RL4HPlNx0vL/xWmfakOL97h+jPUK7MyRu UAw+KVcqL7tgGpzwEG/Bi0cuncZtcQBT/jfcc=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.113.3 with SMTP id q3mr925691rvm.82.1238948498610; Sun, 05  Apr 2009 09:21:38 -0700 (PDT)
In-Reply-To: <49D6A6DB.9040701@webtide.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com> <49D69573.8020508@stpeter.im> <49D6A6DB.9040701@webtide.com>
Date: Sun, 5 Apr 2009 09:21:38 -0700
X-Google-Sender-Auth: 2055b6668a15affc
Message-ID: <1cb725390904050921g196f4979rf2959d2786d06e2d@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: multipart/alternative; boundary=000e0cd22de6279c9a0466d12de3
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 16:20:42 -0000

--000e0cd22de6279c9a0466d12de3
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

Like Roy, I think it's kind of counter-productive that everyone comes to the
list with their own set of requirements and goals.

Do we agree that:

 a) the list should have a wiki?

 b) the requirements should be documented on the wiki?

If so, does anyone have a preference for wiki platform? Choosing a bikeshed
colour is actually marginally better than going in circles.

I'd start the conversation by suggesting Google Code, which combines wiki
with

 * issue tracker (for unanswered questions)
 * version control for more formal specs
 * repository for diagrams or code samples

 Paul Prescod

--000e0cd22de6279c9a0466d12de3
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Like Roy, I think it&#39;s kind of counter-productive that everyone comes t=
o the list with their own set of requirements and goals.<br><div class=3D"g=
mail_quote"><div><br>Do we agree that:<br>=C2=A0<br>=C2=A0a) the list shoul=
d have a wiki?<br>
<br>=C2=A0b) the requirements should be documented on the wiki?<br><br>If s=
o, does anyone have a preference for wiki platform? Choosing a bikeshed col=
our is actually marginally better than going in circles.<br><br>I&#39;d sta=
rt the conversation by suggesting Google Code, which combines wiki with<br>
<br>=C2=A0* issue tracker (for unanswered questions) <br>=C2=A0* version co=
ntrol for more formal specs<br>=C2=A0* repository for diagrams or code samp=
les<br><br>=C2=A0Paul Prescod<br><br></div></div>

--000e0cd22de6279c9a0466d12de3--

From prescod@gmail.com  Sun Apr  5 09:59:22 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 141FE3A681D for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 09:59:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.477
X-Spam-Level: 
X-Spam-Status: No, score=-0.477 tagged_above=-999 required=5 tests=[AWL=-0.915, BAYES_40=-0.185, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yB5V31aIZItm for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 09:59:21 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.239]) by core3.amsl.com (Postfix) with ESMTP id 436A43A677D for <hybi@ietf.org>; Sun,  5 Apr 2009 09:59:21 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1591952rvb.49 for <hybi@ietf.org>; Sun, 05 Apr 2009 10:00:26 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:date :x-google-sender-auth:message-id:subject:from:to:content-type; bh=cWLijH3XCTmNjuAdsAbnpDhHU8RdzNoFTLmbHDlM6Ws=; b=TWF0LAPyrZm+eDPG2fZtgM9LBXPgGOEigIKbUm0+nwPp6ShIt4vJ/X50ziU+Hq7Cml jhKviaD99oThOuCEPVg7lWq9nEFsbiV44Q871NuvZXUgoI8eiZldqh30FxvhRKssQnQI mDzzj+kz8wvDvtGYPtS9yyqo46+sHvTFdyh40=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:date:x-google-sender-auth:message-id:subject :from:to:content-type; b=SwQRsJPhIhA9BqT91wqxsLDX42H0QR8hPQ0MkNS4xcQ+iq6Blsn27SsE99l0kzGeDj QqPpF0GtgwuWQ0FBn1XxdsezJK1+nx7sHS738xkCKjDtBAmaNboHSM8zGv9C0m5wQv17 4LGGGeuihTdiy0BjeNTnfPbbjQjUr3yXj7oNQ=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.50.11 with SMTP id c11mr938738rvk.28.1238950826342; Sun,  05 Apr 2009 10:00:26 -0700 (PDT)
Date: Sun, 5 Apr 2009 10:00:26 -0700
X-Google-Sender-Auth: ddc6c856e2a51f5a
Message-ID: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: hybi@ietf.org
Content-Type: multipart/alternative; boundary=000e0cd14606e5f8fa0466d1b72b
Subject: [hybi] Trains, planes and web protocols
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 16:59:22 -0000

--000e0cd14606e5f8fa0466d1b72b
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

So let's say that I am controlling my model trains from my web browser using
web sockets. The trains set sends a message to say "the train is approaching
the junction: you have 20 seconds to make a decision or the train will
crash." At that second, the connection is lost for any of a thousand
reasons, from a neighbour picking up her wireless phone to a roommate
tripping over a power cord. The problem lasts for one second, but the
message is lost. How will Web Sockets help the client and server to get back
in sync?

My understanding is that the protocol defers this problem to the application
level, despite the fact that almost every application is susceptible to
inaccuracies caused by intermittent message delivery problems. I believe
that it is unacceptable to solve the simple problems in the protocol and
defer all of the difficult stuff to the application.

I propose that the standard should address issues of reliable message
delivery, or build on a standard that has already done so.

 Paul Prescod

--000e0cd14606e5f8fa0466d1b72b
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

So let&#39;s say that I am controlling my model trains from my web browser =
using web sockets. The trains set sends a message to say &quot;the train is=
 approaching the junction: you have 20 seconds to make a decision or the tr=
ain will crash.&quot; At that second, the connection is lost for any of a t=
housand reasons, from a neighbour picking up her wireless phone to a roomma=
te tripping over a power cord. The problem lasts for one second, but the me=
ssage is lost. How will Web Sockets help the client and server to get back =
in sync?<br>
<br>My understanding is that the protocol defers this problem to the applic=
ation level, despite the fact that almost every application is susceptible =
to inaccuracies caused by intermittent message delivery problems. I believe=
 that it is unacceptable to solve the simple problems in the protocol and d=
efer all of the difficult stuff to the application.<br>
<br>I propose that the standard should address issues of reliable message d=
elivery, or build on a standard that has already done so.<br><br>=C2=A0Paul=
 Prescod<br><br>

--000e0cd14606e5f8fa0466d1b72b--

From prescod@gmail.com  Sun Apr  5 10:05:54 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 749B63A6B21 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 10:05:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.64
X-Spam-Level: 
X-Spam-Status: No, score=-0.64 tagged_above=-999 required=5 tests=[AWL=-0.523,  BAYES_20=-0.74, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rDko5R-YfOH2 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 10:05:53 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.238]) by core3.amsl.com (Postfix) with ESMTP id A4A6C3A6A2B for <hybi@ietf.org>; Sun,  5 Apr 2009 10:05:53 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1593056rvb.49 for <hybi@ietf.org>; Sun, 05 Apr 2009 10:06:58 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to :content-type; bh=uIwlHkRqlXsrr4+I5Frj7hkjc1MN49ATlOHJeiTJk40=; b=SPsgOb73FUaZM+CTKaoEtBy67klsm0FUo5lIz6L1Ks8GB2XjPOn3+hQ4dZNvDA//KH NHt6rHxVD56pYUfypcfA0tTt5nrSYchmRNtuXvdWYlFAvBqrcO6Gv8m60EF814QUgv1i ac4pdZUe4MZdVEVqsFB4IklwEc2QVmcnSZHEQ=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:content-type; b=OmpjJcErtr/uV+1zpG54T1xOL4qOfeiLCXpCM/F4cKpIFqLzVqvyGr4My9smzXewS9 PnUydlWEeH4OUa1FRIuZ79RpOKUF5IHHDIp61LfNHQJQVzRhdAtGr3lmLBvGw3ZsXeB5 gYAQnlh7JxhH1Sb6xanSjmILULfQHG5S3wJvc=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.114.19 with SMTP id r19mr937755rvm.24.1238951218720; Sun,  05 Apr 2009 10:06:58 -0700 (PDT)
In-Reply-To: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>
Date: Sun, 5 Apr 2009 10:06:58 -0700
X-Google-Sender-Auth: b316d95ecce0b2fe
Message-ID: <1cb725390904051006o27430f51r19b8f0ba4d1aa7f8@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: hybi@ietf.org
Content-Type: multipart/alternative; boundary=000e0cd29cb049305e0466d1cfa4
Subject: Re: [hybi] Trains, planes and web protocols
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 17:05:54 -0000

--000e0cd29cb049305e0466d1cfa4
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

Just to clarify: I didn't mean to say that every client and server should
support reliable delivery, but those that need it should have an obvious
path to do so, rather than: "Just build it yourself on the protocol."

On Sun, Apr 5, 2009 at 10:00 AM, Paul Prescod <paul@prescod.net> wrote:

> So let's say that I am controlling my model trains from my web browser
> using web sockets. The trains set sends a message to say "the train is
> approaching the junction: you have 20 seconds to make a decision or the
> train will crash." At that second, the connection is lost for any of a
> thousand reasons, from a neighbour picking up her wireless phone to a
> roommate tripping over a power cord. The problem lasts for one second, but
> the message is lost. How will Web Sockets help the client and server to get
> back in sync?
>
> My understanding is that the protocol defers this problem to the
> application level, despite the fact that almost every application is
> susceptible to inaccuracies caused by intermittent message delivery
> problems. I believe that it is unacceptable to solve the simple problems in
> the protocol and defer all of the difficult stuff to the application.
>
> I propose that the standard should address issues of reliable message
> delivery, or build on a standard that has already done so.
>
>  Paul Prescod
>
>

--000e0cd29cb049305e0466d1cfa4
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Just to clarify: I didn&#39;t mean to say that every client and server shou=
ld support reliable delivery, but those that need it should have an obvious=
 path to do so, rather than: &quot;Just build it yourself on the protocol.&=
quot;<br>
<br><div class=3D"gmail_quote">On Sun, Apr 5, 2009 at 10:00 AM, Paul Presco=
d <span dir=3D"ltr">&lt;<a href=3D"mailto:paul@prescod.net">paul@prescod.ne=
t</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"borde=
r-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-le=
ft: 1ex;">
So let&#39;s say that I am controlling my model trains from my web browser =
using web sockets. The trains set sends a message to say &quot;the train is=
 approaching the junction: you have 20 seconds to make a decision or the tr=
ain will crash.&quot; At that second, the connection is lost for any of a t=
housand reasons, from a neighbour picking up her wireless phone to a roomma=
te tripping over a power cord. The problem lasts for one second, but the me=
ssage is lost. How will Web Sockets help the client and server to get back =
in sync?<br>

<br>My understanding is that the protocol defers this problem to the applic=
ation level, despite the fact that almost every application is susceptible =
to inaccuracies caused by intermittent message delivery problems. I believe=
 that it is unacceptable to solve the simple problems in the protocol and d=
efer all of the difficult stuff to the application.<br>

<br>I propose that the standard should address issues of reliable message d=
elivery, or build on a standard that has already done so.<br><font color=3D=
"#888888"><br>=C2=A0Paul Prescod<br><br>
</font></blockquote></div><br>

--000e0cd29cb049305e0466d1cfa4--

From tonyg@lshift.net  Sun Apr  5 11:53:59 2009
Return-Path: <tonyg@lshift.net>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6A1F73A6B90 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 11:53:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.392
X-Spam-Level: 
X-Spam-Status: No, score=-1.392 tagged_above=-999 required=5 tests=[AWL=-1.207, BAYES_40=-0.185]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3zKY3L7dPUMb for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 11:53:58 -0700 (PDT)
Received: from agentk.lshift.net (host226.lshift.net [195.172.218.226]) by core3.amsl.com (Postfix) with ESMTP id 876863A6933 for <hybi@ietf.org>; Sun,  5 Apr 2009 11:53:58 -0700 (PDT)
Received: from nodachi.lshift.net ([10.224.189.13] helo=walk.local) by agentk.lshift.net with esmtpsa (TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.63) (envelope-from <tonyg@lshift.net>) id 1LqXUi-0002Zv-En; Sun, 05 Apr 2009 19:55:00 +0100
Message-ID: <49D8FE80.2010100@lshift.net>
Date: Sun, 05 Apr 2009 11:54:56 -0700
From: Tony Garnock-Jones <tonyg@lshift.net>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Paul Prescod <paul@prescod.net>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>
In-Reply-To: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Score: -1.2
Cc: hybi@ietf.org, rabbitmq info <info@rabbitmq.com>
Subject: Re: [hybi] Trains, planes and web protocols
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 18:53:59 -0000

Paul Prescod wrote:
> I propose that the standard should address issues of reliable message 
> delivery, or build on a standard that has already done so.

I disagree. Reliable delivery (for almost every useful definition of the 
term) is an application-level concern. To me there's a natural hierarchy 
here:

   * transport for getting blobs from named node to named node,
     in a best-effort fashion
   * distributed object system (doesn't have to be complex)
   * messaging system (e.g. pubsub, AMQP etc) and other apps

HTTP serves as the transport. URLs name the nodes. The distributed 
object system is where webhooks come in: the POSTs (etc) activate the 
services (and make requests) of the objects.

The messaging system, with notions of trust and 
transfer-of-responsibility, is a missing piece; here's where, for 
example, AMQP's operations could be transplanted into a webhooks-based 
distributed object system.

When it comes to reliability, it is a sad truth that no matter how much 
help you get from the network, it can only ever be the application's 
final responsibility to ensure that it is properly heard by a remote 
application.

HTTP is actually a very nice stab at a best effort blob delivery 
protocol: it provides positive acknowledgement of acceptance of a 
message, and being based around TCP shares all of the nice reliability 
properties of that.

Your example, of an urgent message that must be reliably delivered, 
requires the use of a persistent identifier for the message, with 
deduplication at the receiving application, and an appropriate 
retransmission strategy at the sender until positive acknowledgement is 
received from the receiver.

Regards,
   Tony

From tonyg@lshift.net  Sun Apr  5 12:02:42 2009
Return-Path: <tonyg@lshift.net>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A506B3A6A29 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 12:02:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.997
X-Spam-Level: 
X-Spam-Status: No, score=-1.997 tagged_above=-999 required=5 tests=[AWL=0.002,  BAYES_00=-2.599, J_CHICKENPOX_35=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5CEDlCkBczT4 for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 12:02:42 -0700 (PDT)
Received: from agentk.lshift.net (host226.lshift.net [195.172.218.226]) by core3.amsl.com (Postfix) with ESMTP id D092B3A6A14 for <hybi@ietf.org>; Sun,  5 Apr 2009 12:02:41 -0700 (PDT)
Received: from nodachi.lshift.net ([10.224.189.13] helo=walk.local) by agentk.lshift.net with esmtpsa (TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.63) (envelope-from <tonyg@lshift.net>) id 1LqXd8-0002eD-Se; Sun, 05 Apr 2009 20:03:43 +0100
Message-ID: <49D9008B.9060606@lshift.net>
Date: Sun, 05 Apr 2009 12:03:39 -0700
From: Tony Garnock-Jones <tonyg@lshift.net>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Paul Prescod <paul@prescod.net>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net>
In-Reply-To: <49D8FE80.2010100@lshift.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Score: -1.2
Cc: hybi@ietf.org, rabbitmq info <info@rabbitmq.com>
Subject: Re: [hybi] Trains, planes and web protocols
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 19:02:42 -0000

On the other hand, let me contradict myself:

Tony Garnock-Jones wrote:
> Your example, of an urgent message that must be reliably delivered, 
> requires the use of a persistent identifier for the message, with 
> deduplication at the receiving application, and an appropriate 
> retransmission strategy at the sender until positive acknowledgement is 
> received from the receiver.

We could absolutely build in the persistent identifier, which is 
necessary but not sufficient for exactly-once delivery. Something along 
the lines of Mark Nottingham's POE, 
http://www.mnot.net/drafts/draft-nottingham-http-poe-00.txt, perhaps. [1]

Or: I've been thinking about (ab)use of Content-Location in a symmetric 
(rHTTP-with-enrolment) HTTP network. Using Content-Location to 
essentially POST links to resources is one thing (e.g. twitter- or 
delicious-style sharing links: "Please go read this"), but it has a dual 
role of providing a URL for the POSTed resource. The HTTP spec 
indicates, but stops short of making it clear, that Content-Location 
could be used in an HTTP request message. It could be our persistent 
identifier.

Regards,
   Tony

[1]: In the POE draft (now long expired!) there's mention of the Link 
header, apropos which 
http://www.mnot.net/drafts/draft-nottingham-http-link-header-03.txt. 
Very cool, am wholly in favour, modulo a quibble or two about rel=first.

From fielding@gbiv.com  Sun Apr  5 12:34:49 2009
Return-Path: <fielding@gbiv.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D61053A685C for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 12:34:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.166
X-Spam-Level: 
X-Spam-Status: No, score=-5.166 tagged_above=-999 required=5 tests=[AWL=-2.567, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id B+YFizGAc7SG for <hybi@core3.amsl.com>; Sun,  5 Apr 2009 12:34:44 -0700 (PDT)
Received: from spaceymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 8B6D53A6869 for <hybi@ietf.org>; Sun,  5 Apr 2009 12:34:44 -0700 (PDT)
Received: from [192.168.0.132] (ip72-211-202-154.oc.oc.cox.net [72.211.202.154]) by spaceymail-a5.g.dreamhost.com (Postfix) with ESMTP id 7E5DA87058; Sun,  5 Apr 2009 12:35:48 -0700 (PDT)
In-Reply-To: <1cb725390904050921g196f4979rf2959d2786d06e2d@mail.gmail.com>
References: <49D1AE22.1080409@webtide.com> <Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com> <49D5AA5F.8030503@webtide.com> <Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com> <DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com> <Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com> <49D69573.8020508@stpeter.im> <49D6A6DB.9040701@webtide.com> <1cb725390904050921g196f4979rf2959d2786d06e2d@mail.gmail.com>
Mime-Version: 1.0 (Apple Message framework v753.1)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <C92C42C7-7696-4531-933B-05F7CCEBC5EE@gbiv.com>
Content-Transfer-Encoding: 7bit
From: "Roy T. Fielding" <fielding@gbiv.com>
Date: Sun, 5 Apr 2009 12:35:51 -0700
To: Paul Prescod <paul@prescod.net>
X-Mailer: Apple Mail (2.753.1)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 05 Apr 2009 19:34:49 -0000

On Apr 5, 2009, at 9:21 AM, Paul Prescod wrote:

> Like Roy, I think it's kind of counter-productive that everyone  
> comes to the list with their own set of requirements and goals.
>
> Do we agree that:
>
>  a) the list should have a wiki?
>
>  b) the requirements should be documented on the wiki?
>
> If so, does anyone have a preference for wiki platform? Choosing a  
> bikeshed colour is actually marginally better than going in circles.
>
> I'd start the conversation by suggesting Google Code, which  
> combines wiki with
>
>  * issue tracker (for unanswered questions)
>  * version control for more formal specs
>  * repository for diagrams or code samples

The IETF already has those things.  As I said before,

    http://trac.tools.ietf.org/bof/trac/wiki/HyBi

is good enough for now.  The rest can happen if a WG is justified.

....Roy


From ian@hixie.ch  Mon Apr  6 01:27:30 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 437913A6BE4 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 01:27:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.318
X-Spam-Level: 
X-Spam-Status: No, score=-2.318 tagged_above=-999 required=5 tests=[AWL=-1.234, BAYES_00=-2.599, J_CHICKENPOX_12=0.6, J_CHICKENPOX_45=0.6, SARE_MILLIONSOF=0.315]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id w-0Z0UCNiGAC for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 01:27:22 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id ECC143A6846 for <hybi@ietf.org>; Mon,  6 Apr 2009 01:27:18 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id EC913125B51; Mon,  6 Apr 2009 01:28:23 -0700 (PDT)
Date: Mon, 6 Apr 2009 08:28:23 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>, Peter Saint-Andre <stpeter@stpeter.im>, Greg Wilkins <gregw@webtide.com>, Sylvain Hellegouarch <sh@defuze.org>, Alex Russell <slightlyoff@google.com>, Tony Garnock-Jones <tonyg@lshift.net>
In-Reply-To: <49D9008B.9060606@lshift.net>
Message-ID: <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 08:27:30 -0000

What I'd like to work on is "TCP for the Web".

I think one of the key points is that when adding things to the Web
platform, IMHO we want to add basic building blocks, rather than
comprehensive protocols.

For example, we have HTML, not DocBook. We have a generic scripting 
language, not a domain-specific language. There are libraries like jQuery, 
Prototype, Y!UI, etc, which provide richer environments for authors. In 
recent times we've added generic SQL database support; despite the main 
use case being client-side e-mail storage, we didn't provide, say, local 
IMAP support. We added generic JavaScript background threads; we didn't 
add, say, a high-level database synchronisation mechanism.

The idea is to provide features that allow people to build libraries that 
provide complex features. This is similar to how the POSIX platform 
provides basic building blocks (libc) and then libraries like libssl, 
libz, libcurl, etc, provide the high-level functionality.

So for instance with bidirectional communication, I think what we should 
be aiming for is something on the level of a TCP socket (modulo security 
requirements), not something like XMPP or HTTP. We're not trying to solve 
a specific problem, we're trying to make it possible to solve a broad 
range of specific problems by providing a simple building block: a socket, 
the Web way.


Now, I have no objection to Jabber or other high-level protocols being 
made available to Web pages, if they can be shown to be securable, but 
that's not the kind of problem I'm trying to solve. WebSocket isn't 
intended to compete with that kind of feature.


With that in mind:

On Fri, 3 Apr 2009, Paul Prescod wrote:
> On Fri, Apr 3, 2009 at 2:07 PM, Ian Hickson <ian@hixie.ch> wrote:
> >
> > [IM is] _an_ application, but I don't think it's the biggest. I would 
> > expect games to be bigger, for example, and would expect background 
> > update notificiations in systems like Google Calendar to be even more 
> > common.
> 
> Most real-time social games will have chat embedded or on the page 
> around them.

Sure, but they'll all have subtly different needs. I imagine there will be 
libraries that provide such functionality for many sites, but for example 
while GMail will want to provide chat features with all the users in the 
greater federated Jabber network, a site like casualcollective will want 
to only have chat within the users of the site, and during a game, only 
the users in the game. A game like that from the Rainbow 6 series of games 
will want to distinguish between players who are still alive and players 
who are not; a game from the Quake series of games will not. Some will 
want some kinds of smileys, others might want audio and video. If we 
provide a protocol that is tuned to the purposes of one kind of chat, 
we'll fail to provide the features desired by another site, and sites will 
have to coerce the provided features into doing what they want.


> For example, how do I send a GIF over a Web Socket? HTTP and Jabber have 
> already solved that problem. How do I correlate a response to a message 
> with the original message?

If the solution HTTP has for sending a file is satisfactory, then 
presumably we don't need one for WebSocket. :-)

There are lots of solutions to the problem of sending files to a server, 
or sending messages in a variety of ways correlated with a request, like 
HTTP and Jabber as you mention, but also like FTP, IMAP, POP, BEEP, and a 
whole host of other high-level application protocols.

I'm interested in providing a feature that is the level below that, like 
TCP is to the above.


> > I'm not convinced that it's really a fair comparison, either. rHTTP as 
> > a protocol is remarkably complex, in that it includes by reference the 
> > whole of HTTP, which is a remarkably complex protocol.
> 
> I did say that the NEW specification is very short and simple.

It doesn't really matter how simple the "NEW" specification is, what 
matters is the complexity the author is exposed to.


> > Actually using rHTTP for true full-duplex connections would be much 
> > more complicated, since you'd need to somehow get two separate 
> > connections from the client to the server into the same CGI script.
> 
> I don't know what you mean.
> 
> These services cannot be served with CGI. CGI is built around 
> request-response. Just as PHP scripts do not in general implement HTTP, 
> I wouldn't expect the handlers of these things to implement the 
> protocol.

On the contrary, I expect the wide variety of uses of this feature to lead 
directly to a situation where people write custom implementations all the 
time. That's one the of the assumptions behind the Web Socket design. 
Interoperable heterogeneity is good, it reduces the impact of any single 
security bug, for instance.


> Here's how I expect it would work: there is a service optimized for 
> long-running TCP connections. Maybe it's written in C or Erlang. There 
> are PHP scripts (or whatever) for info coming in from web clients. When 
> PHP wants to tell the universe that "something happens" it puts a 
> message in a queue. Erlang or C delivers it.

That might work for some uses of bi-directional communication, but it 
wouldn't work for all of them. Luckily, this is the kind of thing that is 
easily built on top of the Web Socket protocol.


> Web developers do not want to become protocol handler developers and 
> they should not. The protocol should not be designed with the 
> presumption that they will be protocol developers. In fact: in my more 
> than 10 years of web development experience, I don't recall ever 
> implementing an HTTP client or server.

Certainly it would be reasonable for people to write servers that can be 
extended in that manner, but that doesn't mean we should preclude the 
possibility of people writing their own. Just like how with CGI scripts it 
is common for libraries to be used for interpreting the CGI environment 
variables, but people still sometimes write their own.


> What I have done frequently though, is trace the text flowing on the 
> wire, which is why a text-based protocol is preferable to me than a 
> binary one.

The Web Socket protocol is trivial to understand (0x00 indicates the start 
of a frame, 0xFF the end, and everything in between is ASCII/UTF-8) so I 
really don't think that it will make be hard to debug on the wire.


> > With Web Socket, there is only one underlying TCP connection.
> 
> So what? How does that make my app simpler? Are you expecting PHP 
> programmers to write long-running TCP server apps?

Yes.


> I don't understand why it matters whether it is one connection or two.

Having two connections is IMHO an misuse of the underlying protocol (TCP).


> > - It's not clear that Jabber's protocol supports the opt-in security
> >   model that is desired for an HTML API.
> 
> Not sure what that means.

We need to make sure that whatever we provide is consistent with the 
browser "same-origin" security model. It's not clear to me that Jabber is 
compatible with this model, since it assumes that the user agent is 
acting on behalf of the user.


> > Also, it's not clear to me that streaming XML is a particularly good 
> > solution. Experience with Web authors suggests that XML's draconian 
> > error handling and XML's namespaces are more confusing than 
> > desireable.
> 
> The web developer is just putting messages into a queue. It's the 
> protocol handler (like Apache, ejabberd, postfix) that talks the 
> protocol.
> 
> Can you point to a single protocol deployed on the Internet that is 
> commonly implemented and re-implemented and then re-re-implemented in 
> scripts, rather than being packaged in servers in libraries?

IRC. The multitude of multiplayer game protocols. The variety of protocols 
layered on top of <iframe>+<script>+XHR. The name/value pairs used in form 
data to write applications over HTML+forms.

The whole point here is to enable more of this for Web application 
developers. Right now they have no way to invent their own protocol at the 
TCP level.


On Fri, 3 Apr 2009, Paul Prescod wrote:
>
> By the way, you mentioned calendars: calDAV is the web protocol for 
> manipulating them. It might be a good fit for reverse HTTP.

We're not taking about manipulating calendars, we're talking about the UI 
side of an application talking to its logic backend. Other examples would 
be Google Docs getting a user's file list, or updating a document or 
spreadsheet as it is being edited. AtomPub and CalDAV wouldn't be 
appropriate, because they are generic inter-application protocols, not 
application-specific internal protocols tuned for the needs of those 
specific applications.


On Fri, 3 Apr 2009, Peter Saint-Andre wrote:
> > 
> >  - It's not clear that one can easily upgrade from HTTP to a 
> >    bidirectional Jabber connection.
> 
> Why upgrade from HTTP to XMPP?

Because in some environments HTTP is all we have, so we need a migration 
path from an HTTP connection to the two-way solution.


> Just build XMPP support into the browser, and you can refer to certain 
> resources as XMPP URIs/IRIs and other resources as HTTP URLs. Or the 
> client (browser) and server can negotiate an XMPP channel between 
> themselves for more advanced bidirectional functionality.

As noted earlier, I have no objection to this happening as well. But it 
dosn't solve all the needs. For example it wouldn't be a very efficient 
way of doing the train control I mentioned. It also would be a pretty poor 
transport for live editing updates as seen in Google Docs -- if the user 
hits one key, and so you need to send one byte to the server, a 200% 
overhead (as with Web Sockets) is already pretty extreme. Forcing the 
application author to frame each character in XML elements and namespaces 
and so forth would not be our proudest moment.


> >  - It's not clear that Jabber's protocol is designed to be resistant to 
> >    the attacks we expect (such as connecting to an SMTP server, or an 
> >    existing Jabber server, or whatnot).
> 
> Have you completed a threat analysis for the projected use cases? I 
> don't see anything about that in draft-hixie-thewebsocketprotocol-07 but 
> perhaps it is defined elsewhere.

I don't have anything written down, but the protocol was examined in this
context, yes. I would welcome an independent analysis though.


> > Experience with Web authors suggests that XML's draconian error 
> > handling and XML's namespaces are more confusing than desireable.
> 
> Whose experience? Where is this documented?

For draconican error handling, the disaster that is XML deployment on the 
Web is the primary documentation.

For namespaces, see: http://wiki.whatwg.org/wiki/Namespace_confusion


> The XMPP developer community has produced a large volume of software 
> (see links above) and runs a large network of servers with millions of 
> IM users. So far we have not experienced any significant interop 
> problems.

Sure, but you're all competent. We're talking about letting random authors 
who are barely able to write CGI scripts create new interactive 
applications here. More than 95% of HTML pages are invalid in some way. 
Simplicity is key here.


> > - The protocol should support running over ports 80 and 443,
> >   ideally with the ability to share the port on the server
> >   with an HTTP server.
> 
> Why? Using a different port for a different protocol seems like the 
> proper approach.

Because in practice, there are environments that block all but port 80, 
and applications like GMail won't use solutions that don't work in those 
environments too.

(Note that Web Sockets default to another port; the requirement is only 
that ports 80 and 443 be supported, not required or the default.)


On Sat, 4 Apr 2009, Greg Wilkins wrote:
> 
> If the IETF does come out with a bidirectional HTTP recommendation, it 
> will be a huge once in a decade opportunity to upgrade the protocols 
> supported by the network infrastructure.

Why would it be a once in a decade opportunity? If there is a need, then 
it can be filled. It's not like we can only create new protocols when the 
planets align in some particular way.


> Websockets sentinel UTF-8 datagrams may well be ideally suited for a 
> specific type of application, but can we really say that this will be 
> the only content type that we need to transport asynchronously server to 
> client for the next decade or two?

The protocol is generic enough (it can do text and known-length binary out 
of the box, and is easily extensible to handle multiplexed infinite 
streams, structured data, finite unknown length binary data via chunking, 
etc), and simple enough (it has virtually no features, leaving everything 
up to the application) to be usable for most bi-directional use cases for 
Web applications.


> What if UTF-16 becomes the preferred encoding? or gzipped UTF-16 or some 
> as-of-yet-un-invented compression algorithm specially designed for 
> packets of UTF-16 data holding JSON becomes in common usage.

We can use the known-length encoding mechanism to transport all of those, 
so they would be easy extensions.


> I'm already working with clients that wish to push images.

We can use the known-length encoding mechanism support by the Web Socket 
protocol as designed today to handle any kind of binary data.


> If web socket supports arbitrary binary data, then why is there a need 
> for the sentinel mechanism and special status for UTF-8 data?

There's a need for a separate "this is text" frame than "this is arbitrary 
data" frame because at the API level the only datatype today is text. 
(This is why the binary mode frames are currently discarded -- JS doesn't 
yet support representing the binary data.)

Once you're sending binary data, you want to avoid length markers because 
authors are very much more likely to get measuring UTF-8 data wrong 
compared to just stick an 0xFF byte on the end.


> Note also with websocket binary transfer, I believe it is a limitation 
> that there cannot be meta data (headers) associated with a datagram. 
> Thus information like mime types, char-sets content-encodings and 
> transfer encodings will need to be handled with stateful mechanisms 
> based on assumption, connection and/or previous sent datagrams.

There can be as much metadata as the applicaton layer wants -- just put 
it before the data, using an application-specific format.

This is just like with TCP, there's no room for MIME types. Protocols that 
use TCP, like HTTP, layer MIME types on top. This isn't a failing of TCP.


On Fri, 3 Apr 2009, Alex Russell wrote:
> >
> > I'm not convinced that it's really a fair comparison, either. rHTTP as 
> > a protocol is remarkably complex, in that it includes by reference the 
> > whole of HTTP, which is a remarkably complex protocol.
> 
> That sort of depends on which version you're talking about. HTTP 1.0 is 
> relatively simple, whereas 1.1 is hugely complicated, particularly when 
> it comes to accommodating chunked encoding and pipelining.

HTTP 1.0 is relatively simple compared to 1.1, but it has all kinds of 
weird subtlties, like header continuations, the many verbs (GET POST 
etc) and respone codes (200, 404, etc), that I wouldn't be at all 
surprised to see misimplemented by random authors. Getting a complete 
conforming implementation is non-trivial, even for 1.0.


> If you're looking to have the complexity of 1.0 as a high-water-mark, 
> that would seem perfectly reasonable. People are going to mis-implement 
> whatever you spec anyway.

Granted, but it's actually pretty hard to misimplement WebSockets as 
specced now and still have it work in the common case. (Though I'm sure 
creative Web designers will prove me wrong.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From ian@hixie.ch  Mon Apr  6 13:21:13 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0643928C0DD for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 13:21:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.022
X-Spam-Level: 
X-Spam-Status: No, score=-3.022 tagged_above=-999 required=5 tests=[AWL=-0.423, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GLTI3YV06tE8 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 13:21:06 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 1A98A3A6D21 for <hybi@ietf.org>; Mon,  6 Apr 2009 13:20:34 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 05ED9125C36; Mon,  6 Apr 2009 13:21:41 -0700 (PDT)
Date: Mon, 6 Apr 2009 20:21:40 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Alex Russell <slightlyoff@google.com>
In-Reply-To: <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>  <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>  <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com> <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 20:21:13 -0000

On Mon, 6 Apr 2009, Alex Russell wrote:
> On Mon, Apr 6, 2009 at 1:28 AM, Ian Hickson <ian@hixie.ch> wrote:
> >
> > What I'd like to work on is "TCP for the Web".
> 
> How about TCP, then?

Unfortunately raw TCP has a number of disadvantages:

 - It doesn't have a good story for preventing cross-origin attacks, 
   e.g. where a hostile page contacts an intranet server. (Web Socket 
   enforces cross-origin integrity by requiring the server to acknowledge 
   the origin of the initiating page.)

 - It doesn't have a good story for upgrading from HTTP, which is one of 
   our requirements.

 - It doesn't play well with virtual hosting, which (in combination with 
   the previous two points) make it hard to use securely in shared 
   hosting environments.

 - It doesn't have protection against connecting to existing services like 
   SMTP servers, HTTP servers, IRC servers, etc.

Note that the SVG working group originally considered using raw TCP, and 
the browser vendors pointed to a number of security problems with this 
which resulted in the feature being dropped.


> I'd also like some form of content-neutral, bi-directional byte 
> packaging format that will let me encapsulate message sending at a 
> higher level. Being able to send binary data seems equally important for 
> the long-term applications (think feeding a <video> element streaming 
> content via a websocket).

Web Socket as designed today supports framing binary data.


> I have a hard time with this line of argumentation. The web has nearly 
> always taken flexibility away in return for ease of implementation. 
> Setting limits and optimizing for the 80% case is *good*, and the web is 
> proof.

I agree, though apparently in this case we reach different conclusions 
based on this principle.


> Not having some sort of metadata at the front of a frame really does 
> seem like a serious oversight. If you want to be TCP for the web, at 
> least copy the parts that worked well = )

Could you elaborate on this? What are the features of TCP that we should 
duplicate at the WebSocket layer? (Isn't it enough that they be available 
at the TCP layer?) I don't really understand. What API are you proposing 
be made available?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From ian@hixie.ch  Mon Apr  6 16:19:22 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 063F83A69F6 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:19:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.148
X-Spam-Level: 
X-Spam-Status: No, score=-3.148 tagged_above=-999 required=5 tests=[AWL=-0.549, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yF7cEG2FOWmx for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:19:21 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 1406F3A6452 for <hybi@ietf.org>; Mon,  6 Apr 2009 16:19:20 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id BF526125B51; Mon,  6 Apr 2009 16:20:26 -0700 (PDT)
Date: Mon, 6 Apr 2009 23:20:26 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Alex Russell <slightlyoff@google.com>
In-Reply-To: <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>  <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>  <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>  <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>  <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 23:19:22 -0000

On Mon, 6 Apr 2009, Alex Russell wrote:
> >
> > Web Socket as designed today supports framing binary data.
> 
> ...and will require senders and receivers to encode/decode (at least) 
> null bytes with some alternate representation that isn't spec'd.

No, not at all. Binary data is sent with a known length declaration, 
there's no reason to escape anything.


> > Could you elaborate on this? What are the features of TCP that we 
> > should duplicate at the WebSocket layer? (Isn't it enough that they be 
> > available at the TCP layer?) I don't really understand. What API are 
> > you proposing be made available?
> 
> Some sort of header that describes a message seems essential. We made 
> the mistake in Bayeux of not including one and I regret it.

I don't understand. Could you elaborate on this?

What would it look like at the API level?

Right now we have:

   socket.postMessage("move:1,2");

Are you saying it should instead be?:

   socket.postMessage("move", "1,2");

...or some such?

I think in the middle-term we will probably want to extend the API to 
support JSON (or some other encoding, e.g. Protocol Buffers) as a message 
type, would that resolve this issue?


> Greg has argued for a content-length descriptor and I can see wanting 
> other kinds of application-level routing information to be made 
> available in a message so that parsing it isn't required to handle/drop 
> it.

I don't understand. Why would a second length field be useful?


> HTTP did a great job here by allowing an arbitrary number of headers for 
> payloads, and we're still reaping the benefits. At least leaving room 
> for a standard-format message header seems useful.

The payload is completely open-ended, the idea is that applications can 
put whatever format they want in there. If they want HTTP-like headers per 
message, then that's quite possible, e.g.:

   socket.postMessage("Action: Move" + crlf +
                      "Player: + player + crlf +
                      crlf +
                      "1,2");

...or whatever. The idea is to leave this up to the author, just like TCP 
leaves the payload completely open to the next layer.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Mon Apr  6 16:24:39 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 098093A6452 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:24:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.579
X-Spam-Level: 
X-Spam-Status: No, score=-2.579 tagged_above=-999 required=5 tests=[AWL=0.020,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ggKW9UpvnP5G for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:24:38 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.189]) by core3.amsl.com (Postfix) with ESMTP id 17C503A69F6 for <hybi@ietf.org>; Mon,  6 Apr 2009 16:24:37 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so2227388tid.25 for <hybi@ietf.org>; Mon, 06 Apr 2009 16:25:43 -0700 (PDT)
Received: by 10.110.46.3 with SMTP id t3mr7154095tit.45.1239060343438; Mon, 06 Apr 2009 16:25:43 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id b7sm1233127tic.15.2009.04.06.16.25.39 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 06 Apr 2009 16:25:42 -0700 (PDT)
Message-ID: <49DA8F6E.7010408@webtide.com>
Date: Tue, 07 Apr 2009 09:25:34 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com> <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org, Alex Russell <slightlyoff@google.com>
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 23:24:39 -0000

Ian Hickson wrote:
>  - It doesn't have a good story for upgrading from HTTP, which is one of 
>    our requirements.

I think upgrade from HTTP is key to these discussions.
If HTTP is our starting point, then upgrade is the natural path.
So I see two key questions:

1) Must we upgrade? are there any recommendations or simple
extension for HTTP that will make it better suited (eg 1xx
responses with content)?

2) If we upgrade what should we upgrade to: rHTTP, websocket,
Beep, XMPP?

In the second case, perhaps we should not try to pick a winner.
Instead we could be looking at what infrastructural and
security issues exist to prevent upgrade from HTTP to an
arbitrary connection based protocol between any
pair of consenting HTTP client/server.

My understanding of the situation today is that upgrade
frequently works because proxy servers often act as simple
TCP tunnels after checking the headers of the first
request/response.  But there are also proxies that will
prevent an upgrade as they wish to inspect/handler every
request.  Thus horrid hacks such as using CONNECT are
contemplated.

It would be great if we could come up with recommendations
for both questions.  How to make HTTP work better bidirectionally
for the current long-pollers of the world, and how to make
upgrade work more generally so that the alternative protocols
can be deployed and compete for who's best.


cheers





From Martin.Thomson@andrew.com  Mon Apr  6 16:33:09 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3B2183A6D0C for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:33:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.529
X-Spam-Level: 
X-Spam-Status: No, score=-2.529 tagged_above=-999 required=5 tests=[AWL=0.070,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GBvHGVvyB2nm for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:33:08 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 402F63A68A0 for <hybi@ietf.org>; Mon,  6 Apr 2009 16:33:08 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_06_18_54_34
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from aopexbh2.andrew.com [10.86.20.25] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Mon, 06 Apr 2009 18:54:34 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by aopexbh2.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Mon, 6 Apr 2009 18:34:13 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Mon, 6 Apr 2009 18:34:12 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF1059C1000@AHQEX1.andrew.com>
In-Reply-To: <49DA8F6E.7010408@webtide.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
Thread-Index: Acm3DwQOSN9q7MrcQTWHA1wdjEGlcwAAJ5eg
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com><49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net><Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com><Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <49DA8F6E.7010408@webtide.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Greg Wilkins" <gregw@webtide.com>, "Ian Hickson" <ian@hixie.ch>
X-OriginalArrivalTime: 06 Apr 2009 23:34:13.0489 (UTC) FILETIME=[31846610:01C9B710]
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 23:33:09 -0000
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From gregw@webtide.com  Mon Apr  6 16:57:17 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BD6D928C178 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:57:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.582
X-Spam-Level: 
X-Spam-Status: No, score=-2.582 tagged_above=-999 required=5 tests=[AWL=0.018,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 86NmbfzhFHZv for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 16:57:16 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.190]) by core3.amsl.com (Postfix) with ESMTP id AE3DA28C17A for <hybi@ietf.org>; Mon,  6 Apr 2009 16:57:14 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so2233712tid.25 for <hybi@ietf.org>; Mon, 06 Apr 2009 16:58:20 -0700 (PDT)
Received: by 10.110.10.16 with SMTP id 16mr7194744tij.27.1239062300122; Mon, 06 Apr 2009 16:58:20 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id u12sm912479tia.6.2009.04.06.16.58.16 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 06 Apr 2009 16:58:18 -0700 (PDT)
Message-ID: <49DA9713.5050407@webtide.com>
Date: Tue, 07 Apr 2009 09:58:11 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com> <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com> <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>	<6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org, Alex Russell <slightlyoff@google.com>
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Apr 2009 23:57:17 -0000

Ian Hickson wrote:

>> HTTP did a great job here by allowing an arbitrary number of headers for 
>> payloads, and we're still reaping the benefits. At least leaving room 
>> for a standard-format message header seems useful.
> 
> The payload is completely open-ended, the idea is that applications can 
> put whatever format they want in there. If they want HTTP-like headers per 
> message, then that's quite possible, e.g.:
> 
>    socket.postMessage("Action: Move" + crlf +
>                       "Player: + player + crlf +
>                       crlf +
>                       "1,2");
> 
> ...or whatever. The idea is to leave this up to the author, just like TCP 
> leaves the payload completely open to the next layer.


Ian,

TCP does not leave it up to the author.   TCP has HTTP, FTP, IMAP,
SMTP, etc. as standards that define content and handling.
TCP has headers and ports to help determine which protocol is in use.

I don't believe Websocket is well suited to become the basis
for other protocols.   It lacks the meta data support and the
UTF-8 favouritism is an issue.    EG I could not use the approach
you advocate above to send a gzipped utf-16 encoded body!

Something like BEEP could indeed be used as a basis for other
protocols, so we should ensure that we can upgrade to both websocket
and protocols like BEEP.


cheers












From jonas@sicking.cc  Mon Apr  6 17:23:15 2009
Return-Path: <jonas@sicking.cc>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 65BE33A6452 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:23:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hZVczJ7VLUKz for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:23:14 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.30]) by core3.amsl.com (Postfix) with ESMTP id 9F6693A693E for <hybi@ietf.org>; Mon,  6 Apr 2009 17:23:14 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 5so3326504ywh.49 for <hybi@ietf.org>; Mon, 06 Apr 2009 17:24:20 -0700 (PDT)
MIME-Version: 1.0
In-Reply-To: <49DA9713.5050407@webtide.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>  <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>  <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>  <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>  <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>  <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>  <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>  <49DA9713.5050407@webtide.com>
Date: Mon, 6 Apr 2009 17:24:05 -0700
Received: by 10.150.53.2 with SMTP id b2mr9339143yba.237.1239063860422; Mon,  06 Apr 2009 17:24:20 -0700 (PDT)
Message-ID: <63df84f0904061724s34d0dca1wb1789d0e4c4a4095@mail.gmail.com>
From: Jonas Sicking <jonas@sicking.cc>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: Alex Russell <slightlyoff@google.com>, hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 00:23:15 -0000

On Mon, Apr 6, 2009 at 4:58 PM, Greg Wilkins <gregw@webtide.com> wrote:
> Ian Hickson wrote:
>
>>> HTTP did a great job here by allowing an arbitrary number of headers fo=
r
>>> payloads, and we're still reaping the benefits. At least leaving room f=
or a
>>> standard-format message header seems useful.
>>
>> The payload is completely open-ended, the idea is that applications can
>> put whatever format they want in there. If they want HTTP-like headers p=
er
>> message, then that's quite possible, e.g.:
>>
>> =A0 socket.postMessage("Action: Move" + crlf +
>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0"Player: + player + crlf +
>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0crlf +
>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0"1,2");
>>
>> ...or whatever. The idea is to leave this up to the author, just like TC=
P
>> leaves the payload completely open to the next layer.
>
>
> Ian,
>
> TCP does not leave it up to the author. =A0 TCP has HTTP, FTP, IMAP,
> SMTP, etc. as standards that define content and handling.
> TCP has headers and ports to help determine which protocol is in use.
>
> I don't believe Websocket is well suited to become the basis
> for other protocols. =A0 It lacks the meta data support and the
> UTF-8 favouritism is an issue.

This is an interesting issue indeed. Maybe a tiny amount of metadata
information could be added without adding a significant amount of
overhead.

/ Jonas

From ian@hixie.ch  Mon Apr  6 17:26:49 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5F95D3A682E for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:26:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.827
X-Spam-Level: 
X-Spam-Status: No, score=-2.827 tagged_above=-999 required=5 tests=[AWL=-0.828, BAYES_00=-2.599, J_CHICKENPOX_73=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7q4LJb9CcTGu for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:26:48 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 9AD233A6CBD for <hybi@ietf.org>; Mon,  6 Apr 2009 17:26:48 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 243CC121F17; Mon,  6 Apr 2009 17:27:54 -0700 (PDT)
Date: Tue, 7 Apr 2009 00:27:54 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49DA9713.5050407@webtide.com>
Message-ID: <Pine.LNX.4.62.0904070020220.25058@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>  <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>  <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>  <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>  <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org, Alex Russell <slightlyoff@google.com>
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 00:26:49 -0000

On Tue, 7 Apr 2009, Greg Wilkins wrote:
> 
> TCP does not leave it up to the author.  TCP has HTTP, FTP, IMAP, SMTP, 
> etc. as standards that define content and handling. TCP has headers and 
> ports to help determine which protocol is in use.

Sure. WebSocket is URI-addressable, which is intended to allow servers to 
distinguish between different services in this way. This leverages the 
ideas behind HTTP's service addressing mechanisms.


> I don't believe Websocket is well suited to become the basis for other 
> protocols.  It lacks the meta data support and the UTF-8 favouritism is 
> an issue.  EG I could not use the approach you advocate above to send a 
> gzipped utf-16 encoded body!

There's no way from JavaScript to get a hold of a gzipped UTF-16 encoded 
blob right now at all. This is why the API is limited to text, and why 
there is UTF-8 favouritism in the protocol. However, once we have binary 
support in JS, I see no reason why we couldn't extend the API to support 
things like this:

   socket.postBinaryMessage(textToUTF8("Filename: " + name + crlf +
                                       "Format: UTF-16" + crlf +
                                       crlf) + 
                            gzip(textToUTF16(data)));

...where textToUTF8(), textToUTF16(), and gzip() are helper methods that 
handle binary data.

The protocol itself already supports binary transport.


> Something like BEEP could indeed be used as a basis for other protocols, 
> so we should ensure that we can upgrade to both websocket and protocols 
> like BEEP.

I agree. WebSocket is not intended to prevent the user of other protocols. 
Right now, HTTP supports Upgrade:ing to any protocol that defines a way to 
handle an HTTP Upgrade, as far as I can tell, so I don't think there is a 
problem here.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From stpeter@stpeter.im  Mon Apr  6 17:27:51 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DB2C23A6CBD for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:27:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.33
X-Spam-Level: 
X-Spam-Status: No, score=-2.33 tagged_above=-999 required=5 tests=[AWL=0.269,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KlbYUkcu3e2D for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:27:44 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 56E313A682E for <hybi@ietf.org>; Mon,  6 Apr 2009 17:27:44 -0700 (PDT)
Received: from squire.local (dsl-251-125.dynamic-dsl.frii.net [216.17.251.125]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id 8DDB8E800D; Mon,  6 Apr 2009 18:28:49 -0600 (MDT)
Message-ID: <49DA9E40.4060002@stpeter.im>
Date: Mon, 06 Apr 2009 18:28:48 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <49DA8F6E.7010408@webtide.com>
In-Reply-To: <49DA8F6E.7010408@webtide.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms090107030201080700010104"
Cc: Alex Russell <slightlyoff@google.com>, hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 00:27:51 -0000

This is a cryptographically signed message in MIME format.

--------------ms090107030201080700010104
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/6/09 5:25 PM, Greg Wilkins wrote:
> Ian Hickson wrote:
>>  - It doesn't have a good story for upgrading from HTTP, which is one
>> of    our requirements.
> 
> I think upgrade from HTTP is key to these discussions.
> If HTTP is our starting point, then upgrade is the natural path.

Perhaps it would help to define precisely what we mean by upgrading from
HTTP. Does this mean that an HTTP client is communicating with an HTTP
server (for some kind of functionality), and the two parties agree to
switch to a different protocol (for some other kind of functionality)?

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms090107030201080700010104
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms090107030201080700010104--

From stpeter@stpeter.im  Mon Apr  6 17:52:39 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B184728C132 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:52:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.352
X-Spam-Level: 
X-Spam-Status: No, score=-2.352 tagged_above=-999 required=5 tests=[AWL=0.247,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ceGIUE3aKNsP for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 17:52:39 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 0C5023A68A0 for <hybi@ietf.org>; Mon,  6 Apr 2009 17:52:39 -0700 (PDT)
Received: from squire.local (dsl-251-125.dynamic-dsl.frii.net [216.17.251.125]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id 29E9DE800D; Mon,  6 Apr 2009 18:53:44 -0600 (MDT)
Message-ID: <49DAA412.2070505@stpeter.im>
Date: Mon, 06 Apr 2009 18:53:38 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms080705060407040802030906"
Cc: hybi@ietf.org, Alex Russell <slightlyoff@google.com>
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 00:52:39 -0000

This is a cryptographically signed message in MIME format.

--------------ms080705060407040802030906
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/6/09 2:28 AM, Ian Hickson wrote:
> What I'd like to work on is "TCP for the Web".

That's what we tried to define with BOSH. Whether we succeeded is an
open question.

/psa

--------------ms080705060407040802030906
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms080705060407040802030906--

From gregw@webtide.com  Mon Apr  6 18:32:39 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 269733A68CF for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 18:32:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.583
X-Spam-Level: 
X-Spam-Status: No, score=-2.583 tagged_above=-999 required=5 tests=[AWL=0.016,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id VcBN3mIbWYyA for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 18:32:32 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.190]) by core3.amsl.com (Postfix) with ESMTP id 0735B28C0E3 for <hybi@ietf.org>; Mon,  6 Apr 2009 18:32:31 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so2255590tid.25 for <hybi@ietf.org>; Mon, 06 Apr 2009 18:33:37 -0700 (PDT)
Received: by 10.110.93.8 with SMTP id q8mr7323146tib.21.1239068017478; Mon, 06 Apr 2009 18:33:37 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm1221293tib.8.2009.04.06.18.33.31 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 06 Apr 2009 18:33:35 -0700 (PDT)
Message-ID: <49DAAD5B.6090303@webtide.com>
Date: Tue, 07 Apr 2009 11:33:15 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: hybi@ietf.org
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <49DA8F6E.7010408@webtide.com> <49DA9E40.4060002@stpeter.im>
In-Reply-To: <49DA9E40.4060002@stpeter.im>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 01:32:39 -0000

Peter Saint-Andre wrote:
> On 4/6/09 5:25 PM, Greg Wilkins wrote:
>> Ian Hickson wrote:
>>>  - It doesn't have a good story for upgrading from HTTP, which is one
>>> of    our requirements.
>> I think upgrade from HTTP is key to these discussions.
>> If HTTP is our starting point, then upgrade is the natural path.
> 
> Perhaps it would help to define precisely what we mean by upgrading from
> HTTP. Does this mean that an HTTP client is communicating with an HTTP
> server (for some kind of functionality), and the two parties agree to
> switch to a different protocol (for some other kind of functionality)?


The upgrade mechanism of HTTP/1.1 is a hop-by-hop mechanism,
so it a user-agent, a server and all the intermediary gateways
and proxies support a specific upgrade, then the protocol
used by a connection can be switched to almost any
connection oriented protocol.

The problem is that all proxies and gateways need to
support the upgraded protocol.  So even if we were able to
make all browsers and servers support websocket, this would
not be sufficient as HTTP/1.1 compliant proxy should not forward
upgrade headers!

(Note that some proxies and gateways do work, as they
ignore upgrade headers and after the first HTTP request/response
some act as just TCP/IP tunnels).


So for non-HTTP upgrades to be widely usable, the network
infrastructure will need to be upgraded to support either
specific protocols (eg websocket) or generic bidirectional
protocols.


I'd prefer the second option (not picking a winner) and
it may be as simple as producing a recommendation
of how a proxy/gateway should handle an upgrade request
for an unknown protocol.   For example, perhaps we could
recommended that unknown upgrades are handled by proxy's and
gateways by them becoming non-caching, non-buffering, TCP/IP
relays.  Ie all packets received are sent without caching,
aggregation, alternation etc.  (like CONNECT).

This legitimises the gateways and proxies that already
work and will encourage more to work over time.  It would
reduce (eventually remove) the need for misusing the CONNECT
mechanism (which has security issues as it allows arbitrary
IP's and ports).

If we can get the proxies and gateways of the world to
support such arbitrary upgrades, then websocket, rHTTP,
XMPP, BEEP should all be able to be easily supported
by consenting pairs of clients and servers.



cheers




















From duerst@it.aoyama.ac.jp  Mon Apr  6 19:38:32 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4166D3A6947 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 19:38:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.389
X-Spam-Level: 
X-Spam-Status: No, score=-0.389 tagged_above=-999 required=5 tests=[AWL=-0.599, BAYES_00=-2.599, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fmUpjRJfNlf8 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 19:38:23 -0700 (PDT)
Received: from acmailgw1.acop.aoyama.ac.jp (acmailgw1.acop.aoyama.ac.jp [133.2.9.176]) by core3.amsl.com (Postfix) with ESMTP id 588E03A6968 for <hybi@ietf.org>; Mon,  6 Apr 2009 19:38:20 -0700 (PDT)
Received: from scmse3.scbb.aoyama.ac.jp ([133.2.253.23]) by acmailgw1.acop.aoyama.ac.jp (secret/secret) with SMTP id n372dOMq018172 for <hybi@ietf.org>; Tue, 7 Apr 2009 11:39:25 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse3.scbb.aoyama.ac.jp with smtp id 295d_4e8c89da_231d_11de_a4a4_001d0969ab06; Tue, 07 Apr 2009 11:39:24 +0900
Received: from [IPv6:::1] ([133.2.210.1]:39007) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC6B705> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Tue, 7 Apr 2009 11:38:27 +0900
Message-ID: <49DABCC5.5060809@it.aoyama.ac.jp>
Date: Tue, 07 Apr 2009 11:39:01 +0900
From: =?ISO-8859-1?Q?=22Martin_J=2E_D=FCrst=22?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>	<6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>	<Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com>
In-Reply-To: <49DA9713.5050407@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Cc: Alex Russell <slightlyoff@google.com>, hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 02:38:32 -0000

> Ian Hickson wrote:

>> ...or whatever. The idea is to leave this up to the author, just like
>> TCP leaves the payload completely open to the next layer.

It seems indeed that that's what's currently happening in many
long-polling applications. I'm not sure whether that's a good
thing or not.

I think one difference between things such as HTTP/SMTP/... and
long polling is that for the former, the main scenario is that
totally unrelated servers and clients have to interoperate.
That's why there is a lot of metadata specified and transferred
in the protocol.

For long polling, it's currently mostly that the server code
(Perl, PHP, Python, Ruby,...) and the client code
(Javascript) are written by the same person or group.
That's why less metadata may be needed, and when needed
(e.g. a timestamp), it's sent ad-hoc.

But once we have a well-accepted protocol for long polling,
I'm sure there will be more ideas and needs to communicate
between different, independently written client and servers,
and the value of agreed-upon conventions for metadata and
such will increase.

Also, please note that in it's first version (later called
HTTP 0.9), HTTP was in many ways about as simple as the
current WebSockets protocol. But that proved to be too
simple.

On 2009/04/07 8:58, Greg Wilkins wrote:

> TCP does not leave it up to the author. TCP has HTTP, FTP, IMAP,
> SMTP, etc. as standards that define content and handling.
> TCP has headers and ports to help determine which protocol is in use.
>
> I don't believe Websocket is well suited to become the basis
> for other protocols.

It would definitely be a good idea to think about how it can be
extended on various levels. But maybe Ian already has ideas,
he just don't want to tell us because otherwise, everybody
will start extending willy-nilly.

It would also be a good idea to think about what kinds of protocols
one would like to run on top of a long-poll connection.

> It lacks the meta data support and the
> UTF-8 favouritism is an issue.

I disagree with "UTF-8 favouritism". 99.9% of programmers
want to get text across when they want to get text across,
they couldn't care less about encoding. Fixing the encoding
in the general case to something that is well known to work
very widely is a very good thing.

Most IETF protocols until a little while ago came with a strong
"ASCII favoritism". The first version(s) of the Web came with
a strong "ISO-8859-1 favoritism". Both are problematic because
they exclude large parts of the world. But both were extremely
helpful in getting interoperability and ease of implementation
for programmers who don't want to deal with character encoding
details. With UTF-8, we get both the benefit of a single encoding
and coverage of the world.

As for UTF-8 being inefficient in some cases, it indeed takes
about 1.5 times the bandwidth for purely East Asian text, about
2 times the bandwidth for Cyrillic, Greek, Arabic, Hebrew, and
about 3 times the bandwidth for Indic scripts and scripts such
as Georgian (and about 4 times the bandwidth for cases such as
Linear B, but I hope not too many are worried about that :-).
But with every bit of framing, XML, JSON, or whatever, which
mostly will be in ASCII, these differences get smaller and
smaller.

Regards,    Martin.

-- 
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From ian@hixie.ch  Mon Apr  6 19:46:57 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C4ECF3A6947 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 19:46:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.946
X-Spam-Level: 
X-Spam-Status: No, score=-2.946 tagged_above=-999 required=5 tests=[AWL=-0.647, BAYES_00=-2.599, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UkTYh7u837Oi for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 19:46:56 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id E41543A687E for <hybi@ietf.org>; Mon,  6 Apr 2009 19:46:56 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id A3FBD125C36; Mon,  6 Apr 2009 19:48:02 -0700 (PDT)
Date: Tue, 7 Apr 2009 02:48:02 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: =?ISO-8859-1?Q?=22Martin_J=2E_D=FCrst=22?= <duerst@it.aoyama.ac.jp>
In-Reply-To: <49DABCC5.5060809@it.aoyama.ac.jp>
Message-ID: <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com> <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com> <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="1909464018-507856712-1239072482=:25058"
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 02:46:57 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--1909464018-507856712-1239072482=:25058
Content-Type: TEXT/PLAIN; charset=iso-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Tue, 7 Apr 2009, "Martin J. D=FCrst" wrote:
>=20
> Also, please note that in it's first version (later called HTTP 0.9),=20
> HTTP was in many ways about as simple as the current WebSockets=20
> protocol. But that proved to be too simple.

The WebSocket protocol is extensible by design. It supports two kinds of=20
frames (0xFF-sentinel-delimited and predeclared-length), and there can be=
=20
up to 128 subtypes of each kind. The current version only exposes one of=20
these, with a payload defined to be UTF-8.


> On 2009/04/07 8:58, Greg Wilkins wrote:
> >=20
> > I don't believe Websocket is well suited to become the basis for other=
=20
> > protocols.
>=20
> It would definitely be a good idea to think about how it can be extended=
=20
> on various levels. But maybe Ian already has ideas, he just don't want=20
> to tell us because otherwise, everybody will start extending=20
> willy-nilly.

I've already listed a variety of probable extensions:

 - simple binary data (known length arbitrary payloads)

 - chunked binary data (unknown length arbitrary payloads)

 - structured data (e.g. JSON)

 - multiplexed video

=2E..etc. The protocol is designed such that these new frame types can be=
=20
introduced later without breaking compatibility with deployed=20
implementations: UAs are required to skip unknown frame types.

--=20
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
--1909464018-507856712-1239072482=:25058--

From gregw@webtide.com  Mon Apr  6 21:02:44 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E4CF53A6D5F for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:02:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.939
X-Spam-Level: 
X-Spam-Status: No, score=-1.939 tagged_above=-999 required=5 tests=[AWL=-0.632, BAYES_00=-2.599, MISSING_HEADERS=1.292]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id vYUIIaOtxpSX for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:02:39 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id 466BF3A6928 for <hybi@ietf.org>; Mon,  6 Apr 2009 21:02:38 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so2299415tid.25 for <hybi@ietf.org>; Mon, 06 Apr 2009 21:03:44 -0700 (PDT)
Received: by 10.110.49.6 with SMTP id w6mr7461765tiw.55.1239077024476; Mon, 06 Apr 2009 21:03:44 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 14sm243808tim.9.2009.04.06.21.03.41 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 06 Apr 2009 21:03:43 -0700 (PDT)
Message-ID: <49DAD098.8060703@webtide.com>
Date: Tue, 07 Apr 2009 14:03:36 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
CC: hybi@ietf.org
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>	<6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>	<Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>	<49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp> <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 04:02:45 -0000

Ian Hickson wrote:
> On Tue, 7 Apr 2009, "Martin J. Dürst" wrote:
>> Also, please note that in it's first version (later called HTTP 0.9), 
>> HTTP was in many ways about as simple as the current WebSockets 
>> protocol. But that proved to be too simple.
> 
> The WebSocket protocol is extensible by design. It supports two kinds of 
> frames (0xFF-sentinel-delimited and predeclared-length), and there can be 
> up to 128 subtypes of each kind. The current version only exposes one of 
> these, with a payload defined to be UTF-8.

Ian,

this is extensibility by the protocol specifiers, so that later versions
of the protocol can introduce new packets without breaking backwards
compatibility.

But HTTP is user extensible.     So if I invent a new content
type, encoding or compression algorithm, I don't need to wait for
HTTP/1.2 to use that.    I don't need to apply to some standards
body to allocate an identifier from a scarce (128) space.

Defining subtypes for  compress-utf8-data,  gzip-utf8-data,
gzip-utf8-data will use up a 128 space very quickly.

Without such subtypes, then you'd have to use one of the
untype binary data types and rely on assumptions of
compression type or stateful exchange of meta data in
prior datagrams.


I'm sorry if I appear to be bashing websocket.  That really is not my
intention (and he who lives in glass houses should not throw
stones).     I'm just saying any general purpose transport needs
to have the flexibility and extensibility of HTTP.

cheers

PS. and I'm also not saying that Ian is proposing websocket
as a general purpose transport.










From stpeter@stpeter.im  Mon Apr  6 21:18:17 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 415E53A6D5E for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:18:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.442
X-Spam-Level: 
X-Spam-Status: No, score=-2.442 tagged_above=-999 required=5 tests=[AWL=0.158,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZflsVgM4vJSE for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:18:09 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 353343A6D65 for <hybi@ietf.org>; Mon,  6 Apr 2009 21:18:09 -0700 (PDT)
Received: from squire.local (ip-65-183-77-201.rev.frii.com [65.183.77.201]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id CED6DE800D; Mon,  6 Apr 2009 22:19:12 -0600 (MDT)
Message-ID: <49DAD440.8090907@stpeter.im>
Date: Mon, 06 Apr 2009 22:19:12 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <49D1AE22.1080409@webtide.com>	<Pine.LNX.4.62.0903310624180.25058@hixie.dreamhostps.com>	<49D5AA5F.8030503@webtide.com>	<Pine.LNX.4.62.0904030757200.25082@hixie.dreamhostps.com>	<DC8B3ACC-329A-40A4-A41D-70F580350CDA@apple.com>	<Pine.LNX.4.62.0904032027050.25082@hixie.dreamhostps.com> <49D69573.8020508@stpeter.im> <49D6A6DB.9040701@webtide.com>
In-Reply-To: <49D6A6DB.9040701@webtide.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms060709040903000207090402"
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 04:18:17 -0000

This is a cryptographically signed message in MIME format.

--------------ms060709040903000207090402
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/3/09 6:16 PM, Greg Wilkins wrote:

> What I'd like to see from the IETF are recommendations about
> how to better use HTTP as it exists today for bidirectional
> purposes: eg.
> 
>   + should long-polling clients be exempt from the 2 connection
>     recommendation of RFC2616? if so why
>   + should we be recommending that HTTP clients allow some
>     application control of pipeling etc.
>   + can we agree a standard header to identify long poll
>     requests.
> 
> Such recommendations could start having an immediate impact.

Agreed. Those would be part of a "best practices" I-D.

Perhaps it's time to start writing I-Ds that capture some of the
conversations on this list?

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms060709040903000207090402
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms060709040903000207090402--

From ian@hixie.ch  Mon Apr  6 21:52:38 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7453A3A6822 for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:52:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.195
X-Spam-Level: 
X-Spam-Status: No, score=-3.195 tagged_above=-999 required=5 tests=[AWL=-0.596, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Sdk3VwNBMUFt for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 21:52:37 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 0292C3A680A for <hybi@ietf.org>; Mon,  6 Apr 2009 21:52:36 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 213D7125B51; Mon,  6 Apr 2009 21:53:43 -0700 (PDT)
Date: Tue, 7 Apr 2009 04:53:43 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49DAD098.8060703@webtide.com>
Message-ID: <Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com> <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com> <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp> <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com> <49DAD098.8060703@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 04:52:38 -0000

On Tue, 7 Apr 2009, Greg Wilkins wrote:
> > 
> > The WebSocket protocol is extensible by design. It supports two kinds 
> > of frames (0xFF-sentinel-delimited and predeclared-length), and there 
> > can be up to 128 subtypes of each kind. The current version only 
> > exposes one of these, with a payload defined to be UTF-8.
> 
> this is extensibility by the protocol specifiers, so that later versions 
> of the protocol can introduce new packets without breaking backwards 
> compatibility.

Right.


> But HTTP is user extensible.  So if I invent a new content type, 
> encoding or compression algorithm, I don't need to wait for HTTP/1.2 to 
> use that.  I don't need to apply to some standards body to allocate an 
> identifier from a scarce (128) space.
>
> Defining subtypes for compress-utf8-data, gzip-utf8-data, gzip-utf8-data 
> will use up a 128 space very quickly.

Sure, but that won't happen. The protocol has one purpose: to be the 
backend for an API exposed by browsers to Web page scripts. Web Sockets 
would be a terrible protocol to use for other purposes -- just use TCP!

Since it is highly unlikely that browser vendors would be interested in 
adding hundreds of different ways of encoding data, the protocol is 
unlikely to need many of its frame types. We wouldn't want anyone to be 
extending the protocol without updating the API; that would be missing 
the point, as browsers wouldn't know how to support those types.

Having said that, if we do find we're running out, we can just define that 
types 0x7F and 0xFF will have further information at the start of the 
payload saying what the type is.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From mridulm80@yahoo.com  Mon Apr  6 22:36:50 2009
Return-Path: <mridulm80@yahoo.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E525B3A686C for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 22:36:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.901
X-Spam-Level: 
X-Spam-Status: No, score=-1.901 tagged_above=-999 required=5 tests=[AWL=-0.698, BAYES_00=-2.599, MIME_QP_LONG_LINE=1.396]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gs6dGEzAKldk for <hybi@core3.amsl.com>; Mon,  6 Apr 2009 22:36:50 -0700 (PDT)
Received: from n1a.bullet.in.yahoo.com (n1a.bullet.in.yahoo.com [202.43.219.18]) by core3.amsl.com (Postfix) with SMTP id 133743A6D63 for <hybi@ietf.org>; Mon,  6 Apr 2009 22:35:49 -0700 (PDT)
Received: from [202.43.219.99] by n1.bullet.in.yahoo.com with NNFMP; 07 Apr 2009 05:36:55 -0000
Received: from [203.104.18.52] by t3.bullet.in.yahoo.com with NNFMP; 07 Apr 2009 05:36:55 -0000
Received: from [127.0.0.1] by omp113.mail.in2.yahoo.com with NNFMP; 07 Apr 2009 05:36:55 -0000
X-Yahoo-Newman-Property: ymail-3
X-Yahoo-Newman-Id: 2761.49059.bm@omp113.mail.in2.yahoo.com
Received: (qmail 77273 invoked by uid 60001); 7 Apr 2009 05:36:54 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1239082614; bh=2z0qq9TLbg0rIZVLOw9D/J/peswzaIgVhdIAo4PEb/I=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=R7GwzHzwpsX66SRY+/+4JUnaCEvfiFKNxBmNmZR38jQjq1VSszJMp7GvtftpwX2Cdth4tCPmdl82VMB5aPT7UytTCluafAoxQg5PFr9ehxwl3jFlZ6DlVHZ7A4wbTrYRBktYDL3QCCbvbidwrf+zmZGxY/xppLPITTHl5uEE/4Y=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=Z47eBOWEcKbUdP4FwINep//4QTYv+K84fm+mGjM8C/m3P11B2c8zwTW5CA6ulyrSegMR/enwcuBVjfCcMatnvwnwVFY1YdK27X3kq0Hd016WXSarL6hrGAyBhzLfSkiTYyGLK+Ot6IbfNidQGtO8cvFMkRj8SHM6ZWGbIRfMAvw=;
Message-ID: <883400.49860.qm@web95404.mail.in2.yahoo.com>
X-YMail-OSG: KP3NvcIVM1msvFCHWmd8X_1mx_uOWAJ9.uJdQjCaMbMB1WMARj0CB68t1932eDde0ygL5N2LiNAH6ToYInm6mPTQ_gtgkbqi5iwbJyiZegYO4DjhVJYsAvrBBKa1DWcSmzL6BTUSuJb2dTuHTWh.0GFBlPNwt.MNuAUmtZvv7Pui4kzqfZWTuIFcFyX0qCLbLBfTRx6iDl6KL6fX2ikw9hUQWJU-
Received: from [122.167.194.242] by web95404.mail.in2.yahoo.com via HTTP; Tue, 07 Apr 2009 11:06:54 IST
X-Mailer: YahooMailClassic/5.2.15 YahooMailWebService/0.7.289.1
Date: Tue, 7 Apr 2009 11:06:54 +0530 (IST)
From: Mridul Muralidharan <mridulm80@yahoo.com>
To: Greg Wilkins <gregw@webtide.com>, Ian Hickson <ian@hixie.ch>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 05:36:51 -0000

=0ANote that browser based javascript clients are not the only consumers of=
 bi-directional http; others like mobile clients, firewalled thick clients,=
 etc are also possible (and actually exist with BOSH for example).=0A=0ASo =
possibly targeting specific implementation technology might not be appropri=
ate for designing the spec ?=0A=0ARegards,=0AMridul=0A=0APS: Not particular=
ly responding to this message - the thread as such seemed to have a strong =
javascript bias - particularly for WebSockets design parts.=0A=0A--- On Tue=
, 7/4/09, Ian Hickson <ian@hixie.ch> wrote:=0A=0A> From: Ian Hickson <ian@h=
ixie.ch>=0A> Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty p=
erspective.)=0A> To: "Greg Wilkins" <gregw@webtide.com>=0A> Cc: hybi@ietf.o=
rg=0A> Date: Tuesday, 7 April, 2009, 10:23 AM=0A> On Tue, 7 Apr 2009, Greg =
Wilkins=0A> wrote:=0A> > > =0A> > > The WebSocket protocol is extensible by=
 design.=0A> It supports two kinds =0A> > > of frames (0xFF-sentinel-delimi=
ted and=0A> predeclared-length), and there =0A> > > can be up to 128 subtyp=
es of each kind. The=0A> current version only =0A> > > exposes one of these=
, with a payload defined to=0A> be UTF-8.=0A> > =0A> > this is extensibilit=
y by the protocol specifiers, so=0A> that later versions =0A> > of the prot=
ocol can introduce new packets without=0A> breaking backwards =0A> > compat=
ibility.=0A> =0A> Right.=0A> =0A> =0A> > But HTTP is user extensible.=A0 So=
 if I invent a=0A> new content type, =0A> > encoding or compression algorit=
hm, I don't need to=0A> wait for HTTP/1.2 to =0A> > use that.=A0 I don't ne=
ed to apply to some=0A> standards body to allocate an =0A> > identifier fro=
m a scarce (128) space.=0A> >=0A> > Defining subtypes for compress-utf8-dat=
a,=0A> gzip-utf8-data, gzip-utf8-data =0A> > will use up a 128 space very q=
uickly.=0A> =0A> Sure, but that won't happen. The protocol has one purpose:=
=0A> to be the =0A> backend for an API exposed by browsers to Web page scri=
pts.=0A> Web Sockets =0A> would be a terrible protocol to use for other pur=
poses --=0A> just use TCP!=0A> =0A> Since it is highly unlikely that browse=
r vendors would be=0A> interested in =0A> adding hundreds of different ways=
 of encoding data, the=0A> protocol is =0A> unlikely to need many of its fr=
ame types. We wouldn't want=0A> anyone to be =0A> extending the protocol wi=
thout updating the API; that would=0A> be missing =0A> the point, as browse=
rs wouldn't know how to support those=0A> types.=0A> =0A> Having said that,=
 if we do find we're running out, we can=0A> just define that =0A> types 0x=
7F and 0xFF will have further information at the=0A> start of the =0A> payl=
oad saying what the type is.=0A> =0A> -- =0A> Ian Hickson=A0 =A0 =A0 =A0 =
=A0 =A0=0A> =A0=A0=A0U+1047E=A0 =A0 =A0 =A0 =A0=0A> =A0 =A0 =A0 )\._.,--...=
..,'``.=A0 =A0 fL=0A> http://ln.hixie.ch/=A0 =A0=0A> =A0=A0=A0U+263A=A0 =A0 =
=A0 =A0 =A0=0A> =A0 =A0 =A0 /,=A0=A0=A0_..=0A> \=A0=A0=A0_\=A0 ;`._ ,.=0A> =
Things that are impossible just take=0A> longer.=A0=A0=A0`._.-(,_..'--(,_..=
'`-.;.'=0A> _______________________________________________=0A> hybi mailin=
g list=0A> hybi@ietf.org=0A> https://www.ietf.org/mailman/listinfo/hybi=0A>=
 =0A=0A=0A      Bring your gang together. Do your thing. Find your favourit=
e Yahoo! group at http://in.promos.yahoo.com/groups/


From sh@defuze.org  Tue Apr  7 02:25:59 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8382228C189 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 02:25:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.115
X-Spam-Level: 
X-Spam-Status: No, score=-2.115 tagged_above=-999 required=5 tests=[AWL=0.150,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fsjl6Ouom6Zc for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 02:25:58 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 8604728C132 for <hybi@ietf.org>; Tue,  7 Apr 2009 02:25:58 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n376suQJ030894 for <hybi@ietf.org>; Tue, 7 Apr 2009 01:54:56 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n376su1Y030892; Tue, 7 Apr 2009 08:54:56 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Tue, 7 Apr 2009 08:54:56 +0200 (CEST)
Message-ID: <56551.193.253.216.132.1239087296.squirrel@mail1.webfaction.com>
In-Reply-To: <Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net> <Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com> <6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com> <Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp> <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com> <49DAD098.8060703@webtide.com> <Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com>
Date: Tue, 7 Apr 2009 08:54:56 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: hybi@ietf.org
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 09:25:59 -0000

>
>> But HTTP is user extensible.  So if I invent a new content type,
>> encoding or compression algorithm, I don't need to wait for HTTP/1.2 to
>> use that.  I don't need to apply to some standards body to allocate an
>> identifier from a scarce (128) space.
>>
>> Defining subtypes for compress-utf8-data, gzip-utf8-data, gzip-utf8-data
>> will use up a 128 space very quickly.
>
> Sure, but that won't happen. The protocol has one purpose: to be the
> backend for an API exposed by browsers to Web page scripts. Web Sockets
> would be a terrible protocol to use for other purposes -- just use TCP!

I think this long exchange comes from different set of expectations. You
initiated this discussion with a few common use cases you consider being
worthy enough of a new, however small, protocol.

Unfortunately different people have different background. As a developer I
find it miserable not to have direct access to TCP sockets from Javascript
so that I could converse with a XMPP server without having to relay on
BOSH. As an infrastructure/system person I would probably see it with a
keen eye that web applications can't open sockets up at will from an
already prone to security issue that is the browser.

In this context I can understand why your draft is so short and that's
probably a good sign because it'll be that easier to see deployed.
Nonetheless this long discussion makes me wonder whether or not having a
new protocol is a better idea than picking one of the existing
alternatives (Comet, BOSH). Wouldn't it be better to focus energy on best
practices when it comes to long polling connections based on what's
working now?

Indeed, it'd be already a good starting point listing the shortcomings of
those existing alternatives. What are the showstoppers? In which use case?

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From stpeter@stpeter.im  Tue Apr  7 08:16:46 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 249A33A6DB3 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 08:16:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.505
X-Spam-Level: 
X-Spam-Status: No, score=-2.505 tagged_above=-999 required=5 tests=[AWL=0.094,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id V5Q-zuUi+-Lo for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 08:16:45 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 33D513A6860 for <hybi@ietf.org>; Tue,  7 Apr 2009 08:16:45 -0700 (PDT)
Received: from wrk165.corp.jabber.com (dencfw1.jabber.com [207.182.164.5]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id 3CBD4E800D; Tue,  7 Apr 2009 09:17:50 -0600 (MDT)
Message-ID: <49DB6E9D.3080700@stpeter.im>
Date: Tue, 07 Apr 2009 09:17:49 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: sh@defuze.org
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net> <49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>	<6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>	<Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>	<49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp>	<Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com>	<49DAD098.8060703@webtide.com>	<Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com> <56551.193.253.216.132.1239087296.squirrel@mail1.webfaction.com>
In-Reply-To: <56551.193.253.216.132.1239087296.squirrel@mail1.webfaction.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms050301050903050807020401"
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 15:16:46 -0000

This is a cryptographically signed message in MIME format.

--------------ms050301050903050807020401
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/7/09 12:54 AM, Sylvain Hellegouarch wrote:

> Nonetheless this long discussion makes me wonder whether or not having a
> new protocol is a better idea than picking one of the existing
> alternatives (Comet, BOSH). Wouldn't it be better to focus energy on best
> practices when it comes to long polling connections based on what's
> working now?
> 
> Indeed, it'd be already a good starting point listing the shortcomings of
> those existing alternatives. What are the showstoppers? In which use case?

This is why we need to write an I-D that describes the use cases. :)

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms050301050903050807020401
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms050301050903050807020401--

From frank.salim@kaazing.com  Tue Apr  7 10:47:56 2009
Return-Path: <frank.salim@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D89A93A68DB for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 10:47:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.976
X-Spam-Level: 
X-Spam-Status: No, score=-1.976 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id E+CnMns7abBh for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 10:47:56 -0700 (PDT)
Received: from yx-out-2324.google.com (yx-out-2324.google.com [74.125.44.28]) by core3.amsl.com (Postfix) with ESMTP id ED31F3A68E7 for <hybi@ietf.org>; Tue,  7 Apr 2009 10:47:55 -0700 (PDT)
Received: by yx-out-2324.google.com with SMTP id 8so3768791yxm.49 for <hybi@ietf.org>; Tue, 07 Apr 2009 10:49:02 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.88.17 with SMTP id l17mr422007agb.8.1239126541933; Tue, 07  Apr 2009 10:49:01 -0700 (PDT)
In-Reply-To: <49DB6E9D.3080700@stpeter.im>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com> <6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com> <Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com> <49DA9713.5050407@webtide.com> <49DABCC5.5060809@it.aoyama.ac.jp> <Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com> <49DAD098.8060703@webtide.com> <Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com> <56551.193.253.216.132.1239087296.squirrel@mail1.webfaction.com> <49DB6E9D.3080700@stpeter.im>
Date: Tue, 7 Apr 2009 10:49:01 -0700
Message-ID: <b7197f6c0904071049i93a3effxa3ab460dade9b301@mail.gmail.com>
From: Frank Salim <frank.salim@kaazing.com>
To: Peter Saint-Andre <stpeter@stpeter.im>
Content-Type: multipart/alternative; boundary=0016361649e75d1e4d0466faa12e
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 07 Apr 2009 17:47:56 -0000

--0016361649e75d1e4d0466faa12e
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

> Indeed, it'd be already a good starting point listing the shortcomings of
> those existing alternatives. What are the showstoppers? In which use case?

I started filling out the wiki with that information.
http://trac.tools.ietf.org/bof/trac/wiki/HyBi

-Frank

On Tue, Apr 7, 2009 at 8:17 AM, Peter Saint-Andre <stpeter@stpeter.im>wrote:

> On 4/7/09 12:54 AM, Sylvain Hellegouarch wrote:
>
> > Nonetheless this long discussion makes me wonder whether or not having a
> > new protocol is a better idea than picking one of the existing
> > alternatives (Comet, BOSH). Wouldn't it be better to focus energy on best
> > practices when it comes to long polling connections based on what's
> > working now?
> >
> > Indeed, it'd be already a good starting point listing the shortcomings of
> > those existing alternatives. What are the showstoppers? In which use
> case?
>
> This is why we need to write an I-D that describes the use cases. :)
>
> Peter
>
> --
> Peter Saint-Andre
> https://stpeter.im/
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>
>

--0016361649e75d1e4d0466faa12e
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

&gt; Indeed, it&#39;d be already a good starting point listing the shortcom=
ings of<br>
&gt; those existing alternatives. What are the showstoppers? In which use c=
ase?<br><br>I started filling out the wiki with that information.<br><a hre=
f=3D"http://trac.tools.ietf.org/bof/trac/wiki/HyBi">http://trac.tools.ietf.=
org/bof/trac/wiki/HyBi</a><br>
<br>-Frank<br><br><div class=3D"gmail_quote">On Tue, Apr 7, 2009 at 8:17 AM=
, Peter Saint-Andre <span dir=3D"ltr">&lt;<a href=3D"mailto:stpeter@stpeter=
.im">stpeter@stpeter.im</a>&gt;</span> wrote:<br><blockquote class=3D"gmail=
_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt=
 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">On 4/7/09 12:54 AM, Sylvain Hellegouarch wrote:<br>
<br>
&gt; Nonetheless this long discussion makes me wonder whether or not having=
 a<br>
&gt; new protocol is a better idea than picking one of the existing<br>
&gt; alternatives (Comet, BOSH). Wouldn&#39;t it be better to focus energy =
on best<br>
&gt; practices when it comes to long polling connections based on what&#39;=
s<br>
&gt; working now?<br>
&gt;<br>
&gt; Indeed, it&#39;d be already a good starting point listing the shortcom=
ings of<br>
&gt; those existing alternatives. What are the showstoppers? In which use c=
ase?<br>
<br>
</div>This is why we need to write an I-D that describes the use cases. :)<=
br>
<div><div></div><div class=3D"h5"><br>
Peter<br>
<br>
--<br>
Peter Saint-Andre<br>
<a href=3D"https://stpeter.im/" target=3D"_blank">https://stpeter.im/</a><b=
r>
<br>
</div></div><br>_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
<br></blockquote></div><br>

--0016361649e75d1e4d0466faa12e--

From duerst@it.aoyama.ac.jp  Tue Apr  7 18:46:32 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 53E553A6E13 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 18:46:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.365
X-Spam-Level: 
X-Spam-Status: No, score=-0.365 tagged_above=-999 required=5 tests=[AWL=-0.575, BAYES_00=-2.599, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AU5bUW+XcEv1 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 18:46:30 -0700 (PDT)
Received: from scmailgw1.scop.aoyama.ac.jp (scmailgw1.scop.aoyama.ac.jp [133.2.251.194]) by core3.amsl.com (Postfix) with ESMTP id 2FD223A6E11 for <hybi@ietf.org>; Tue,  7 Apr 2009 18:46:29 -0700 (PDT)
Received: from scmse3.scbb.aoyama.ac.jp ([133.2.253.23]) by scmailgw1.scop.aoyama.ac.jp (secret/secret) with SMTP id n381lYKo027752 for <hybi@ietf.org>; Wed, 8 Apr 2009 10:47:35 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse3.scbb.aoyama.ac.jp with smtp id 7ef4_3b491aa6_23df_11de_9540_001d0969ab06; Wed, 08 Apr 2009 10:47:34 +0900
Received: from [IPv6:::1] ([133.2.210.1]:35940) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC7B419> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Wed, 8 Apr 2009 10:46:40 +0900
Message-ID: <49DC0222.8020903@it.aoyama.ac.jp>
Date: Wed, 08 Apr 2009 10:47:14 +0900
From: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Peter Saint-Andre <stpeter@stpeter.im>
References: <1cb725390904051000m45e9db66geb038acb452d6479@mail.gmail.com>	<49D8FE80.2010100@lshift.net>	<49D9008B.9060606@lshift.net>	<Pine.LNX.4.62.0904051912520.25058@hixie.dreamhostps.com>	<6fc58d0d0904061308r374490c5w43164b24b67a9c73@mail.gmail.com>	<Pine.LNX.4.62.0904062011220.25058@hixie.dreamhostps.com>	<6fc58d0d0904061545s4a5d10afh331f602037415cc8@mail.gmail.com>	<Pine.LNX.4.62.0904062315280.25058@hixie.dreamhostps.com>	<49DA9713.5050407@webtide.com>	<49DABCC5.5060809@it.aoyama.ac.jp>	<Pine.LNX.4.62.0904070241560.25058@hixie.dreamhostps.com>	<49DAD098.8060703@webtide.com>	<Pine.LNX.4.62.0904070439460.18470@hixie.dreamhostps.com>	<56551.193.253.216.132.1239087296.squirrel@mail1.webfaction.com> <49DB6E9D.3080700@stpeter.im>
In-Reply-To: <49DB6E9D.3080700@stpeter.im>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSockets feedback (Was: Bayeux / Jetty perspective.)
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 01:46:32 -0000

> On 4/7/09 12:54 AM, Sylvain Hellegouarch wrote:
>
>> Nonetheless this long discussion makes me wonder whether or not having a
>> new protocol is a better idea than picking one of the existing
>> alternatives (Comet, BOSH). Wouldn't it be better to focus energy on best
>> practices when it comes to long polling connections based on what's
>> working now?

Writing this up is definitely a good idea. But the things that are 
working now, and the things that people might want to do once long-time 
connections are available might differ considerably.

Regards,   Martin.

-- 
#-# Martin J. DÃ¼rst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From duerst@it.aoyama.ac.jp  Tue Apr  7 18:50:41 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A5E163A6E05 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 18:50:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.864
X-Spam-Level: 
X-Spam-Status: No, score=0.864 tagged_above=-999 required=5 tests=[AWL=-1.760,  BAYES_40=-0.185, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id NlAd1BrVcXV1 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 18:50:40 -0700 (PDT)
Received: from scmailgw2.scop.aoyama.ac.jp (scmailgw2.scop.aoyama.ac.jp [133.2.251.195]) by core3.amsl.com (Postfix) with ESMTP id 15FAC3A6E29 for <hybi@ietf.org>; Tue,  7 Apr 2009 18:50:39 -0700 (PDT)
Received: from scmse2.scbb.aoyama.ac.jp (scmse2 [133.2.253.17]) by scmailgw2.scop.aoyama.ac.jp (secret/secret) with SMTP id n381pjw6021443 for <hybi@ietf.org>; Wed, 8 Apr 2009 10:51:45 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse2.scbb.aoyama.ac.jp with smtp id 0eaa_d0b8bfd8_23df_11de_8f40_0019b9e2b3d9; Wed, 08 Apr 2009 10:51:45 +0900
Received: from [IPv6:::1] ([133.2.210.1]:35941) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC7B5F5> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Wed, 8 Apr 2009 10:50:51 +0900
Message-ID: <49DC031D.9090007@it.aoyama.ac.jp>
Date: Wed, 08 Apr 2009 10:51:25 +0900
From: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>, hybi@ietf.org
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Subject: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 01:50:41 -0000

Hello Ian,

In your postings, you gave quite a few examples of the API a browser 
would use to send data on the WebSockets protocol. That seems quite easy 
and straightforward. However, what I wonder is how the API on the 
receiver side (because we are bidirectional, this may also be the 
browser) may/will look. By experience, the receiving side is always more 
tricky than the sending side.

Regards,    Martin.

-- 
#-# Martin J. DÃ¼rst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From ian@hixie.ch  Tue Apr  7 22:22:47 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3234F3A6A5C for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:22:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.135
X-Spam-Level: 
X-Spam-Status: No, score=-3.135 tagged_above=-999 required=5 tests=[AWL=-0.836, BAYES_00=-2.599, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id F5SRJORpR5tn for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:22:46 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 3FC083A6AC7 for <hybi@ietf.org>; Tue,  7 Apr 2009 22:22:46 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 2D349125B51; Tue,  7 Apr 2009 22:23:53 -0700 (PDT)
Date: Wed, 8 Apr 2009 05:23:52 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
In-Reply-To: <49DC031D.9090007@it.aoyama.ac.jp>
Message-ID: <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com>
References: <49DC031D.9090007@it.aoyama.ac.jp>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="1909464018-2104290481-1239168232=:18470"
Cc: hybi@ietf.org
Subject: Re: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 05:22:47 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--1909464018-2104290481-1239168232=:18470
Content-Type: TEXT/PLAIN; charset=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Wed, 8 Apr 2009, "Martin J. D=C3=BCrst" wrote:
>=20
> In your postings, you gave quite a few examples of the API a browser=20
> would use to send data on the WebSockets protocol. That seems quite easy=
=20
> and straightforward. However, what I wonder is how the API on the=20
> receiver side (because we are bidirectional, this may also be the=20
> browser) may/will look. By experience, the receiving side is always more=
=20
> tricky than the sending side.

The server-side implementation is up to the author.

--=20
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
--1909464018-2104290481-1239168232=:18470--

From prescod@gmail.com  Tue Apr  7 22:38:59 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B2F6D3A6E1F for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:38:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.512
X-Spam-Level: 
X-Spam-Status: No, score=-1.512 tagged_above=-999 required=5 tests=[AWL=0.465,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BnJc1gFNQtq0 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:38:59 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.178]) by core3.amsl.com (Postfix) with ESMTP id 029F13A6E36 for <hybi@ietf.org>; Tue,  7 Apr 2009 22:38:58 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so1506246waf.5 for <hybi@ietf.org>; Tue, 07 Apr 2009 22:40:06 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=QFwjbpgRzsZ9yjux2cfsyLDPqFLbNialtz0/ilVkoN8=; b=r4mlyfMMA6TgrtOizQIKf/vSD1h+MV+uoagC+ItyEh/uCP3MBxwogY8Pa2XmN2T2Ys 4MutQR19OcMjZH2R0FurZl29pwJAm/M+SNCATHcmGqCLC5eF40LTwsZje+yh6W9On66U 917yjVHxX3inr90ontVeNjo+d6m7T7leo45ms=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=NddA6QJfAtRHOslWHcToEC5fyETIMWjYyYy8JTNPABaWy+NcRe047y9twYaUuFtbZZ OPkG120PlgkOnm3Rnkf79ufSba7SyhSVrFAX7O2COiENEYAFBgfiDlWJgyj1/lijjTKO EYdpJ0ubdTraPUHpI4B34DLGEbYWosmAx4E+M=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.140.174.20 with SMTP id w20mr2780558rve.0.1239169205942; Tue,  07 Apr 2009 22:40:05 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com>
References: <49DC031D.9090007@it.aoyama.ac.jp> <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com>
Date: Tue, 7 Apr 2009 22:40:05 -0700
X-Google-Sender-Auth: 6933402c8e35bb5a
Message-ID: <1cb725390904072240o340230a5p2f772b35df910d9b@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 05:38:59 -0000

On Tue, Apr 7, 2009 at 10:23 PM, Ian Hickson <ian@hixie.ch> wrote:
> On Wed, 8 Apr 2009, "Martin J. D=C3=BCrst" wrote:
>>
>> In your postings, you gave quite a few examples of the API a browser
>> would use to send data on the WebSockets protocol. That seems quite easy
>> and straightforward. However, what I wonder is how the API on the
>> receiver side (because we are bidirectional, this may also be the
>> browser) may/will look. By experience, the receiving side is always more
>> tricky than the sending side.
>
> The server-side implementation is up to the author.

I believe that Martin was asking about the details onmessage event
handler and the MessageEvent interface.

 Paul

From ian@hixie.ch  Tue Apr  7 22:47:40 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B2EAF3A69A2 for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:47:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.259
X-Spam-Level: 
X-Spam-Status: No, score=-3.259 tagged_above=-999 required=5 tests=[AWL=-0.660, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UJ3uW8quOUZm for <hybi@core3.amsl.com>; Tue,  7 Apr 2009 22:47:39 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id A2B5B3A69AD for <hybi@ietf.org>; Tue,  7 Apr 2009 22:45:38 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id E18A2125B51; Tue,  7 Apr 2009 22:46:45 -0700 (PDT)
Date: Wed, 8 Apr 2009 05:46:45 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904072240o340230a5p2f772b35df910d9b@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904080545200.18470@hixie.dreamhostps.com>
References: <49DC031D.9090007@it.aoyama.ac.jp>  <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com> <1cb725390904072240o340230a5p2f772b35df910d9b@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="1909464018-304922226-1239169605=:18470"
Cc: hybi@ietf.org
Subject: Re: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 05:47:40 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--1909464018-304922226-1239169605=:18470
Content-Type: TEXT/PLAIN; charset=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Tue, 7 Apr 2009, Paul Prescod wrote:
> On Tue, Apr 7, 2009 at 10:23 PM, Ian Hickson <ian@hixie.ch> wrote:
> > On Wed, 8 Apr 2009, "Martin J. D=C3=BCrst" wrote:
> >>
> >> In your postings, you gave quite a few examples of the API a browser
> >> would use to send data on the WebSockets protocol. That seems quite ea=
sy
> >> and straightforward. However, what I wonder is how the API on the
> >> receiver side (because we are bidirectional, this may also be the
> >> browser) may/will look. By experience, the receiving side is always mo=
re
> >> tricky than the sending side.
> >
> > The server-side implementation is up to the author.
>=20
> I believe that Martin was asking about the details onmessage event=20
> handler and the MessageEvent interface.

Oh. This is defined in the same place as the sending API:

   http://dev.w3.org/html5/websockets/

The MessageEvent interface is defined in HTML5:

   http://www.whatwg.org/specs/web-apps/current-work/#messageevent

HTH,
--=20
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
--1909464018-304922226-1239169605=:18470--

From duerst@it.aoyama.ac.jp  Wed Apr  8 02:26:32 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 354903A6E4F for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 02:26:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.278
X-Spam-Level: 
X-Spam-Status: No, score=-0.278 tagged_above=-999 required=5 tests=[AWL=-0.488, BAYES_00=-2.599, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hCTeDnAltSrm for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 02:26:31 -0700 (PDT)
Received: from scmailgw2.scop.aoyama.ac.jp (scmailgw2.scop.aoyama.ac.jp [133.2.251.195]) by core3.amsl.com (Postfix) with ESMTP id BC8F93A6836 for <hybi@ietf.org>; Wed,  8 Apr 2009 02:26:30 -0700 (PDT)
Received: from scmse3.scbb.aoyama.ac.jp ([133.2.253.23]) by scmailgw2.scop.aoyama.ac.jp (secret/secret) with SMTP id n389RaNJ014953 for <hybi@ietf.org>; Wed, 8 Apr 2009 18:27:37 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse3.scbb.aoyama.ac.jp with smtp id 5bf1_7f600822_241f_11de_befa_001d0969ab06; Wed, 08 Apr 2009 18:27:36 +0900
Received: from [IPv6:::1] ([133.2.210.1]:36377) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC81012> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Wed, 8 Apr 2009 18:26:42 +0900
Message-ID: <49DC6DF4.6080305@it.aoyama.ac.jp>
Date: Wed, 08 Apr 2009 18:27:16 +0900
From: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49DC031D.9090007@it.aoyama.ac.jp> <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 09:26:32 -0000

I was not (only) talking about the server side implementation.
I'm assuming that WebSockets, like TCP, is bidirectional. So I'm
assuming that the server can send the client the same kind of packages
that the client can send the server (UTF-8 and length-limited binary,
the later at least planned).

Based on that assumption, I'm wondering what the receiving API on the 
*client* side is going to look. I was thinking that was clear enough 
from my mail, and I hope it's clear now.

Regards,   Martin.

On 2009/04/08 14:23, Ian Hickson wrote:
> On Wed, 8 Apr 2009, "Martin J. DÃ¼rst" wrote:
>> In your postings, you gave quite a few examples of the API a browser
>> would use to send data on the WebSockets protocol. That seems quite easy
>> and straightforward. However, what I wonder is how the API on the
>> receiver side (because we are bidirectional, this may also be the
>> browser) may/will look. By experience, the receiving side is always more
>> tricky than the sending side.
>
> The server-side implementation is up to the author.
>

-- 
#-# Martin J. DÃ¼rst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From markl@lindenlab.com  Wed Apr  8 08:46:23 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A65F23A6A73 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 08:46:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.559
X-Spam-Level: 
X-Spam-Status: No, score=-3.559 tagged_above=-999 required=5 tests=[AWL=0.040,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id LQ5D7Ffk0ys1 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 08:46:22 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id E6EDC3A6856 for <hybi@ietf.org>; Wed,  8 Apr 2009 08:46:22 -0700 (PDT)
Received: from 945battery-guestB-139.lindenlab.com (945battery-guestB-139.lindenlab.com [10.1.11.139]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 094D73DBC070 for <hybi@ietf.org>; Wed,  8 Apr 2009 08:47:30 -0700 (PDT)
Message-Id: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Wed, 8 Apr 2009 08:47:29 -0700
X-Mailer: Apple Mail (2.930.3)
Subject: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 15:46:23 -0000

 From the point of view of our customers, the application level  
authors (both client and server side), I think there are two major  
questions they would ask to characterize any potential protocol:

1) What is the data model?
	Minimal (binary blob and/or UTF-8 string)
	Structured (JSON object or XML stanza)
	Message (data and meta-data)

2) What is interaction model?
	Event (datagram at the senders whim)
	Request/Response (request at senders whim, with response from the  
receiver)

While issues of data framing, efficient use of TCP, and ease of  
implementation, will influence their view, I don't think they will be  
nearly as important as these first two questions[*].

The protocols we're discussing can be characterized by these two  
questions:
	WebSockets: Minimal, Event
	BOSH, Bayeaux: Structured, Event
	rHTTP:  Message, Req/Resp

Other protocols can be fit into this framework, with perhaps some  
extension:
	XMPP: Structured , Event
	TCP: Minimal, Stream
	SMTP: Message, Event
	HTTP: Message, Req/Resp

I think the first thing to look at is which 'buckets' are we trying to  
serve? And only then compare items within a bucket. I don't think it  
makes sense to discuss WebSocket vs. rHTTP -- The discussion will  
inevitable be more about Minimal Event vs. Message Req/Resp.

Now, while clearly any of these be used to implement the others, there  
is a cost. Where simpler models are implemented with more complex,  
then there is overhead on the wire and processing. Where more complex  
models are implemented on simpler, then the burden falls to the  
application developer.

I'm think it will be productive to discuss which models we are trying  
to serve. I suppose this is the same sort of thing Roy has been asking  
for. I'll try to get this into the wiki, too. I'll leave my personal  
views for a second post...

	- Mark

[*] Okay, I admit, there is actually one even more important question:

0) Is the protocol available in the environment that the application  
is built on?

For protocols described here, some are trying to support the web-page  
environment, others are not, though often could. Kudos for WebSockets  
for describing a DOM level API. I realize that it isn't generally IETF  
practice to do so, but having a description of a base expected API  
(not to mention implementations) is actually very important to  
developers.


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From ian@hixie.ch  Wed Apr  8 12:47:36 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E128C3A6859 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 12:47:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.431
X-Spam-Level: 
X-Spam-Status: No, score=-3.431 tagged_above=-999 required=5 tests=[AWL=-0.832, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oXMgS3jcNHcz for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 12:47:36 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 10D043A68B5 for <hybi@ietf.org>; Wed,  8 Apr 2009 12:47:35 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 45EC7125C85; Wed,  8 Apr 2009 12:48:43 -0700 (PDT)
Date: Wed, 8 Apr 2009 19:48:43 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Mark Lentczner <markl@lindenlab.com>
In-Reply-To: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
Message-ID: <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 19:47:37 -0000

On Wed, 8 Apr 2009, Mark Lentczner wrote:
> 
> I think the first thing to look at is which 'buckets' are we trying to serve?

I think it would make sense to serve several of these buckets. Personally 
I am interested in "Minimal, Event", but I don't see any reason why we 
wouldn't also have the other models.

It's not clear to me how we should proceed, though. Would it make sense to 
split into subgroups to work on the various buckets that people are 
interested in?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Wed Apr  8 13:01:03 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 37D943A6A2D for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 13:01:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.559
X-Spam-Level: 
X-Spam-Status: No, score=-1.559 tagged_above=-999 required=5 tests=[AWL=0.418,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uZZ2TJt7rRc2 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 13:01:02 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.234]) by core3.amsl.com (Postfix) with ESMTP id 783B73A6A24 for <hybi@ietf.org>; Wed,  8 Apr 2009 13:01:02 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so221699rvb.49 for <hybi@ietf.org>; Wed, 08 Apr 2009 13:02:10 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=pKmNByi4C17VDWwd1t8ZAbfx9kN2KDH2UDYdXohJQKE=; b=Uy+qX9228NAooFFw8E2uv+acgCpUWj2HALUAY8Ap3PfUhCJq/UFN1C78Nj2bFuMhFN Pj3Gr4hEIqDy/W4QcoZFZ2sdmX3XoLewZSVnw0IUcLZXNumnXKIOXhqiGjtdyH4/+WFK b224D+vPJ/AaQSeC6YjBOUIgvd9+HAqzv+fE8=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=eld/AY7FFY2Aaqt0gaXyMUYRDZYU4cKMRPrat3lbPtkdrrkHijnubMvpfQMjm4EcwO 9Z0Qa9hMKU2YrEA3JvnWR6I9+Mhf4owpDQLXvcAT3Hw5lHoCYeoj4wKYbxpZ0OoTkqWv EEbmrlZsnUU0fudnAHdEKY4Ejk8tnuo2wxdWs=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.140.207.2 with SMTP id e2mr664675rvg.148.1239220930028; Wed,  08 Apr 2009 13:02:10 -0700 (PDT)
In-Reply-To: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
Date: Wed, 8 Apr 2009 13:02:10 -0700
X-Google-Sender-Auth: 062ff1d7324cbc66
Message-ID: <1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Mark Lentczner <markl@lindenlab.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 20:01:03 -0000

On Wed, Apr 8, 2009 at 8:47 AM, Mark Lentczner <markl@lindenlab.com> wrote:
> From the point of view of our customers, the application level authors (b=
oth
> client and server side), I think there are two major questions they would
> ask to characterize any potential protocol:
>
> 1) What is the data model?
> =C2=A0 =C2=A0 =C2=A0 =C2=A0Minimal (binary blob and/or UTF-8 string)
> =C2=A0 =C2=A0 =C2=A0 =C2=A0Structured (JSON object or XML stanza)
> =C2=A0 =C2=A0 =C2=A0 =C2=A0Message (data and meta-data)

I would add a "data synchronization" data model. I find that most
often the goal is to have the web browser report the state of a
resource or resource collection to a user, and then inform the user of
changes to it.

The resource might be "chat transcript", "current location of a
train", "objects currently visible in the game", "state of the game
board in a Chess game", "set of currently visible emails."

For example, one could imagine it would be a list of key-value pairs,
or SQL rows or DOM nodes (I worked on such a protocol for DOM nodes a
long time ago).

 Paul Prescod

From gregw@webtide.com  Wed Apr  8 14:13:07 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A134028C1FC for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 14:13:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.886
X-Spam-Level: 
X-Spam-Status: No, score=-1.886 tagged_above=-999 required=5 tests=[AWL=-0.579, BAYES_00=-2.599, MISSING_HEADERS=1.292]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WvGG7tT1CRM0 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 14:13:06 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id 7306128C1CC for <hybi@ietf.org>; Wed,  8 Apr 2009 14:12:59 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so203479tid.25 for <hybi@ietf.org>; Wed, 08 Apr 2009 14:14:06 -0700 (PDT)
Received: by 10.110.40.8 with SMTP id n8mr2281951tin.28.1239225245926; Wed, 08 Apr 2009 14:14:05 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id u8sm1228168tia.30.2009.04.08.14.14.02 (version=TLSv1/SSLv3 cipher=RC4-MD5); Wed, 08 Apr 2009 14:14:04 -0700 (PDT)
Message-ID: <49DD1397.9030903@webtide.com>
Date: Thu, 09 Apr 2009 07:13:59 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
CC: hybi@ietf.org
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 21:13:07 -0000

Ian Hickson wrote:
> It's not clear to me how we should proceed, though. Would it make sense to 
> split into subgroups to work on the various buckets that people are 
> interested in?

I see that there are 3 areas where the IETF could help, so perhaps
we could form subgroups or sub-discussions about:



1) What can we do if we can't change HTTP
-----------------------------------------

Ie is there anything that we can do to help best practises with
the HTTP specification as it is.

We could define common headers that identify long-polls and thus give
hints to infrastructure to better handle them (eg pipeline only other
longpolls behind a pipeline).

Another example of what could be done is that the open ajax alliance wrote
an open list of demands to the browser vendors listing the key pain points
for Ajax / Comet development (eg lack of pipeline control, need for
better interframe eventing etc.).    Perhaps the IETF could do
similar from the network point of view.

Specifically I think the IETF should be saying
is something about the 2 connection limit (either it is or
it is no OK to raise it and the IETF should be able to
clearly say which it is).    We could also recommend how
pipelining as it is currently defined could be better
used for long polling... if only the clients allowed
control of pipelining.  etc.



2) What can we do to improve HTTP for bidirection usage
-------------------------------------------------------

If there was a new HTTP spec developed (and I have no idea
if one is on the horizon or not),  what could we change
to improve it's use for bidirectional use-cases.

My favourite here is 1xx methods with content, but
there are sure to be other things we could change.



3) What can we do to improve usage of other protocols
-----------------------------------------------------

Are there recommendations we can make to improve the
generic handling of upgrade so that proxies and
gateways could be expected to better handle upgrade
to specific (XMPP, WebSocket, BEEP) and arbitrary
TCP based protocols.


cheers






















From ian@hixie.ch  Wed Apr  8 14:47:06 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AE86E3A6998 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 14:47:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.576
X-Spam-Level: 
X-Spam-Status: No, score=-3.576 tagged_above=-999 required=5 tests=[AWL=-0.977, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9RKpRwW9k1pa for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 14:47:05 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id BAC513A672F for <hybi@ietf.org>; Wed,  8 Apr 2009 14:47:05 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 3C0DA27B95; Wed,  8 Apr 2009 14:48:13 -0700 (PDT)
Date: Wed, 8 Apr 2009 21:48:11 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49DD1397.9030903@webtide.com>
Message-ID: <Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com> <49DD1397.9030903@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Apr 2009 21:47:06 -0000

On Thu, 9 Apr 2009, Greg Wilkins wrote:
> Ian Hickson wrote:
> > It's not clear to me how we should proceed, though. Would it make 
> > sense to split into subgroups to work on the various buckets that 
> > people are interested in?
> 
> I see that there are 3 areas where the IETF could help, so perhaps we 
> could form subgroups or sub-discussions about:
> 
> 1) What can we do if we can't change HTTP
> 
> We could define common headers that identify long-polls [...]
>
> 2) What can we do to improve HTTP for bidirection usage
> 
> If there was a new HTTP spec developed (and I have no idea if one is on 
> the horizon or not), what could we change to improve it's use for 
> bidirectional use-cases. [...]
> 
> 3) What can we do to improve usage of other protocols
> 
> Are there recommendations we can make to improve the generic handling of 
> upgrade so that proxies and gateways could be expected to better handle 
> upgrade to specific (XMPP, WebSocket, BEEP) and arbitrary TCP based 
> protocols.

These seem to all boil down to extensions or changes to HTTP. It seems 
that those should probably be discussed in the ietf-http-wg mailing list 
as part of HTTP working group discussions.


> Another example of what could be done is that the open ajax alliance 
> wrote an open list of demands to the browser vendors listing the key 
> pain points for Ajax / Comet development (eg lack of pipeline control, 
> need for better interframe eventing etc.).  Perhaps the IETF could do 
> similar from the network point of view.

The browser vendors are present here on this list, I don't think there's 
any need to write "demands" to them.


I would be interested in taking a specification for a "minimal, event" 
protocol for implementation directly by Web browsers to the standards 
track. This seems orthogonal to the above work. What would be the best way 
to move forward with such a goal?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From markl@lindenlab.com  Wed Apr  8 17:05:33 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C84063A6BB7 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:05:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.562
X-Spam-Level: 
X-Spam-Status: No, score=-3.562 tagged_above=-999 required=5 tests=[AWL=0.037,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 94xpGibewsVO for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:05:32 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id D34EE3A6BB1 for <hybi@ietf.org>; Wed,  8 Apr 2009 17:05:32 -0700 (PDT)
Received: from 945battery-guestB-139.lindenlab.com (945battery-guestB-139.lindenlab.com [10.1.11.139]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 90E383DBC05F for <hybi@ietf.org>; Wed,  8 Apr 2009 17:06:40 -0700 (PDT)
Message-Id: <559D718A-1A14-4DBB-8DFE-F38773466E6F@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Wed, 8 Apr 2009 17:06:40 -0700
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 00:05:33 -0000

Now, for part two, my personal view... (apologies for length)

I see the attraction of just giving something very simple to  
application developers. Giving them a protocol that does Minimal  
Events is easy to explain and would have a simple API. However, I have  
seen that over time, applications often need to build most of the  
framework that is already provided by more complex abstractions. Where  
the cost of those abstractions is small, and the ability to use them  
in simpler ways easy, it is often better to supply a richer  
abstraction to all, since the uniformity of the often implemented bits  
means better code for all.

For example, the overhead of structured data (in XML or JSON or some  
such) is generally seen as small when talking about internet needs  
(other than, say, streaming media). As such, many protocols and APIs  
these days are happy to use a structured data form exclusively, even  
when some applications will only make use of it to transport simpler  
data.

Similarly, Request/Response can be used to implement events, by  
treating the requests as the events, and giving only a minimal  
acknowledgement of receipt as the response (which could be ignored.) I  
note that many event systems at the lower level actually have this  
acknowledgement response (i.e. AMQP).


At Linden Lab, we've found the level of abstraction offered by HTTP's  
messages and request/response framework to be right in the sweet spot:  
It has a lot of the common things applications want (or eventually  
need) in their communications, and adds them in a rather low overhead  
manner. Judging by the large number of APIs written to work over HTTP,  
even where a browser isn't involved, I don't think we're the only ones.

Hence, I'm interested in looking at protocols in Message, Req/Resp.  
bucket.


Now, while clearly any protocol in any of these buckets can be used to  
implement the others, there is a cost. Where simpler models are  
implemented with more complex, then there is overhead on the wire and  
processing. Where more complex models are implemented on simpler, then  
the burden falls to the application developer.

Example: Implementing a Minimal, Event protocol with Message, Req/Resp  
is just a matter of accepting the overhead of the framing of the  
simple data blob, and the cost of the minimal response message. On the  
other hand, implementing a Message, Req/Resp. protocol with Minimal,  
Events, means building a whole framework for encoding message parts  
(address, meta-data and data) into the minimal data units (strings?  
binary blobs?) and coming up with a way to correlate responses to  
requests, etc... all in the application layer.

If you think that the features of Message, Req/Resp. are common  
enough, and the overhead of those features when not used low enough,  
then I think building a Message, Req/Resp. protocol is the winner.  
Having common support in server frameworks, client libraries, and  
Internet infrastructure will give rise to wider interoperability than  
if each application has to build the features they need.


So, how to proceed? How can the IETF process help? We could determine,  
for each bucket that there is interest in, how to achieve it a)  
without changing HTTP (the fall-back), and b) using newly defined  
protocol (via either wholesale new protocol, or HTTP extension). This  
would provide guidance to software authors for best practices, lead to  
interoperable server and client components, ensure applications work  
well with infrastructure, and enable infrastructure to better support  
these applications.

I'm deeply interesting in supporting applications that want Message,  
Req/Resp. semantics. I'd like to help with applications that have  
needs in other buckets. I actually think we could do this with just  
one protocol (and fall-back methodology), and would love to try to  
achieve that.

	- Mark

Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com


From markl@lindenlab.com  Wed Apr  8 17:22:17 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 43AC73A6B9E for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:22:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.264
X-Spam-Level: 
X-Spam-Status: No, score=-3.264 tagged_above=-999 required=5 tests=[AWL=-0.266, BAYES_00=-2.599, HTML_MESSAGE=0.001, J_CHICKENPOX_64=0.6, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id befwmfqyFbRm for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:22:16 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 48DC03A6AD4 for <hybi@ietf.org>; Wed,  8 Apr 2009 17:22:16 -0700 (PDT)
Received: from 945battery-guestB-139.lindenlab.com (945battery-guestB-139.lindenlab.com [10.1.11.139]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id E9C063DBC073 for <hybi@ietf.org>; Wed,  8 Apr 2009 17:23:23 -0700 (PDT)
Message-Id: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
Content-Type: multipart/alternative; boundary=Apple-Mail-14--1063601149
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Wed, 8 Apr 2009 17:23:23 -0700
X-Mailer: Apple Mail (2.930.3)
Subject: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 00:22:17 -0000

--Apple-Mail-14--1063601149
Content-Type: text/plain;
	charset=US-ASCII;
	format=flowed;
	delsp=yes
Content-Transfer-Encoding: 7bit

Sorry about the hodge-podge here: After IETF-74 I had a zillion things  
to do at work, so it wasn't until this morning that I was able to  
fully catch up on my backlog of reading hybi mail. Here are a number  
of bit-and-bobs that I wanted to reply to:

> the whole point of bidirectional communication is that it's not a  
> request-response model.
Actually, I don't agree. At Linden Lab, we've seen a great need for  
bidirectional communication that is request/response model.

> IMHO, is that it be possible to implement the server-side in a few  
> dozen lines of code, and I don't see any way to implement a  
> compliant HTTP server and client in such a small code footprint.
We've successfully implemented rHTTP server and client components in  
just a few dozen lines of code because we are able to just reuse the  
existing HTTP client and server components that exist in the  
environments we work in (C++, Python).

> The rHTTP proposal is more attractive as a wire protocl as it uses  
> normal HTTP framing and should be trivial to implement with existing  
> code bases. However, rHTTP would not significantly change bayuex  
> long polling other than to make long-poll responses into real  
> requests and long-poll requests into real responses. The same data  
> framing and latency would go over the wire. The semantic that is  
> missing from HTTP that would improve Bayeux/cometd is the ability to  
> send multiple responses to a single request. If for example, a 1xx  
> response could be sent with headers and content without terminating  
> the current request, then bayeux long-polling would avoid the need  
> to issue another long-polling request. This would reduce data  
> overhead and avoid protocol induced worst case latency.

I think this statement is worthy of greater study!  While I don't like  
overloading the 1xx responses... It seems to me that this is speaking  
to the crux of the matter:Many of us like HTTP framing, but want only  
one TCP connection, and asynchrony of request and response, in both  
directions. I feel a draft in here....


> What would you consider normal HTTP style framing here? What would  
> HTTP framing give you that isn't available with the framing  
> currently proposed with Web Sockets?

Addressing (URLs), Headers, Content specification, Transfer  
specification, Status

	 - Mark


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






--Apple-Mail-14--1063601149
Content-Type: text/html;
	charset=US-ASCII
Content-Transfer-Encoding: quoted-printable

<html><body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; =
-webkit-line-break: after-white-space; "><div><pre><pre><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal; ">Sorry about the hodge-podge here: After IETF-74 I had a =
zillion things to do at work, so it wasn't until this morning that I was =
able to fully catch up on my backlog of reading hybi mail. Here are a =
number of bit-and-bobs that I wanted to reply =
to:</span></font></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: =
normal;"><br></span></font></pre></pre><pre><blockquote =
type=3D"cite"><pre><font class=3D"Apple-style-span" face=3D"Helvetica" =
size=3D"3"><span class=3D"Apple-style-span" style=3D"font-size: 12px; =
white-space: normal;">the whole point of=20
bidirectional communication is that it's not a request-response =
model.</span></font></pre></blockquote><pre><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal;">Actually, I don't agree. At Linden Lab, we've seen a great need =
for bidirectional communication that is request/response =
model.</span></font></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: =
normal;"><br></span></font></pre><pre></pre><blockquote =
type=3D"cite"><pre><span class=3D"Apple-style-span" style=3D"font-family: =
Helvetica; font-size: 12px; white-space: normal; ">IMHO, is that it be =
possible=20
to implement the server-side in a few dozen lines of code, and I don't =
see=20
any way to implement a compliant HTTP server and client in such a small=20=

code footprint.</span></pre></blockquote><pre><pre><pre><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal;">We've successfully implemented rHTTP server and client =
components in just a few dozen lines of code because we are able to just =
reuse the existing HTTP client and server components that exist in the =
environments we work in (C++, Python).</span></font></pre><pre><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal;"><br></span></font></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: normal;"><span =
class=3D"Apple-style-span" style=3D"font-family: -webkit-monospace; =
font-size: 10px; white-space: pre; =
"><pre><pre><pre></pre><pre><blockquote type=3D"cite"><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal; ">The rHTTP proposal is more attractive as a wire protocl as it =
uses normal HTTP framing and should be trivial to implement with =
existing code bases. However, rHTTP would not significantly change =
bayuex long polling other than to make long-poll responses into real =
requests and long-poll requests into real responses. The same data =
framing and latency would go over the wire. The semantic that is missing =
from HTTP that would improve Bayeux/cometd is the ability to send =
multiple responses to a single request. If for example, a 1xx response =
could be sent with headers and content without terminating the current =
request, then bayeux long-polling would avoid the need to issue another =
long-polling request. This would reduce data overhead and avoid protocol =
induced worst case latency.</span></font></blockquote><br><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal; "><span class=3D"Apple-style-span" style=3D"color: rgb(0, 0, 0); =
">I think this statement is worthy of greater study! &nbsp;While I don't =
like overloading the 1xx responses... It seems to me that this is =
speaking to the crux of the matter:Many of us like HTTP framing, but =
want only one TCP connection, and asynchrony of request and response, in =
both directions. I feel a draft in =
here....&nbsp;</span></span></font></pre><pre><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal; "><span class=3D"Apple-style-span" style=3D"color: rgb(0, 0, 0); =
"><br></span></span></font><br><blockquote type=3D"cite"><font =
class=3D"Apple-style-span" face=3D"Helvetica" size=3D"3"><span =
class=3D"Apple-style-span" style=3D"font-size: 12px; white-space: =
normal; "><span class=3D"Apple-style-span" style=3D"color: rgb(0, 0, 0); =
">What would you consider normal HTTP style framing here? What would =
HTTP=20
framing give you that isn't available with the framing currently =
proposed=20
with Web =
Sockets?</span></span></font></blockquote></pre></pre></pre></span></span>=
</font></pre></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: normal;">Addressing (URLs), =
Headers, Content specification, Transfer specification, =
Status</span></font></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: =
normal;"><br></span></font></pre><pre><font class=3D"Apple-style-span" =
face=3D"Helvetica" size=3D"3"><span class=3D"Apple-style-span" =
style=3D"font-size: 12px; white-space: normal;"><span =
class=3D"Apple-tab-span" style=3D"white-space:pre">	</span>&nbsp;- =
Mark<br></span></font></pre></pre></pre></div><br><br><div =
apple-content-edited=3D"true"> <span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; color: rgb(0, 0, 0); font-family: =
Helvetica; font-size: 12px; font-style: normal; font-variant: normal; =
font-weight: normal; letter-spacing: normal; line-height: normal; =
orphans: 2; text-align: auto; text-indent: 0px; text-transform: none; =
white-space: normal; widows: 2; word-spacing: 0px; =
-webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: =
0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: =
auto; -webkit-text-stroke-width: 0; "><div style=3D"word-wrap: =
break-word; -webkit-nbsp-mode: space; -webkit-line-break: =
after-white-space; "><span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; color: rgb(0, 0, 0); font-family: =
Helvetica; font-size: 12px; font-style: normal; font-variant: normal; =
font-weight: normal; letter-spacing: normal; line-height: normal; =
orphans: 2; text-indent: 0px; text-transform: none; white-space: normal; =
widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; =
-webkit-border-vertical-spacing: 0px; =
-webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: =
auto; -webkit-text-stroke-width: 0px; "><div style=3D"word-wrap: =
break-word; -webkit-nbsp-mode: space; -webkit-line-break: =
after-white-space; "><span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; -webkit-border-horizontal-spacing: =
0px; -webkit-border-vertical-spacing: 0px; color: rgb(0, 0, 0); =
font-family: Helvetica; font-size: 12px; font-style: normal; =
font-variant: normal; font-weight: normal; letter-spacing: normal; =
line-height: normal; -webkit-text-decorations-in-effect: none; =
text-indent: 0px; -webkit-text-size-adjust: auto; text-transform: none; =
orphans: 2; white-space: normal; widows: 2; word-spacing: 0px; "><div =
style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; =
-webkit-line-break: after-white-space; "><div>Mark =
Lentczner</div><div>Sr. Systems Architect</div><div>Technology =
Integration</div><div>Linden Lab</div><div><br =
class=3D"khtml-block-placeholder"></div><div><a =
href=3D"mailto:markl@lindenlab.com">markl@lindenlab.com</a></div><div><br =
class=3D"khtml-block-placeholder"></div><div>Zero Linden</div><div><a =
href=3D"mailto:zero.linden@secondlife.com">zero.linden@secondlife.com</a><=
/div><div><br class=3D"khtml-block-placeholder"></div></div><br =
class=3D"Apple-interchange-newline"></span></div></span><br =
class=3D"Apple-interchange-newline"></div></span><br =
class=3D"Apple-interchange-newline"> </div><br></body></html>=

--Apple-Mail-14--1063601149--

From ian@hixie.ch  Wed Apr  8 17:33:18 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7CA0C3A6BC1 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:33:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.551
X-Spam-Level: 
X-Spam-Status: No, score=-3.551 tagged_above=-999 required=5 tests=[AWL=-0.952, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id H2jE8oIim0My for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 17:33:17 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 88A983A6BBE for <hybi@ietf.org>; Wed,  8 Apr 2009 17:33:17 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 4B11E121EFC; Wed,  8 Apr 2009 17:34:25 -0700 (PDT)
Date: Thu, 9 Apr 2009 00:34:22 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Mark Lentczner <markl@lindenlab.com>
In-Reply-To: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>
Message-ID: <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 00:33:18 -0000

On Wed, 8 Apr 2009, Mark Lentczner wrote:
> >
> > the whole point of bidirectional communication is that it's not a 
> > request-response model.
>
> Actually, I don't agree. At Linden Lab, we've seen a great need for 
> bidirectional communication that is request/response model.

I misspoke, I should have said that the point of full-duplex communication 
is that it's not request-response -- either side can spontaneously send a 
message.


> > IMHO, is that it be possible to implement the server-side in a few 
> > dozen lines of code, and I don't see any way to implement a compliant 
> > HTTP server and client in such a small code footprint.
>
> We've successfully implemented rHTTP server and client components in 
> just a few dozen lines of code because we are able to just reuse the 
> existing HTTP client and server components that exist in the 
> environments we work in (C++, Python).

Sure, if you can use libraries then implementing any protocol is trivial. 
The point is that the protocol should be simple enough that one doesn't 
need to rely on libraries to get a compliant implementation.


> > What would you consider normal HTTP style framing here? What would 
> > HTTP framing give you that isn't available with the framing currently 
> > proposed with Web Sockets?
> 
> Addressing (URLs), Headers, Content specification, Transfer 
> specification, Status

The socket in WebSocket can be (indeed, must be) addressed via URL.

At the message level, why do we need to force those features on all users? 
They seem like features we should allow people to have if they want, but 
they don't seem like features we should require all clients and servers to 
implement just because someone might want them.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From duerst@it.aoyama.ac.jp  Wed Apr  8 20:18:54 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 41D103A68CC for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 20:18:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.532
X-Spam-Level: 
X-Spam-Status: No, score=0.532 tagged_above=-999 required=5 tests=[AWL=-1.167,  BAYES_05=-1.11, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bE+6jDn+MYQB for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 20:18:53 -0700 (PDT)
Received: from scmailgw2.scop.aoyama.ac.jp (scmailgw2.scop.aoyama.ac.jp [133.2.251.195]) by core3.amsl.com (Postfix) with ESMTP id 36E673A6781 for <hybi@ietf.org>; Wed,  8 Apr 2009 20:18:51 -0700 (PDT)
Received: from scmse2.scbb.aoyama.ac.jp (scmse2 [133.2.253.17]) by scmailgw2.scop.aoyama.ac.jp (secret/secret) with SMTP id n393Jw4O000214 for <hybi@ietf.org>; Thu, 9 Apr 2009 12:19:58 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse2.scbb.aoyama.ac.jp with smtp id 5bd9_4df3ce2c_24b5_11de_a31a_0019b9e2b3d9; Thu, 09 Apr 2009 12:19:58 +0900
Received: from [IPv6:::1] ([133.2.210.1]:56137) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC905B9> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Thu, 9 Apr 2009 12:19:03 +0900
Message-ID: <49DD6947.30402@it.aoyama.ac.jp>
Date: Thu, 09 Apr 2009 12:19:35 +0900
From: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49DC031D.9090007@it.aoyama.ac.jp> <Pine.LNX.4.62.0904080523311.18470@hixie.dreamhostps.com> <1cb725390904072240o340230a5p2f772b35df910d9b@mail.gmail.com> <Pine.LNX.4.62.0904080545200.18470@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904080545200.18470@hixie.dreamhostps.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Receiving side browser API for WebSockets?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 03:18:54 -0000

Hello Ian,

On 2009/04/08 14:46, Ian Hickson wrote:

>> I believe that Martin was asking about the details onmessage event
>> handler and the MessageEvent interface.
>
> Oh. This is defined in the same place as the sending API:
>
>     http://dev.w3.org/html5/websockets/

I see. That seems reasonable. I was wondering about it mostly because of 
the comparison with TCP. For TCP, as far as I understand, the idea is 
that packets are internal to TCP, that applications do not know and do 
not care whether some data came all in one packet or was split up.

For WebSockets, that's different, each datagram is handled separately to 
the client. That makes parsing on the receiver side reasonably easy.

> The MessageEvent interface is defined in HTML5:
>
>     http://www.whatwg.org/specs/web-apps/current-work/#messageevent

[That crashed my browser (Opera 9.64 on Vista) several times yesterday, 
but today, I was able to look at it.]

Regards,    Martin.

-- 
#-# Martin J. DÃ¼rst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From duerst@it.aoyama.ac.jp  Wed Apr  8 20:28:13 2009
Return-Path: <duerst@it.aoyama.ac.jp>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 34D2F3A696F for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 20:28:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.571
X-Spam-Level: 
X-Spam-Status: No, score=0.571 tagged_above=-999 required=5 tests=[AWL=-1.128,  BAYES_05=-1.11, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 6Vx-84qz+0Up for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 20:28:12 -0700 (PDT)
Received: from scmailgw2.scop.aoyama.ac.jp (scmailgw2.scop.aoyama.ac.jp [133.2.251.195]) by core3.amsl.com (Postfix) with ESMTP id CA2923A68B1 for <hybi@ietf.org>; Wed,  8 Apr 2009 20:28:11 -0700 (PDT)
Received: from scmse3.scbb.aoyama.ac.jp ([133.2.253.23]) by scmailgw2.scop.aoyama.ac.jp (secret/secret) with SMTP id n393TI7Z004136 for <hybi@ietf.org>; Thu, 9 Apr 2009 12:29:18 +0900 (JST)
Received: from (unknown [133.2.206.133]) by scmse3.scbb.aoyama.ac.jp with smtp id 5d7d_9bc375b6_24b6_11de_a326_001d0969ab06; Thu, 09 Apr 2009 12:29:18 +0900
Received: from [IPv6:::1] ([133.2.210.1]:44191) by itmail.it.aoyama.ac.jp with [XMail 1.22 ESMTP Server] id <SC906A6> for <hybi@ietf.org> from <duerst@it.aoyama.ac.jp>; Thu, 9 Apr 2009 12:28:24 +0900
Message-ID: <49DD6B77.50104@it.aoyama.ac.jp>
Date: Thu, 09 Apr 2009 12:28:55 +0900
From: =?UTF-8?B?Ik1hcnRpbiBKLiBEw7xyc3Qi?= <duerst@it.aoyama.ac.jp>
Organization: Aoyama Gakuin University
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b3pre) Gecko/20090108 Eudora/3.0b1pre
MIME-Version: 1.0
To: Paul Prescod <paul@prescod.net>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com>
In-Reply-To: <1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 03:28:13 -0000

On 2009/04/09 5:02, Paul Prescod wrote:

> I would add a "data synchronization" data model. I find that most
> often the goal is to have the web browser report the state of a
> resource or resource collection to a user, and then inform the user of
> changes to it.

That is indeed a very good point. Most kinds of collaborative activity 
can easily be looked at that way. In its full incarnation, maybe it 
would just allow an application programmer to say: synchronize this part 
of this page (i.e. DOM) with this part of that page over at that client.

Regards,    Martin.

-- 
#-# Martin J. DÃ¼rst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

From gregw@webtide.com  Wed Apr  8 22:30:39 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 95F483A6B18 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 22:30:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.488
X-Spam-Level: 
X-Spam-Status: No, score=-2.488 tagged_above=-999 required=5 tests=[AWL=0.111,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id TaLtRH+32-0X for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 22:30:39 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.183]) by core3.amsl.com (Postfix) with ESMTP id EC31A3A6878 for <hybi@ietf.org>; Wed,  8 Apr 2009 22:30:38 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so231790waf.5 for <hybi@ietf.org>; Wed, 08 Apr 2009 22:31:46 -0700 (PDT)
Received: by 10.114.133.1 with SMTP id g1mr1219002wad.21.1239255106551; Wed, 08 Apr 2009 22:31:46 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id z20sm16114465pod.21.2009.04.08.22.31.44 (version=TLSv1/SSLv3 cipher=RC4-MD5); Wed, 08 Apr 2009 22:31:45 -0700 (PDT)
Message-ID: <49DD883C.2050004@webtide.com>
Date: Thu, 09 Apr 2009 15:31:40 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com> <49DD1397.9030903@webtide.com> <Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 05:30:39 -0000

Ian Hickson wrote:

> These seem to all boil down to extensions or changes to HTTP. It seems 
> that those should probably be discussed in the ietf-http-wg mailing list 
> as part of HTTP working group discussions.

I was hoping that this list could have a discussion about
the general requirements and possible solutions of bidirectional
HTTP.   Then if some consensus views were developed, these could
be taken to the appropriate forums: ietf-http-wg etc.

But then I'm not familiar with ietf processes.  Should I be taking the
concerns I've raised here to the http wg list?


> The browser vendors are present here on this list, I don't think there's 
> any need to write "demands" to them.

"demands" was not a good choice of words.
The point was that the OAA prepared a qualified/vetted list of issues
as a consensus document to take to the browser vendors.


> I would be interested in taking a specification for a "minimal, event" 
> protocol for implementation directly by Web browsers to the standards 
> track. This seems orthogonal to the above work. What would be the best way 
> to move forward with such a goal?

How would this protocol be different from other existing protocols
such as BEEP?

cheers







From ian@hixie.ch  Wed Apr  8 23:23:31 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C8A383A6B10 for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 23:23:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.527
X-Spam-Level: 
X-Spam-Status: No, score=-3.527 tagged_above=-999 required=5 tests=[AWL=-0.928, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bdjQ4s1KsdZG for <hybi@core3.amsl.com>; Wed,  8 Apr 2009 23:23:31 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id F0E293A680E for <hybi@ietf.org>; Wed,  8 Apr 2009 23:23:30 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 921E7122510; Wed,  8 Apr 2009 23:24:38 -0700 (PDT)
Date: Thu, 9 Apr 2009 06:24:36 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49DD883C.2050004@webtide.com>
Message-ID: <Pine.LNX.4.62.0904090542130.19453@hixie.dreamhostps.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com> <49DD1397.9030903@webtide.com> <Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com> <49DD883C.2050004@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 06:23:31 -0000

On Thu, 9 Apr 2009, Greg Wilkins wrote:
> > 
> > I would be interested in taking a specification for a "minimal, event" 
> > protocol for implementation directly by Web browsers to the standards 
> > track. This seems orthogonal to the above work. What would be the best 
> > way to move forward with such a goal?
> 
> How would this protocol be different from other existing protocols such 
> as BEEP?

BEEP isn't minimal.

(Also, BEEP would have to be extended to support appropriate security 
checks for the Web's same-origin model.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From stpeter@stpeter.im  Thu Apr  9 07:13:27 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A55753A6BC9 for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 07:13:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.52
X-Spam-Level: 
X-Spam-Status: No, score=-2.52 tagged_above=-999 required=5 tests=[AWL=0.079,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QafsK07pPCiR for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 07:13:26 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id B76D83A682A for <hybi@ietf.org>; Thu,  9 Apr 2009 07:13:26 -0700 (PDT)
Received: from squire.local (ip-65-183-77-201.rev.frii.com [65.183.77.201]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id D1947E800D; Thu,  9 Apr 2009 08:14:23 -0600 (MDT)
Message-ID: <49DE02BE.5080908@stpeter.im>
Date: Thu, 09 Apr 2009 08:14:22 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>	<Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com>	<49DD1397.9030903@webtide.com>	<Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com>	<49DD883C.2050004@webtide.com> <Pine.LNX.4.62.0904090542130.19453@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904090542130.19453@hixie.dreamhostps.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms060809000902060601020603"
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 14:13:27 -0000

This is a cryptographically signed message in MIME format.

--------------ms060809000902060601020603
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/9/09 12:24 AM, Ian Hickson wrote:
> On Thu, 9 Apr 2009, Greg Wilkins wrote:
>>> I would be interested in taking a specification for a "minimal, event" 
>>> protocol for implementation directly by Web browsers to the standards 
>>> track. This seems orthogonal to the above work. What would be the best 
>>> way to move forward with such a goal?
>> How would this protocol be different from other existing protocols such 
>> as BEEP?
> 
> BEEP isn't minimal.
> 
> (Also, BEEP would have to be extended to support appropriate security 
> checks for the Web's same-origin model.)

The same-origin model is often invoked here. Is there a good definition
of it? Perhaps it would be helpful for someone to write a small I-D
about this.

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms060809000902060601020603
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms060809000902060601020603--

From stpeter@stpeter.im  Thu Apr  9 08:01:50 2009
Return-Path: <stpeter@stpeter.im>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D20613A68BE for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 08:01:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.523
X-Spam-Level: 
X-Spam-Status: No, score=-2.523 tagged_above=-999 required=5 tests=[AWL=0.076,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id p1DNDdt5VCYu for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 08:01:50 -0700 (PDT)
Received: from dizzyd.com (dizzyd.com [207.210.219.225]) by core3.amsl.com (Postfix) with ESMTP id 5B79D3A682A for <hybi@ietf.org>; Thu,  9 Apr 2009 08:01:24 -0700 (PDT)
Received: from wrk165.corp.jabber.com (dencfw1.jabber.com [207.182.164.5]) (Authenticated sender: stpeter) by dizzyd.com (Postfix) with ESMTPSA id E2B19E803E; Thu,  9 Apr 2009 09:02:31 -0600 (MDT)
Message-ID: <49DE0A0B.1040603@stpeter.im>
Date: Thu, 09 Apr 2009 08:45:31 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Thunderbird 2.0.0.21 (Macintosh/20090302)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
X-Enigmail-Version: 0.95.7
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=sha1; boundary="------------ms070309060902020805060304"
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 15:01:50 -0000

This is a cryptographically signed message in MIME format.

--------------ms070309060902020805060304
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 4/8/09 6:34 PM, Ian Hickson wrote:
> On Wed, 8 Apr 2009, Mark Lentczner wrote:
>>
>>> IMHO, is that it be possible to implement the server-side in a few 
>>> dozen lines of code, and I don't see any way to implement a compliant 
>>> HTTP server and client in such a small code footprint.
>> We've successfully implemented rHTTP server and client components in 
>> just a few dozen lines of code because we are able to just reuse the 
>> existing HTTP client and server components that exist in the 
>> environments we work in (C++, Python).
> 
> Sure, if you can use libraries then implementing any protocol is trivial. 
> The point is that the protocol should be simple enough that one doesn't 
> need to rely on libraries to get a compliant implementation.

Where does this requirement come from? I have never seen a similar
requirement in any RFC.

And are you assuming that developers should be able to implement this
protocol without any libraries whatsoever? Not even for TCP, UTF-8, or
anything else?

Peter

-- 
Peter Saint-Andre
https://stpeter.im/


--------------ms070309060902020805060304
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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--------------ms070309060902020805060304--


From markl@lindenlab.com  Thu Apr  9 09:05:17 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 32CA73A6CA7 for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 09:05:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.546
X-Spam-Level: 
X-Spam-Status: No, score=-3.546 tagged_above=-999 required=5 tests=[AWL=0.053,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PpiSf5fvXtwK for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 09:05:16 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 428CF3A683A for <hybi@ietf.org>; Thu,  9 Apr 2009 09:05:16 -0700 (PDT)
Received: from nil.lindenlab.com (nil.lindenlab.com [10.1.16.4]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 321443DBC09A for <hybi@ietf.org>; Thu,  9 Apr 2009 09:06:24 -0700 (PDT)
Message-Id: <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Thu, 9 Apr 2009 09:06:23 -0700
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 16:05:17 -0000

On Apr 8, 2009, at 5:34 PM, Ian Hickson wrote:
> I misspoke, I should have said that the point of full-duplex  
> communication
> is that it's not request-response -- either side can spontaneously  
> send a
> message.

Then I mis-replied. I should have said that we have a need for full- 
duplex request-response -- either side can spontaneously send a  
request message, and either side can, when ready, reply to a prior  
request.

> Sure, if you can use libraries then implementing any protocol is  
> trivial.
> The point is that the protocol should be simple enough that one  
> doesn't
> need to rely on libraries to get a compliant implementation.

I'm not sure you understood my point: We weren't using rHTTP  
libraries, but extant HTTP libraries. You had made prior comments  
about the ease of implementation, which as a practical consideration,  
I agree is relevant. There was indication that you thought other  
proposals in this space required complex implementations. I was  
pointing out that our rHTTP experience was quite the opposite: Just a  
few lines of code to get all the features of request/response  
messages. That we were able to reuse ubiquitous HTTP libraries I think  
is precisely to the point.

> The socket in WebSocket can be (indeed, must be) addressed via URL.

Only the establishment in WebSocket is addressed via URL.  In rHTTP  
each request can be addressed by URL.  This is useful as it is a very  
common piece of meta-data that applications want. For example, using  
WebSockets, for any complex page, an application is almost certainly  
going to want a way to address which widget on the page is the  
intended recipient of the event. Hence, each application is going to  
have to devise an addressing scheme, and more importantly, a method of  
encoding that address in the simple string event.

> At the message level, why do we need to force those features on all  
> users?
> They seem like features we should allow people to have if they want,  
> but
> they don't seem like features we should require all clients and  
> servers to
> implement just because someone might want them.

Of course, but you must recognize there is a trade-off here that all  
protocols make. The design needs to choose a feature set wide enough  
that many different applications can make use of it, but must do so  
balancing complexity and overhead. For example, WebSockets has chosen  
to support both binary and UTF-8 payloads. The overhead of the lead-in  
type byte was seen as worth it, even though there are many  
applications that will only use the UTF-8 strings. So, I see the  
choices here as just a matter of degree. I think popularity of HTTP  
for many kinds of applications has shown that one or more features of  
HTTP(*) are often applicable to an application's needs, while the  
overhead of these features is low enough to be worth including them in  
the protocol for all.(**)

	- Mark

(*) Addressing (URLs), Method, Headers, Content specification,  
Transfer specification, Status
(**) I imagine the wire overhead for putting a minimal string event  
into an HTTP message would be something like:
	POST / HTTP/1.1
	Host:
	Content-Type:text/plain;charset=UTF-8
	Content-Length:xxx

and
	204 ACK

That's 96 bytes overhead for a 100 to 999 byte payload. Seems high,  
but for the usage cases that web pages present, this probably is not  
relevant.  Further, once one has this framing, it is very likely that  
rather than build addressing into the string payload, applications  
would just make use of the URL field. Plus, this framing already has  
character encoding support.

The overhead in code is certainly more than parsing a single sentinel  
delimited string. However, the ability to parse an HTTP message is  
well understood, easily implemented, and nearly ubiquitously available.


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From David.Linner9ab33xy531fokus.fraunhofer.de@bounce.antispameurope.de  Thu Apr  9 08:07:42 2009
Return-Path: <David.Linner9ab33xy531fokus.fraunhofer.de@bounce.antispameurope.de>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 830DD3A6A19 for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 08:07:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.599
X-Spam-Level: 
X-Spam-Status: No, score=-3.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bFkBxm4fTjnv for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 08:07:41 -0700 (PDT)
Received: from relay03-haj2.antispameurope.com (relay03-haj2.antispameurope.com [83.246.65.53]) by core3.amsl.com (Postfix) with ESMTP id C3F053A6AE8 for <hybi@ietf.org>; Thu,  9 Apr 2009 08:07:40 -0700 (PDT)
Received: by relay03-haj2.antispameurope.com (ASE-Secure-MTA, from userid 1000) id 24A4D63C087; Thu,  9 Apr 2009 17:08:46 +0200 (CEST)
Received: from pluto.fokus.fraunhofer.de (pluto.fokus.fraunhofer.de [195.37.77.164]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by relay03-haj2.antispameurope.com (ASE-Secure-MTA) with ESMTP id 8DBC463C087; Thu,  9 Apr 2009 17:08:45 +0200 (CEST)
Received: from EXCHSRV.fokus.fraunhofer.de (bohr [10.147.9.231]) by pluto.fokus.fraunhofer.de (8.13.7/8.13.7) with SMTP id n39F8iwA011423; Thu, 9 Apr 2009 17:08:45 +0200 (MEST)
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Thu, 9 Apr 2009 17:08:35 +0200
Message-ID: <79C0EA6E7AD7CE4A85EDAF482B5456B202173B59@EXCHSRV.fokus.fraunhofer.de>
In-Reply-To: <49DD6B77.50104@it.aoyama.ac.jp>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] Apples and Orangutans
Thread-Index: Acm4w2K2i3y9fZ7ORDqonljoUiR7yAAXg1yA
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com><1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com> <49DD6B77.50104@it.aoyama.ac.jp>
From: "Linner, David" <David.Linner@fokus.fraunhofer.de>
To: <hybi@ietf.org>
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2009 16:19:02 -0000
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From gregw@webtide.com  Thu Apr  9 18:05:54 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B41FB3A6C9E for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 18:05:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.495
X-Spam-Level: 
X-Spam-Status: No, score=-2.495 tagged_above=-999 required=5 tests=[AWL=0.104,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MP0e9LqXGPRA for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 18:05:53 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id 249683A6AE2 for <hybi@ietf.org>; Thu,  9 Apr 2009 18:05:52 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so599626tid.25 for <hybi@ietf.org>; Thu, 09 Apr 2009 18:07:00 -0700 (PDT)
Received: by 10.110.42.17 with SMTP id p17mr3033659tip.51.1239325620381; Thu, 09 Apr 2009 18:07:00 -0700 (PDT)
Received: from ?10.10.1.9? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id b7sm3227573tic.15.2009.04.09.18.06.57 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 09 Apr 2009 18:06:58 -0700 (PDT)
Message-ID: <49DE9BAC.5050905@webtide.com>
Date: Fri, 10 Apr 2009 11:06:52 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <Pine.LNX.4.62.0904081849380.19453@hixie.dreamhostps.com> <49DD1397.9030903@webtide.com> <Pine.LNX.4.62.0904082142580.19453@hixie.dreamhostps.com> <49DD883C.2050004@webtide.com> <Pine.LNX.4.62.0904090542130.19453@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904090542130.19453@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 01:05:54 -0000

Ian Hickson wrote:
> On Thu, 9 Apr 2009, Greg Wilkins wrote:
>>> I would be interested in taking a specification for a "minimal, event" 
>>> protocol for implementation directly by Web browsers to the standards 
>>> track. This seems orthogonal to the above work. What would be the best 
>>> way to move forward with such a goal?
>> How would this protocol be different from other existing protocols such 
>> as BEEP?
> 
> BEEP isn't minimal.

BEEPs framing is pretty minimal, yet full featured with respect optional
headers, chunking, multi response etc etc.

But I do agree that the channel negotiation stuff is a bit non-minimal

However, IF we are in the game of picking winners or designing new protocols,
we could do a lot worse than starting with a close inspection of BEEP and
the concerns that they address.

Maybe a BEEP-lite protocol would be a good upgrade target and one
capable of carrying arbitrary other protocols through firewalls
and gateways?



> (Also, BEEP would have to be extended to support appropriate security 
> checks for the Web's same-origin model.)

If we were using the upgrade mechanism, then I believe that most
of these concerns are addressed.  The standard HTTP security model
would be used to establish a connection and then the upgrade is
unable to break out of that connection etc.

This applies for websocket, rHTTP or any other upgrade protocol.
If there are security issues that exist even within such an
upgrade model, then it would be good task for this group
to identify them as a measure for any candidate upgrade protocols.


cheers

From mike@dierken.com  Thu Apr  9 21:22:37 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D60093A65A5 for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 21:22:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ji5qHgxRJdM9 for <hybi@core3.amsl.com>; Thu,  9 Apr 2009 21:22:37 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.228]) by core3.amsl.com (Postfix) with ESMTP id E994B3A67D7 for <hybi@ietf.org>; Thu,  9 Apr 2009 21:22:36 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so755397rvb.49 for <hybi@ietf.org>; Thu, 09 Apr 2009 21:23:45 -0700 (PDT)
Received: by 10.141.162.1 with SMTP id p1mr1327843rvo.277.1239337424975; Thu, 09 Apr 2009 21:23:44 -0700 (PDT)
Received: from ?192.168.1.5? (c-76-121-195-84.hsd1.wa.comcast.net [76.121.195.84]) by mx.google.com with ESMTPS id k41sm2756502rvb.56.2009.04.09.21.23.42 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 09 Apr 2009 21:23:44 -0700 (PDT)
From: Mike Dierken <mike@dierken.com>
To: Mark Lentczner <markl@lindenlab.com>
In-Reply-To: <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
X-Mailer: iPhone Mail (5H11)
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
Message-Id: <C848C205-339E-4BCC-A52C-AD05C7A36C13@dierken.com>
Content-Type: text/plain; charset=us-ascii; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (iPhone Mail 5H11)
Date: Thu, 9 Apr 2009 21:23:33 -0700
Cc: "hybi@ietf.org" <hybi@ietf.org>
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 04:22:37 -0000

Using HTTP framing for these request/response messages is an  
interesting idea and something I like a lot, especially when the  
request receiver is on the user's machine.


On Apr 9, 2009, at 9:06 AM, Mark Lentczner <markl@lindenlab.com> wrote:

>
> On Apr 8, 2009, at 5:34 PM, Ian Hickson wrote:
>> I misspoke, I should have said that the point of full-duplex  
>> communication
>> is that it's not request-response -- either side can spontaneously  
>> send a
>> message.
>
> Then I mis-replied. I should have said that we have a need for full- 
> duplex request-response -- either side can spontaneously send a  
> request message, and either side can, when ready, reply to a prior  
> request.
>
>> Sure, if you can use libraries then implementing any protocol is  
>> trivial.
>> The point is that the protocol should be simple enough that one  
>> doesn't
>> need to rely on libraries to get a compliant implementation.
>
> I'm not sure you understood my point: We weren't using rHTTP  
> libraries, but extant HTTP libraries. You had made prior comments  
> about the ease of implementation, which as a practical  
> consideration, I agree is relevant. There was indication that you  
> thought other proposals in this space required complex  
> implementations. I was pointing out that our rHTTP experience was  
> quite the opposite: Just a few lines of code to get all the features  
> of request/response messages. That we were able to reuse ubiquitous  
> HTTP libraries I think is precisely to the point.
>
>> The socket in WebSocket can be (indeed, must be) addressed via URL.
>
> Only the establishment in WebSocket is addressed via URL.  In rHTTP  
> each request can be addressed by URL.  This is useful as it is a  
> very common piece of meta-data that applications want. For example,  
> using WebSockets, for any complex page, an application is almost  
> certainly going to want a way to address which widget on the page is  
> the intended recipient of the event. Hence, each application is  
> going to have to devise an addressing scheme, and more importantly,  
> a method of encoding that address in the simple string event.
>
>> At the message level, why do we need to force those features on all  
>> users?
>> They seem like features we should allow people to have if they  
>> want, but
>> they don't seem like features we should require all clients and  
>> servers to
>> implement just because someone might want them.
>
> Of course, but you must recognize there is a trade-off here that all  
> protocols make. The design needs to choose a feature set wide enough  
> that many different applications can make use of it, but must do so  
> balancing complexity and overhead. For example, WebSockets has  
> chosen to support both binary and UTF-8 payloads. The overhead of  
> the lead-in type byte was seen as worth it, even though there are  
> many applications that will only use the UTF-8 strings. So, I see  
> the choices here as just a matter of degree. I think popularity of  
> HTTP for many kinds of applications has shown that one or more  
> features of HTTP(*) are often applicable to an application's needs,  
> while the overhead of these features is low enough to be worth  
> including them in the protocol for all.(**)
>
>    - Mark
>
> (*) Addressing (URLs), Method, Headers, Content specification,  
> Transfer specification, Status
> (**) I imagine the wire overhead for putting a minimal string event  
> into an HTTP message would be something like:
>    POST / HTTP/1.1
>    Host:
>    Content-Type:text/plain;charset=UTF-8
>    Content-Length:xxx
>
> and
>    204 ACK
>
> That's 96 bytes overhead for a 100 to 999 byte payload. Seems high,  
> but for the usage cases that web pages present, this probably is not  
> relevant.  Further, once one has this framing, it is very likely  
> that rather than build addressing into the string payload,  
> applications would just make use of the URL field. Plus, this  
> framing already has character encoding support.
>
> The overhead in code is certainly more than parsing a single  
> sentinel delimited string. However, the ability to parse an HTTP  
> message is well understood, easily implemented, and nearly  
> ubiquitously available.
>
>
> Mark Lentczner
> Sr. Systems Architect
> Technology Integration
> Linden Lab
>
> markl@lindenlab.com
>
> Zero Linden
> zero.linden@secondlife.com
>
>
>
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi

From rsayre@mozilla.com  Fri Apr 10 00:11:43 2009
Return-Path: <rsayre@mozilla.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 282EF3A6D43 for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 00:11:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.74
X-Spam-Level: 
X-Spam-Status: No, score=-0.74 tagged_above=-999 required=5 tests=[BAYES_20=-0.74]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XHYE9wk7grms for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 00:11:42 -0700 (PDT)
Received: from dm-mail01.mozilla.org (dm-mail01.mozilla.org [63.245.208.150]) by core3.amsl.com (Postfix) with ESMTP id 8C24B3A6D0F for <hybi@ietf.org>; Fri, 10 Apr 2009 00:11:42 -0700 (PDT)
Received: from robert-sayres-macbook-pro.local (cm-vpn01.mozilla.org [10.2.72.11]) (Authenticated sender: rsayre@mozilla.com) by dm-mail01.mozilla.org (Postfix) with ESMTP id DE895B8024 for <hybi@ietf.org>; Fri, 10 Apr 2009 00:12:50 -0700 (PDT)
Message-ID: <49DEF171.4080506@mozilla.com>
Date: Fri, 10 Apr 2009 03:12:49 -0400
From: Rob Sayre <rsayre@mozilla.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b3pre) Gecko/20090223 Thunderbird/3.0b2
MIME-Version: 1.0
To: hybi@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 07:11:43 -0000

Ian Hickson wrote:
>  Personally I am interested in "Minimal, Event", but I don't see any
>  reason why we wouldn't also have the other models.

I am interested in "Minimal, Event" as well.

- Rob


From john.fallows@kaazing.com  Fri Apr 10 00:18:24 2009
Return-Path: <john.fallows@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 28FA43A6D49 for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 00:18:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FLsGQt2YnGjR for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 00:18:23 -0700 (PDT)
Received: from mail-bw0-f169.google.com (mail-bw0-f169.google.com [209.85.218.169]) by core3.amsl.com (Postfix) with ESMTP id 388F53A6CEC for <hybi@ietf.org>; Fri, 10 Apr 2009 00:18:22 -0700 (PDT)
Received: by bwz17 with SMTP id 17so923892bwz.37 for <hybi@ietf.org>; Fri, 10 Apr 2009 00:19:30 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.102.247.4 with SMTP id u4mr1697887muh.104.1239347970672; Fri,  10 Apr 2009 00:19:30 -0700 (PDT)
In-Reply-To: <49DEF171.4080506@mozilla.com>
References: <49DEF171.4080506@mozilla.com>
Date: Fri, 10 Apr 2009 00:19:30 -0700
Message-ID: <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>
From: John Fallows <john.fallows@kaazing.com>
To: hybi@ietf.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 07:18:24 -0000

I am also primarily interested in "Minimal, Event".

Kind Regards,
John Fallows.

On Fri, Apr 10, 2009 at 12:12 AM, Rob Sayre <rsayre@mozilla.com> wrote:
> Ian Hickson wrote:
>>
>> =A0Personally I am interested in "Minimal, Event", but I don't see any
>> =A0reason why we wouldn't also have the other models.
>
> I am interested in "Minimal, Event" as well.
>
> - Rob
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>
--=20
>|< Kaazing Corporation >|<
John Fallows | CTO | +1.650.960.8148
888 Villa St, Ste 410 | Mountain View, CA 94041, USA

From prescod@gmail.com  Fri Apr 10 12:32:44 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 941833A63EB for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 12:32:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.597
X-Spam-Level: 
X-Spam-Status: No, score=-1.597 tagged_above=-999 required=5 tests=[AWL=0.380,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HeM3W1VykaMC for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 12:32:43 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.232]) by core3.amsl.com (Postfix) with ESMTP id DEDD83A6973 for <hybi@ietf.org>; Fri, 10 Apr 2009 12:32:43 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so972609rvb.49 for <hybi@ietf.org>; Fri, 10 Apr 2009 12:33:52 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=evXe684yNA0/N/q7ko82A0KOfhchRZ8T3C2DXPbKOQ8=; b=D6x9zhTR7N2m9RB3uyEoExDzKr+kTes8YDY+wipA9PLtVfTz0jkorigxVy56xKAYCY bcLHmGax2PEdXdd1hh06eakdrESQ3OQwSAjT02pPBHfLNmBqYT4goWdSmlK2kWqoQjPD 3XX76BaX3uCMvMu9Ha+6J72xVuJ9DlP1AXlyo=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=wJWITxLnDTgAXBbqSK0V8C9cmmZc25fKrYxL0dUcWHhwHvkZ/ZdcKQafvnTtoeBITm cSMLAXr1h1+paC+mYhknJvvd0rO8nMBN4GaH4f61UL3qsQZzPDQxgyJY5Rem7Z3ZUih1 wYYWgEkKCtIKx6h+BrV+n8GfLHiK/PDslkX78=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.180.16 with SMTP id h16mr1598963rvp.68.1239392032705; Fri,  10 Apr 2009 12:33:52 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com>
Date: Fri, 10 Apr 2009 12:33:52 -0700
X-Google-Sender-Auth: 0d808ef3b5e9f754
Message-ID: <1cb725390904101233r402f4b9cicb1cdb3d7bfeaec7@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 19:32:44 -0000

On Wed, Apr 8, 2009 at 5:34 PM, Ian Hickson <ian@hixie.ch> wrote:
> ...
>> We've successfully implemented rHTTP server and client components in
>> just a few dozen lines of code because we are able to just reuse the
>> existing HTTP client and server components that exist in the
>> environments we work in (C++, Python).
>
> Sure, if you can use libraries then implementing any protocol is trivial.
> The point is that the protocol should be simple enough that one doesn't
> need to rely on libraries to get a compliant implementation.

He's not talking about random libraries. He's talking specifically
about the libraries required to implement Web browsers and Web Server
*already*.

As someone else pointed out, if we presume no libraries then TCP,
UTF-8 and URLs are all intimidating standards that Web Sockets pulls
in "by reference". Nobody objects because we all know that libraries
for these are very widely deployed in Web systems. HTTP is also in
that category.

 Paul Prescod

From prescod@gmail.com  Fri Apr 10 12:56:35 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C23FD3A6985 for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 12:56:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.628
X-Spam-Level: 
X-Spam-Status: No, score=-1.628 tagged_above=-999 required=5 tests=[AWL=0.349,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ebOm7P+K6mZI for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 12:56:34 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.238]) by core3.amsl.com (Postfix) with ESMTP id E37833A6A68 for <hybi@ietf.org>; Fri, 10 Apr 2009 12:55:57 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so977779rvb.49 for <hybi@ietf.org>; Fri, 10 Apr 2009 12:57:06 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=a3F9hvf+GTNPl7al79911X9LdBkorbeeFRewxpd54JQ=; b=E1N53lKYT4eCPrpGdWcZotjDNOofBI24e5HiAK0cGlHID5ge3DqOxJ6sSMXDAG/Zpw a4moMcGYVdKAglqVtTZLrNTE7wkLEYxqFx4feZ44RDP/P24eR6hLmuWp5ecs24N27DeG zKROD6EiH7BEE5FqCO1hJ1KHuoGAGLhSBoRX4=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=LXAjbLraERQT2h2QpwbRth7CfcBSnvi6iEdEuhX78HPUSB82SLUDay62DzKi/PibdV kT5AkLurylj5SuW+521sCS45gxhocyoMqG4/0R+2agLvOARTYx/OthQ4GvMZzGwdQDh1 hh8UW9pmduNC9lvpgJIz0O449Y7oqe29VEY8Y=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.201.1 with SMTP id d1mr1600965rvq.142.1239393426730; Fri,  10 Apr 2009 12:57:06 -0700 (PDT)
In-Reply-To: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>
Date: Fri, 10 Apr 2009 12:57:06 -0700
X-Google-Sender-Auth: c4673120d3930375
Message-ID: <1cb725390904101257q57828e37t1b7a6d135fcce739@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Mark Lentczner <markl@lindenlab.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Apr 2009 19:56:35 -0000

On Wed, Apr 8, 2009 at 5:23 PM, Mark Lentczner <markl@lindenlab.com> wrote:
> Sorry about the hodge-podge here: After IETF-74 I had a zillion things to do
> at work, so it wasn't until this morning that I was able to fully catch up
> on my backlog of reading hybi mail. Here are a number of bit-and-bobs that I
> wanted to reply to:
>
> the whole point of
> bidirectional communication is that it's not a request-response model.
>
> Actually, I don't agree. At Linden Lab, we've seen a great need for
> bidirectional communication that is request/response model.

I am curious what you typically put in responses to messages from the
"central server" to a browser running on an edge device.

I'd also be curious whether you expect the response to be provided by
Javascript or by an underlying layer that just says: "yeah, I got the
message: I'll pass it to Javascript when that thread is available."

As I think about it today I wonder if it is really helpful for the
server to wait for a response from the client, which it will typically
ignore (IMO).

 Paul Prescod

From rektide@voodoowarez.com  Fri Apr 10 23:26:11 2009
Return-Path: <rektide@voodoowarez.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0BFBD3A6A87 for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 23:26:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.299
X-Spam-Level: 
X-Spam-Status: No, score=-2.299 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id selw-yobVYhi for <hybi@core3.amsl.com>; Fri, 10 Apr 2009 23:26:10 -0700 (PDT)
Received: from deneb.voodoowarez.com (eldergods.com [72.9.228.236]) by core3.amsl.com (Postfix) with ESMTP id EAA8B3A6AB2 for <hybi@ietf.org>; Fri, 10 Apr 2009 23:26:00 -0700 (PDT)
Received: by deneb.voodoowarez.com (Postfix, from userid 1000) id 022C1B219A; Fri, 10 Apr 2009 23:33:45 -0400 (EDT)
Date: Fri, 10 Apr 2009 23:33:45 -0400
From: rektide <rektide@voodoowarez.com>
To: Martin =?iso-8859-1?Q?J=2E_D=FCrst?= <duerst@it.aoyama.ac.jp>
Message-ID: <20090411033345.GD9817@voodoowarez.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com> <49DD6B77.50104@it.aoyama.ac.jp>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <49DD6B77.50104@it.aoyama.ac.jp>
User-Agent: Mutt/1.5.17+20080114 (2008-01-14)
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 11 Apr 2009 06:26:11 -0000

On Thu, Apr 09, 2009 at 12:28:55PM +0900, "Martin J. Dürst" wrote:
> On 2009/04/09 5:02, Paul Prescod wrote:
>
>> I would add a "data synchronization" data model. I find that most
>> often the goal is to have the web browser report the state of a
>> resource or resource collection to a user, and then inform the user of
>> changes to it.
>
> That is indeed a very good point. Most kinds of collaborative activity  
> can easily be looked at that way. In its full incarnation, maybe it  
> would just allow an application programmer to say: synchronize this part  
> of this page (i.e. DOM) with this part of that page over at that client.
>
> Regards,    Martin.

This seems like something that would be ideal to layer on top of a HyBi/"Server Initiated
HTTP" transport.  It doesnt sound like something that everyone using duplex HTTP would want.
Whether or not the conversation about this possible protocol belongs here, I truly do not
know, but I want to call it out right now as something I see as decidedly "not transport",
which was the question I interpretted as lying at the heart of Apples and Orangutans.

rektide




From rektide@voodoowarez.com  Sat Apr 11 00:36:09 2009
Return-Path: <rektide@voodoowarez.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EEB4F3A6AF8 for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 00:36:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.277
X-Spam-Level: 
X-Spam-Status: No, score=-2.277 tagged_above=-999 required=5 tests=[AWL=-0.022, BAYES_00=-2.599, DATE_IN_PAST_03_06=0.044, MIME_8BIT_HEADER=0.3]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id EaxTTOEc7Cr8 for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 00:36:09 -0700 (PDT)
Received: from deneb.voodoowarez.com (eldergods.com [72.9.228.236]) by core3.amsl.com (Postfix) with ESMTP id 2C6F43A6AF0 for <hybi@ietf.org>; Sat, 11 Apr 2009 00:36:00 -0700 (PDT)
Received: by deneb.voodoowarez.com (Postfix, from userid 1000) id B9A0DB219C; Fri, 10 Apr 2009 23:29:40 -0400 (EDT)
Date: Fri, 10 Apr 2009 23:29:40 -0400
From: rektide <rektide@voodoowarez.com>
To: Martin =?iso-8859-1?Q?J=2E_D=FCrst?= <duerst@it.aoyama.ac.jp>
Message-ID: <20090411032940.GC9817@voodoowarez.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com> <1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com> <49DD6B77.50104@it.aoyama.ac.jp>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <49DD6B77.50104@it.aoyama.ac.jp>
User-Agent: Mutt/1.5.17+20080114 (2008-01-14)
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 11 Apr 2009 07:36:10 -0000

On Thu, Apr 09, 2009 at 12:28:55PM +0900, "Martin J. Dürst" wrote:
>
>
> On 2009/04/09 5:02, Paul Prescod wrote:
>
>> I would add a "data synchronization" data model. I find that most
>> often the goal is to have the web browser report the state of a
>> resource or resource collection to a user, and then inform the user of
>> changes to it.
>
> That is indeed a very good point. Most kinds of collaborative activity  
> can easily be looked at that way. In its full incarnation, maybe it  
> would just allow an application programmer to say: synchronize this part  
> of this page (i.e. DOM) with this part of that page over at that client.
>
> Regards,    Martin.

This seems like something that would be ideal to layer on top of a HyBi/"Server Initiated
HTTP" transport.  It doesnt sound like something that everyone using duplex HTTP would want.
Whether or not the conversation about this possible protocol belongs here, I truly do not 
know, but I want to call it out right now as something I see as decidedly "not transport",
which was the question I interpretted as lying at the heart of Apples and Orangutans.

rektide




From rektide@voodoowarez.com  Sat Apr 11 00:59:11 2009
Return-Path: <rektide@voodoowarez.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0BAA93A6E15 for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 00:59:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.416
X-Spam-Level: 
X-Spam-Status: No, score=-2.416 tagged_above=-999 required=5 tests=[AWL=0.139,  BAYES_00=-2.599, DATE_IN_PAST_03_06=0.044]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 24HLdQ7xe8qe for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 00:59:09 -0700 (PDT)
Received: from voodoowarez.com (eldergods.com [72.9.228.236]) by core3.amsl.com (Postfix) with ESMTP id CD2883A6C7D for <hybi@ietf.org>; Sat, 11 Apr 2009 00:59:09 -0700 (PDT)
Received: by deneb.voodoowarez.com (Postfix, from userid 1000) id EBB10B2191; Fri, 10 Apr 2009 23:23:24 -0400 (EDT)
Date: Fri, 10 Apr 2009 23:23:24 -0400
From: rektide <rektide@voodoowarez.com>
To: Mark Lentczner <markl@lindenlab.com>
Message-ID: <20090411032324.GB9817@voodoowarez.com>
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com>
User-Agent: Mutt/1.5.17+20080114 (2008-01-14)
Cc: hybi@ietf.org
Subject: Re: [hybi] asychronous HTTP [was:Apples and Orangutans]
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 11 Apr 2009 07:59:11 -0000

On Wed, Apr 08, 2009 at 08:47:29AM -0700, Mark Lentczner wrote:
> From the point of view of our customers, the application level authors 
> (both client and server side), I think there are two major questions they 
> would ask to characterize any potential protocol:
>
> 1) What is the data model?
> 	Minimal (binary blob and/or UTF-8 string)
> 	Structured (JSON object or XML stanza)
> 	Message (data and meta-data)
>
> 2) What is interaction model?
> 	Event (datagram at the senders whim)
> 	Request/Response (request at senders whim, with response from the  
> receiver)

A messaging protocol ought be able to meet all of the interaction & data models above, 
ought be designed to convey any content, and potentially ails one of the greatest pitfalls 
of the web: the two/multiple connection limit[1].  

If the web ecosystem evolves towards w3c widgets and integrative software networks ala 
OpenSocial, the connection limit must be addressed.  And not by juvenile hacks to permit 
more connections (eg by creating subdomains); the real answer is that the web needs a 
multiplexed communication channel.  I can think of nothing more suitable for holding multiple 
conversations over a single wire than a messaging protocol.

Recently in #persevere (April 7), reflecting on HyBi:
<rektide> i really love greg wilkins constant chime ins that, http is a great model, why
break complete from it

HTTP already meets many of the critereon for being a suitable messaging protocol, and I
think this is of capital importance.  HTTP pipelining allows asychronous data sending, Kris
Zyp has developed HTTP Channels[2] to permit decoupled asychornous data recieving, and if
these are leveraged together, it leaves only the outstanding issue of proxies and multiplexing:
adapting HTTP to be a multichannel bidirectional messaging system supporting meta-data. 

Kris' model in HTTP Channels introduces a client-id and a sequence-id that can be used to
maintain conversations while clients are hurling a stream of pipelined HTTP requests and a
server is sending data to the clients in an out of order manner.  These tokens are sent as
HTTP headers, and although not presently a part of the server response in Kris' HTTP Channels 
spec, I believe ought be adapted as the primary multiplexing mechanism.

The one issue I think isnt particularly well addressed is broadcast messaging; can the
server send events to multiple or all client-id's, and how do these broadcast messages get
serialized?

Assuming the proxy issue can be adapted to be non terminal, assuming a persistent TCP
connection is not a deal breaker (i.e. a non connection oriented transport such as UDP is not
required), it ought be possible to fully extend HTTP to be a asychronous bidirectional
multiplexed communication channel, and further, I believe it would be possible to implement
a client side implementation on top of existing web browsers entirely client side (although
browser code for strictly enforcing multiplexing constraints would be "a good thing").

[1] http://www.openajax.org/runtime/wiki/The_Two_HTTP_Connection_Limit_Issue
[2] http://cometdaily.com/2008/05/13/http-channels-2/


> While issues of data framing, efficient use of TCP, and ease of  
> implementation, will influence their view, I don't think they will be  
> nearly as important as these first two questions[*].
>
> The protocols we're discussing can be characterized by these two  
> questions:
> 	WebSockets: Minimal, Event
> 	BOSH, Bayeaux: Structured, Event
> 	rHTTP:  Message, Req/Resp
>
> Other protocols can be fit into this framework, with perhaps some  
> extension:
> 	XMPP: Structured , Event
> 	TCP: Minimal, Stream
> 	SMTP: Message, Event
> 	HTTP: Message, Req/Resp

To emphasize, we already have the tooling to make HTTP sending and (most of the tools for
making) HTTP recieving asychronous.  This Event model HTTP still supports Req/Resp http
communication as well.


> Now, while clearly any of these be used to implement the others, there  
> is a cost. Where simpler models are implemented with more complex, then 
> there is overhead on the wire and processing. Where more complex models 
> are implemented on simpler, then the burden falls to the application 
> developer.

The cost savings I'd like to see most is being able to leverage HTTP, by simply removing
its fault of being so sychronous & undirectional.  If anyone believes a fully duplex multiplexed
communication channel is of itself insufficient or ill suited for your needs, please speak 
as to why.  If you find the HTTP framing unsuitable, eg a header & a payload, that too would
be intesting to hear.


Good day all, 
rektide




From rektide@voodoowarez.com  Sat Apr 11 10:19:51 2009
Return-Path: <rektide@voodoowarez.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D9BF23A6B20 for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 10:19:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.484
X-Spam-Level: 
X-Spam-Status: No, score=-2.484 tagged_above=-999 required=5 tests=[AWL=0.115,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id peyKgjK+wX3r for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 10:19:51 -0700 (PDT)
Received: from voodoowarez.com (eldergods.com [72.9.228.236]) by core3.amsl.com (Postfix) with ESMTP id E16BE3A6B0C for <hybi@ietf.org>; Sat, 11 Apr 2009 10:19:50 -0700 (PDT)
Received: by voodoowarez.com (Postfix, from userid 1000) id 154ECB21A6; Sat, 11 Apr 2009 13:21:00 -0400 (EDT)
Date: Sat, 11 Apr 2009 13:20:59 -0400
From: rektide <rektide@voodoowarez.com>
To: Mark Lentczner <markl@lindenlab.com>
Message-ID: <20090411172059.GE9817@voodoowarez.com>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
User-Agent: Mutt/1.5.17+20080114 (2008-01-14)
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 11 Apr 2009 17:19:51 -0000

On Thu, Apr 09, 2009 at 09:06:23AM -0700, Mark Lentczner wrote:
>
> On Apr 8, 2009, at 5:34 PM, Ian Hickson wrote:
>> I misspoke, I should have said that the point of full-duplex  
>> communication
>> is that it's not request-response -- either side can spontaneously  
>> send a
>> message.
>
> Then I mis-replied. I should have said that we have a need for full- 
> duplex request-response -- either side can spontaneously send a request 
> message, and either side can, when ready, reply to a prior request.

I'd inject that a "worthy" transport also has to be able to permit either side to ad-hoc
initiate communication at any time as well.  I'd also say either end must be able to opt to
never reply without violating the transport spec.

How the reader dices this is anyone's guess: is this is a request-response transport with
optional reply, or is this a event transport with optional request-reply? I dont think it
matters, I think a good transport ought to support both.

>> The socket in WebSocket can be (indeed, must be) addressed via URL.
>
> Only the establishment in WebSocket is addressed via URL.  In rHTTP each 
> request can be addressed by URL.  This is useful as it is a very common 
> piece of meta-data that applications want. For example, using  
> WebSockets, for any complex page, an application is almost certainly  
> going to want a way to address which widget on the page is the intended 
> recipient of the event. Hence, each application is going to have to 
> devise an addressing scheme, and more importantly, a method of encoding 
> that address in the simple string event.

Addressability to the client side is something I had not considered before.  I would be very
interested to hear how this would work in a collaborative environment, how the server would
deal with multiple clients.

When I was heavily investigating Bayeux, I often thought each client should have a service
channel to enable cross system addressability.  I never had a concrete API sketched out for
how the endpoint would function, just the notion that there would be a URL template for
talking to clients.

>> At the message level, why do we need to force those features on all  
>> users?
>> They seem like features we should allow people to have if they want,  
>> but
>> they don't seem like features we should require all clients and  
>> servers to
>> implement just because someone might want them.
>
> Of course, but you must recognize there is a trade-off here that all  
> protocols make. The design needs to choose a feature set wide enough  
> that many different applications can make use of it, but must do so  
> balancing complexity and overhead. For example, WebSockets has chosen to 
> support both binary and UTF-8 payloads. The overhead of the lead-in type 
> byte was seen as worth it, even though there are many applications that 
> will only use the UTF-8 strings. So, I see the choices here as just a 
> matter of degree. I think popularity of HTTP for many kinds of 
> applications has shown that one or more features of HTTP(*) are often 
> applicable to an application's needs, while the overhead of these 
> features is low enough to be worth including them in the protocol for 
> all.(**)

Agreed, I think HTTP's framing is well worth the cost.

-rektide




From gregw@webtide.com  Sat Apr 11 19:26:26 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B5F953A6929 for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 19:26:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.502
X-Spam-Level: 
X-Spam-Status: No, score=-2.502 tagged_above=-999 required=5 tests=[AWL=0.097,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5WaHDaIVGDth for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 19:26:25 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id 812903A684D for <hybi@ietf.org>; Sat, 11 Apr 2009 19:26:25 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id j3so1155857tid.25 for <hybi@ietf.org>; Sat, 11 Apr 2009 19:27:34 -0700 (PDT)
Received: by 10.110.63.17 with SMTP id l17mr4221934tia.48.1239503254646; Sat, 11 Apr 2009 19:27:34 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id a14sm4401554tia.27.2009.04.11.19.27.31 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sat, 11 Apr 2009 19:27:33 -0700 (PDT)
Message-ID: <49E1518E.7050402@webtide.com>
Date: Sun, 12 Apr 2009 12:27:26 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Mark Lentczner <markl@lindenlab.com>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com>	<Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
In-Reply-To: <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 02:26:26 -0000

Mark Lentczner wrote:
> 
>  I should have said that we have a need for 
> full-duplex request-response -- either side can spontaneously send a 
> request message, and either side can, when ready, reply to a prior request.

BEEP!

> for any complex page, an application is almost certainly 
> going to want a way to address which widget on the page is the intended 
> recipient of the event. Hence, each application is going to have to 
> devise an addressing scheme, and more importantly, a method of encoding 
> that address in the simple string event.

BEEP!

 > I think popularity of HTTP for many kinds of
> applications has shown that one or more features of HTTP(*) are often 
> applicable to an application's needs, while the overhead of these 
> features is low enough to be worth including them in the protocol for 
> all.(**)
> 
>     - Mark
> 
> (*) Addressing (URLs), Method, Headers, Content specification, Transfer 
> specification, Status

BEEP!

> (**) I imagine the wire overhead for putting a minimal string event into 
> an HTTP message would be something like:
>     POST / HTTP/1.1
>     Host:
>     Content-Type:text/plain;charset=UTF-8
>     Content-Length:xxx
> 
> and
>     204 ACK
> That's 96 bytes overhead for a 100 to 999 byte payload. Seems high, but 
> for the usage cases that web pages present, this probably is not 
> relevant.  Further, once one has this framing, it is very likely that 
> rather than build addressing into the string payload, applications would 
> just make use of the URL field. Plus, this framing already has character 
> encoding support.

BEEP!


Generally agreeing with all your points.  Yet the attributes you
describe do sound pretty much like BEEP protocol - or at least
beep framing.

BEEP is very much a HTTP style protocol with addition of
bidirectionality, multiple response requests and multiple
channels over a single connection.

If BEEP was extended with a few default assumptions (eg
UTF-8 text body if no content type specified), then it would
approach the byte efficiency of websocket, while preserving
the extensibility and flexibility of HTTP.

If the security and authentication aspects of BEEP were
stripped, then a pretty simply protocol would remain.


IF I was forced to pick a winner.... then I'd certainly
look very closely at BEEP or a BEEP derivative.


cheers










From mike@dierken.com  Sat Apr 11 20:01:05 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 32A4A3A689F for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 20:01:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=-0.001, BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 27dhPk6DaUqU for <hybi@core3.amsl.com>; Sat, 11 Apr 2009 20:01:04 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.239]) by core3.amsl.com (Postfix) with ESMTP id BF22E3A67E1 for <hybi@ietf.org>; Sat, 11 Apr 2009 20:00:41 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id b25so1165902rvf.5 for <hybi@ietf.org>; Sat, 11 Apr 2009 20:01:51 -0700 (PDT)
Received: by 10.143.7.5 with SMTP id k5mr2016254wfi.176.1239505310355; Sat, 11 Apr 2009 20:01:50 -0700 (PDT)
Received: from mercury (c-76-121-195-84.hsd1.wa.comcast.net [76.121.195.84]) by mx.google.com with ESMTPS id k37sm9614866rvb.0.2009.04.11.20.01.49 (version=SSLv3 cipher=RC4-MD5); Sat, 11 Apr 2009 20:01:49 -0700 (PDT)
From: "S. Mike Dierken" <mike@dierken.com>
To: <hybi@ietf.org>
Date: Sat, 11 Apr 2009 20:02:16 -0700
Message-ID: <F813335479CB4F70A58EC314CF9D362B@mercury>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----=_NextPart_000_001C_01C9BAE0.6A695B60"
X-Mailer: Microsoft Office Outlook 11
Thread-Index: Acm7GxK/UkDXJP+cQf6D5zE05/aM/A==
x-mimeole: Produced By Microsoft MimeOLE V6.00.2900.5579
Subject: [hybi] Question about charter/mission/summary/etc.
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 03:01:05 -0000

This is a multi-part message in MIME format.

------=_NextPart_000_001C_01C9BAE0.6A695B60
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

I may have joined this mailing list a bit late, and I'm a bit lost trying to
find a place with a summary of the goal, the players, overview, history and
whatnot.
So far, what I'm reading sounds just like the what-wg 'event source' stuff
from a few years back and I was wondering what happened with that, and how
HyBi relates to that. Can someone share this background and/or point me to a
page with more info?

Back then, I had written down some thoughts and these may be relevant to the
current hybi discussion, especially using HTTP message framing for the
server-to-client messages (though I see it as still client-to-server but the
server is on the desktop)
http://www.searchalert.net/dierken/eventsource/
http://korrespondence.blogspot.com/2005/06/whatwg-and-event-source-design.ht
ml


------=_NextPart_000_001C_01C9BAE0.6A695B60
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
6.5.7036.0">
<TITLE>Question about charter/mission/summary/etc.</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/rtf format -->

<P><FONT SIZE=3D2 FACE=3D"Arial">I may have joined this mailing list a =
bit late, and I'm a bit lost trying to find a place with a summary of =
the goal, the players, overview, history and whatnot.</FONT></P>

<P><FONT SIZE=3D2 FACE=3D"Arial">So far, what I'm reading sounds just =
like the what-wg 'event source' stuff from a few years back and I was =
wondering what happened with that, and how HyBi relates to that. Can =
someone share this background and/or point me to a page with more =
info?</FONT></P>

<P><FONT SIZE=3D2 FACE=3D"Arial">Back then, I had written down some =
thoughts and these may be relevant to the current hybi discussion, =
especially using HTTP message framing for the server-to-client messages =
(though I see it as still client-to-server but the server is on the =
desktop)</FONT></P>

<P><A HREF=3D"http://www.searchalert.net/dierken/eventsource/"><U><FONT =
COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Arial">http://www.searchalert.net/dierken/eventsource/</FONT></U>=
</A>

<BR><A =
HREF=3D"http://korrespondence.blogspot.com/2005/06/whatwg-and-event-sourc=
e-design.html"><U><FONT COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Arial">http://korrespondence.blogspot.com/2005/06/whatwg-and-even=
t-source-design.html</FONT></U></A>
</P>

</BODY>
</HTML>
------=_NextPart_000_001C_01C9BAE0.6A695B60--


From mridulm80@yahoo.com  Sun Apr 12 05:41:50 2009
Return-Path: <mridulm80@yahoo.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3088A3A6B79 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 05:41:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.25
X-Spam-Level: 
X-Spam-Status: No, score=-2.25 tagged_above=-999 required=5 tests=[AWL=0.349,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id y7m+pkgwl6Q9 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 05:41:49 -0700 (PDT)
Received: from n8-vm0.bullet.mail.in.yahoo.com (n8-vm0.bullet.mail.in.yahoo.com [202.86.4.189]) by core3.amsl.com (Postfix) with SMTP id C22053A697E for <hybi@ietf.org>; Sun, 12 Apr 2009 05:41:48 -0700 (PDT)
Received: from [202.86.4.170] by n8.bullet.mail.in.yahoo.com with NNFMP; 12 Apr 2009 12:42:57 -0000
Received: from [203.104.17.88] by t1.bullet.in.yahoo.com with NNFMP; 12 Apr 2009 12:42:57 -0000
Received: from [127.0.0.1] by omp102.mail.in2.yahoo.com with NNFMP; 12 Apr 2009 12:42:57 -0000
X-Yahoo-Newman-Property: ymail-3
X-Yahoo-Newman-Id: 554666.9551.bm@omp102.mail.in2.yahoo.com
Received: (qmail 37218 invoked by uid 60001); 12 Apr 2009 12:42:57 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1239540177; bh=RAORkZpfQusK3exHSzaR6vI3xN5OCSY4tqxUcI8bYsE=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=rw6hO4Kn0SQj/G7MUmIfMYV2bipmrzlIss1uPGKHTIJM5utiCO1WJe8KwnS5Sxxsxy+pgyxocKLkc+fQr//XOuOYJNOq08kEDOqikbmJsXVe4ccbDF+LMPYhIWA5NNjxhEOXwKLD33wZD2JrwoJ1ZF4fQBAcntM7YM/E8xly5Bs=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=rYNOw5rQeijq8Imt8LcSrIpB65m3zvnAEkz+5UQeXzFagBZtSAna4Vf/UzTGygYwmq7GvD5CSXdt064t6Sq6odSea/Hl006OYboxYGhDGfgKdTyETH2QTXFO3AWXpA28tRIX7hDajGzj1RYzT/92Up/7Q5Sr+VJ5Ur73+CrFMCE=;
Message-ID: <377449.36122.qm@web95405.mail.in2.yahoo.com>
X-YMail-OSG: BFbMf7sVM1nAHEwInpxuOVt2vjOgzCxJuwBkaR1tjzmozAni7slULSt1YP0q8ni5yvGNBycI6tFxmAhX1vj9PxSGc1iL3EgehogRGPBjn4_LyHz_p_2Jsh.mTu.MAZ1.boayxI5vZEHXgNNBuyO6HBrEhtT3Hjha4BSctatuowRZ6o8Pp36udmm3.kmNAOe11bJ.H_1.96le6w--
Received: from [122.167.176.126] by web95405.mail.in2.yahoo.com via HTTP; Sun, 12 Apr 2009 18:12:57 IST
X-Mailer: YahooMailClassic/5.2.15 YahooMailWebService/0.7.289.1
Date: Sun, 12 Apr 2009 18:12:57 +0530 (IST)
From: Mridul Muralidharan <mridulm80@yahoo.com>
To: Mark Lentczner <markl@lindenlab.com>, Greg Wilkins <gregw@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 12:41:50 -0000

=0A=0AYes, BEEP does fit the requirements discussed so far ... (not to ment=
ion support for flow control which becomes immediately necessary in bi-dire=
ctional streams).=0A=0AFlip side would be, I am not sure if it is upgradabl=
e from HTTP with current infrastructure - other than through CONNECT that i=
s.=0A=0ARegards,=0AMridul=0A=0A--- On Sun, 12/4/09, Greg Wilkins <gregw@web=
tide.com> wrote:=0A=0A> From: Greg Wilkins <gregw@webtide.com>=0A> Subject:=
 Re: [hybi] Random Responses=0A> To: "Mark Lentczner" <markl@lindenlab.com>=
=0A> Cc: hybi@ietf.org=0A> Date: Sunday, 12 April, 2009, 7:57 AM=0A> Mark L=
entczner wrote:=0A> > =0A> >=A0 I should have said that we have a need for=
=0A> full-duplex request-response -- either side can=0A> spontaneously send=
 a request message, and either side can,=0A> when ready, reply to a prior r=
equest.=0A> =0A> BEEP!=0A> =0A> > for any complex page, an application is a=
lmost=0A> certainly going to want a way to address which widget on the=0A> =
page is the intended recipient of the event. Hence, each=0A> application is=
 going to have to devise an addressing scheme,=0A> and more importantly, a =
method of encoding that address in=0A> the simple string event.=0A> =0A> BE=
EP!=0A> =0A> > I think popularity of HTTP for many kinds of=0A> > applicati=
ons has shown that one or more features of=0A> HTTP(*) are often applicable=
 to an application's needs,=0A> while the overhead of these features is low=
 enough to be=0A> worth including them in the protocol for all.(**)=0A> > =
=0A> >=A0 =A0=A0=A0- Mark=0A> > =0A> > (*) Addressing (URLs), Method, Heade=
rs, Content=0A> specification, Transfer specification, Status=0A> =0A> BEEP=
!=0A> =0A> > (**) I imagine the wire overhead for putting a minimal=0A> str=
ing event into an HTTP message would be something like:=0A> >=A0 =A0=A0=A0P=
OST / HTTP/1.1=0A> >=A0 =A0=A0=A0Host:=0A> >=A0=0A> =A0=A0=A0Content-Type:t=
ext/plain;charset=3DUTF-8=0A> >=A0 =A0=A0=A0Content-Length:xxx=0A> > =0A> >=
 and=0A> >=A0 =A0=A0=A0204 ACK=0A> > That's 96 bytes overhead for a 100 to =
999 byte=0A> payload. Seems high, but for the usage cases that web pages=0A=
> present, this probably is not relevant.=A0 Further, once=0A> one has this=
 framing, it is very likely that rather than=0A> build addressing into the =
string payload, applications would=0A> just make use of the URL field. Plus=
, this framing already=0A> has character encoding support.=0A> =0A> BEEP!=
=0A> =0A> =0A> Generally agreeing with all your points.=A0 Yet the=0A> attr=
ibutes you=0A> describe do sound pretty much like BEEP protocol - or at=0A>=
 least=0A> beep framing.=0A> =0A> BEEP is very much a HTTP style protocol w=
ith addition of=0A> bidirectionality, multiple response requests and multip=
le=0A> channels over a single connection.=0A> =0A> If BEEP was extended wit=
h a few default assumptions (eg=0A> UTF-8 text body if no content type spec=
ified), then it=0A> would=0A> approach the byte efficiency of websocket, wh=
ile=0A> preserving=0A> the extensibility and flexibility of HTTP.=0A> =0A> =
If the security and authentication aspects of BEEP were=0A> stripped, then =
a pretty simply protocol would remain.=0A> =0A> =0A> IF I was forced to pic=
k a winner.... then I'd certainly=0A> look very closely at BEEP or a BEEP d=
erivative.=0A> =0A> =0A> cheers=0A> =0A> =0A> =0A> =0A> =0A> =0A> =0A> =0A>=
 =0A> _______________________________________________=0A> hybi mailing list=
=0A> hybi@ietf.org=0A> https://www.ietf.org/mailman/listinfo/hybi=0A> =0A=
=0A=0A      Add more friends to your messenger and enjoy! Go to http://mess=
enger.yahoo.com/invite/


From annevk@opera.com  Sun Apr 12 10:33:47 2009
Return-Path: <annevk@opera.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 564043A6B01 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 10:33:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.185
X-Spam-Level: 
X-Spam-Status: No, score=-4.185 tagged_above=-999 required=5 tests=[BAYES_40=-0.185, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id vn8w-yNk5Wsx for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 10:33:46 -0700 (PDT)
Received: from smtp.opera.com (sam.opera.com [213.236.208.81]) by core3.amsl.com (Postfix) with ESMTP id 3B0A93A6A6E for <hybi@ietf.org>; Sun, 12 Apr 2009 10:33:45 -0700 (PDT)
Received: from anne-van-kesterens-macbook.local (37.Red-83-45-204.dynamicIP.rima-tde.net [83.45.204.37]) (authenticated bits=0) by smtp.opera.com (8.13.4/8.13.4/Debian-3sarge3) with ESMTP id n3CHYa9f009236 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Sun, 12 Apr 2009 17:34:52 GMT
Date: Sun, 12 Apr 2009 19:34:35 +0200
To: "John Fallows" <john.fallows@kaazing.com>, hybi@ietf.org
From: "Anne van Kesteren" <annevk@opera.com>
Organization: Opera Software ASA
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>
Content-Transfer-Encoding: 8bit
Message-ID: <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>
In-Reply-To: <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>
User-Agent: Opera Mail/9.63 (MacIntel)
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 17:33:47 -0000

Me too.

On Fri, 10 Apr 2009 09:19:30 +0200, John Fallows  
<john.fallows@kaazing.com> wrote:
> I am also primarily interested in "Minimal, Event".
>
> Kind Regards,
> John Fallows.
>
> On Fri, Apr 10, 2009 at 12:12 AM, Rob Sayre <rsayre@mozilla.com> wrote:
>> Ian Hickson wrote:
>>>
>>> Â Personally I am interested in "Minimal, Event", but I don't see any
>>> Â reason why we wouldn't also have the other models.
>>
>> I am interested in "Minimal, Event" as well.
>>
>> - Rob


-- 
Anne van Kesteren
http://annevankesteren.nl/

From jamie@shareable.org  Sun Apr 12 11:27:55 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2AE873A6930 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 11:27:55 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.799
X-Spam-Level: 
X-Spam-Status: No, score=-5.799 tagged_above=-999 required=5 tests=[AWL=-3.200, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FlRVGKAK-zHO for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 11:27:54 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 507DC3A67EB for <hybi@ietf.org>; Sun, 12 Apr 2009 11:27:54 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lt4QO-0001Dl-8n; Sun, 12 Apr 2009 19:29:00 +0100
Date: Sun, 12 Apr 2009 19:29:00 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mridul Muralidharan <mridulm80@yahoo.com>
Message-ID: <20090412182900.GA4394@shareable.org>
References: <377449.36122.qm@web95405.mail.in2.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <377449.36122.qm@web95405.mail.in2.yahoo.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 18:27:55 -0000

Mridul Muralidharan wrote:
> Yes, BEEP does fit the requirements discussed so far ...

BEEP doesn't meet the requirement for simple implementation because it
requires every implementation to parse XML.  HTTP does not.

Also BEEP's per-channel setup latency is rather large: one redundant
round trip.  Using BEEP for a simple HTTP request would take noticably
longer than HTTP itself on high latency networks.

> (not to mention support for flow control which becomes immediately
> necessary in bi-directional streams).

BEEP's flow control is sensible, but that's not the reason for it.

For bi-directional streams, TCP/IP's built in flow control is enough.

If you imagine a bi-directional HTTP (easy because requests and
responses are easily distinguished), it does not need any BEEP-style
flow control.  (I have implemented this).

BEEP-style per channel flow control is needed because of its ability
to multiplex multiple channels in the _same_ direction on a single
TCP/IP stream.  Without per-channel flow control you get deadlocks or
infinite buffering needed in some practical uses.

Unfortunately BEEPs per channel flow control is sensible, but it's per
channel setup is expensive both in needing XML and in network round
trip time.  It would be good to keep the flow control but use smarter
channel setup.

> Flip side would be, I am not sure if it is upgradable from HTTP with
> current infrastructure - other than through CONNECT that is.

It's not, but all protocols discussed can be overlaid onto the current
infrastructure using the HTTP hacks we use today (long GET etc.),
while giving new proxies the chance to recognise the overlaid protocol
and participate in it properly (dropping the hacks, proxying better).

Imho it would be good to discuss how we can overlay any new protocol
that we're talking about compatibly onto the existing HTTP infrastructure.

-- Jamie

From mridulm80@yahoo.com  Sun Apr 12 12:21:13 2009
Return-Path: <mridulm80@yahoo.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E10DA3A6B21 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 12:21:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.668
X-Spam-Level: 
X-Spam-Status: No, score=-1.668 tagged_above=-999 required=5 tests=[AWL=-0.465, BAYES_00=-2.599, MIME_QP_LONG_LINE=1.396]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RLox+5T7hfoB for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 12:21:13 -0700 (PDT)
Received: from n11-vm0.bullet.mail.in.yahoo.com (n11-vm0.bullet.mail.in.yahoo.com [202.86.4.227]) by core3.amsl.com (Postfix) with SMTP id D6B953A6965 for <hybi@ietf.org>; Sun, 12 Apr 2009 12:21:11 -0700 (PDT)
Received: from [202.43.219.99] by n11.bullet.mail.in.yahoo.com with NNFMP; 12 Apr 2009 19:22:20 -0000
Received: from [203.104.18.50] by t3.bullet.in.yahoo.com with NNFMP; 12 Apr 2009 19:22:20 -0000
Received: from [127.0.0.1] by omp111.mail.in2.yahoo.com with NNFMP; 12 Apr 2009 19:22:20 -0000
X-Yahoo-Newman-Property: ymail-3
X-Yahoo-Newman-Id: 584332.92588.bm@omp111.mail.in2.yahoo.com
Received: (qmail 84180 invoked by uid 60001); 12 Apr 2009 19:15:39 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1239563739; bh=krHpE6w3yQlXh5i1PM4HHWO6vMp7ebLAts7K9B/4+AA=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=r6rWxnnEaaniGJgQB042XRiZSli9EONhOz3O8KrDxtNPEuqZRV2m3dMVp/jR4K9XwVw6wDmuZ05S1CUUmyWxsXMe4PeyQwkKlphVzGuEQYNCeiEYhlIcEEkGDksV1wY/5Q8UlNIzRqzVuX1lcIvK+7DTEALqJgdfGt0J6pecjkg=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=5Fe1+1ojGwuM+4pf3fiBOJUuqs8D3NNfx6WsCBCu0eaDq1VGfL8QIP/OjYhJ6GTJf5b6TckUp505u/jOT7uFGnVYPO4UPygtvDEnlS2LGs76EvonQ/P335TyT0g77Rjl7WhYPEs6MnilqyHIoMfpjTGPEvLBiY60GpVmmaHTch0=;
Message-ID: <605442.82172.qm@web95416.mail.in2.yahoo.com>
X-YMail-OSG: qDeU3kcVM1mvjD05gUhifw7zVMmfE16HLk_yEVyRJVPXKlPnpZtcDH1rFCdC0PCtTiG.Kt9dCgKE6pdJkQFravqUD52P1PxaBewu2gMj5rrfSeHPgq9_SQN39VmMy4e38t6Yt0LO4dxgtwL83jLSX7id.hruHfZY5Dkju40ebJ17PPN_73b0Jnzdqq_97WOH_JmnV9odUjao9g--
Received: from [122.167.176.126] by web95416.mail.in2.yahoo.com via HTTP; Mon, 13 Apr 2009 00:45:39 IST
X-Mailer: YahooMailClassic/5.2.15 YahooMailWebService/0.7.289.1
Date: Mon, 13 Apr 2009 00:45:39 +0530 (IST)
From: Mridul Muralidharan <mridulm80@yahoo.com>
To: Jamie Lokier <jamie@shareable.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 19:21:14 -0000

=0A=0A--- On Sun, 12/4/09, Jamie Lokier <jamie@shareable.org> wrote:=0A=0A>=
 From: Jamie Lokier <jamie@shareable.org>=0A> Subject: Re: [hybi] Random Re=
sponses=0A> To: "Mridul Muralidharan" <mridulm80@yahoo.com>=0A> Cc: "Mark L=
entczner" <markl@lindenlab.com>, "Greg Wilkins" <gregw@webtide.com>, hybi@i=
etf.org=0A> Date: Sunday, 12 April, 2009, 11:59 PM=0A> Mridul Muralidharan =
wrote:=0A> > Yes, BEEP does fit the requirements discussed so far=0A> ...=
=0A> =0A> BEEP doesn't meet the requirement for simple implementation=0A> b=
ecause it=0A> requires every implementation to parse XML.=A0 HTTP does=0A> =
not.=0A> =0A> Also BEEP's per-channel setup latency is rather large: one=0A=
> redundant=0A> round trip.=A0 Using BEEP for a simple HTTP request=0A> wou=
ld take noticably=0A> longer than HTTP itself on high latency networks.=0A=
=0A=0AThe usecases discussed are not one request/response per socket case, =
but a significantly longer (in time) channel - so the setup overhead mentio=
ned would not be that significant.=0AFor simple request/response, http is f=
ine - for really really simple (few bytes) request/response, even http can =
be quite an overhead, but that is not the point :-)=0A=0A> =0A> > (not to m=
ention support for flow control which becomes=0A> immediately=0A> > necessa=
ry in bi-directional streams).=0A> =0A> BEEP's flow control is sensible, bu=
t that's not the reason=0A> for it.=0A> =0A> For bi-directional streams, TC=
P/IP's built in flow control=0A> is enough.=0A=0AI think you are assuming a=
 request/response modal - where in server does not 'read' next request unti=
l previous one has been processed and response sent ?=0A=0A=0AConsider the =
following needs :=0A=0Aa) Multiple widgets talking to server and vice versa=
 - so, hopefully, sharing connection. (else the cost if really high in term=
s of number of connections, etc)=0Ab) Not limit to just request/response ba=
sed modal, but also including event/notifications which dont require an ack=
..=0A=0A=0ANot to mention corner case handling like error detection/recovery=
 due to abrupt connection loss, etc.=0A=0A=0A=0A> =0A> If you imagine a bi-=
directional HTTP (easy because requests=0A> and=0A> responses are easily di=
stinguished), it does not need any=0A> BEEP-style=0A> flow control.=A0 (I h=
ave implemented this).=0A=0APlease see above.=0A=0A> =0A> BEEP-style per ch=
annel flow control is needed because of=0A> its ability=0A> to multiplex mu=
ltiple channels in the _same_ direction on a=0A> single=0A> TCP/IP stream.=
=A0 Without per-channel flow control you=0A> get deadlocks or=0A> infinite =
buffering needed in some practical uses.=0A=0A=0ASee above on why this migh=
t be needed in HTTP case too.=0A=0A> =0A> Unfortunately BEEPs per channel f=
low control is sensible,=0A> but it's per=0A> channel setup is expensive bo=
th in needing XML and in=0A> network round=0A> trip time.=A0 It would be go=
od to keep the flow control=0A> but use smarter=0A> channel setup.=0A=0A=0A=
Use of XML, though sensible, is just part of BEEP framing.=0ANot sure why X=
ML parsing is an issue, but we can always replace it in other means to carr=
y metadata - for example, the meta-data getting exchanged is representable =
in HTTP header form too imo (*) - =0A=0Ahttp style headers (if required) \r=
\n=0Araw content=0A=0A=0A=0A(*) The headers could include both the framing =
rules and flow control rules. Since I was not involved in BEEP design, or n=
ot very intimately involved in using BEEP, not if this is a valid assertion=
 though.)=0A=0A> =0A> > Flip side would be, I am not sure if it is upgradab=
le=0A> from HTTP with=0A> > current infrastructure - other than through CON=
NECT=0A> that is.=0A> =0A> It's not, but all protocols discussed can be ove=
rlaid onto=0A> the current=0A> infrastructure using the HTTP hacks we use t=
oday (long GET=0A> etc.),=0A> while giving new proxies the chance to recogn=
ise the=0A> overlaid protocol=0A> and participate in it properly (dropping =
the hacks,=0A> proxying better).=0A> =0A> Imho it would be good to discuss =
how we can overlay any new=0A> protocol=0A> that we're talking about compat=
ibly onto the existing HTTP=0A> infrastructure.=0A=0A=0AYes, this might be =
the key issue - getting it to inter-operate might outweigh other considerat=
ions, atleast initially.=0AThis was probably the only reason why we stuck h=
ad used BOSH's long poll based approach to other possibilities.=0A=0ARegard=
s,=0AMridul=0A=0A=0A=0A> =0A> -- Jamie=0A> =0A=0A=0A      Add more friends =
to your messenger and enjoy! Go to http://messenger.yahoo.com/invite/


From jamie@shareable.org  Sun Apr 12 13:08:52 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3111F3A6C62 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 13:08:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.456
X-Spam-Level: 
X-Spam-Status: No, score=-5.456 tagged_above=-999 required=5 tests=[AWL=-2.857, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 8T5uM48VPBCF for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 13:08:51 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 268A03A6C00 for <hybi@ietf.org>; Sun, 12 Apr 2009 13:08:50 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lt601-0001kS-8c; Sun, 12 Apr 2009 21:09:53 +0100
Date: Sun, 12 Apr 2009 21:09:53 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mridul Muralidharan <mridulm80@yahoo.com>
Message-ID: <20090412200953.GG4394@shareable.org>
References: <605442.82172.qm@web95416.mail.in2.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <605442.82172.qm@web95416.mail.in2.yahoo.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 20:08:52 -0000

Mridul Muralidharan wrote:
> > Also BEEP's per-channel setup latency is rather large: one
> > redundant
> > round trip.  Using BEEP for a simple HTTP request
> > would take noticably
> > longer than HTTP itself on high latency networks.
> 
> The usecases discussed are not one request/response per socket case,
> but a significantly longer (in time) channel - so the setup overhead
> mentioned would not be that significant.

No, it's the latency per BEEP channel, not socket.  BEEP channel setup
involves a requests "give me a channel please [+XML profile]", and
response "here's your channel [+XML profile]".  Then you're allowed to
send your real request over the channel.

The trouble is you need a separate BEEP channel for every potentially
overlapping message.

> I think you are assuming a request/response modal - where in server does not 'read' next request until previous one has been processed and response sent ?

No, I'm not.  I'm firmly in the asynchronous messaging (interleaved
without arbitrary size limits and ACKs optional) camp :-)

> Consider the following needs :
> 
> a) Multiple widgets talking to server and vice versa - so,
hopefully, sharing connection. (else the cost if really high in terms
of number of connections, etc)
>
> b) Not limit to just request/response based modal, but also
including event/notifications which dont require an ack..

Yes, I'm talking about those sort of things, over a single socket.

With BEEP you need separate channels (not sockets) for every message
which may be multiplexed with other messages, whether there's a
response or not.  You can't start those BEEP channels just by using
them: first you have to request them all.  To multiplex messages
independently, ou don't know how many channels you need in advance, so
you request them on demand and that's the extra delay from BEEP.

Multiplexed HTTP proposals are much better: to start a new channel,
you just pick a new channel number and send a message using it.

> Not to mention corner case handling like error detection/recovery
> due to abrupt connection loss, etc.

I'm deeply familiar with such things.

> > If you imagine a bi-directional HTTP (easy because requests and
> > responses are easily distinguished), it does not need any
> > BEEP-style flow control.  (I have implemented this).
> 
> Please see above.

Agreed.  The point is interleaving (multiple in-flight independent
messages) is what makes per channel flow control needed.
Bi-directionality by itself doesn't, but bi-directionality by itself
is much less useful (like the RHTTP proposal).

> > BEEP-style per channel flow control is needed because of its
> > ability to multiplex multiple channels in the _same_ direction on
> > a single TCP/IP stream.  Without per-channel flow control you get
> > deadlocks or infinite buffering needed in some practical uses.
> 
> See above on why this might be needed in HTTP case too.

Yes it is, when we let messages be interleaved with each other.  It's
interleaving which makes the requirement for per channel flow control.

Interleaving, by the way, is important to support large messages (like
video files, say) which don't block small messages.  Without
interleaving we'd force each application developer to produce their
own splitting and sliding window protocol whenever they _might_ have a
large message, which is silly when we can do it in the protocol nearly
for free.

> > Unfortunately BEEPs per channel flow control is sensible,
> > but it's per
> > channel setup is expensive both in needing XML and in
> > network round
> > trip time.  It would be good to keep the flow control
> > but use smarter
> > channel setup.
>
> Use of XML, though sensible, is just part of BEEP framing.

BEEP framing is simple, a short bit of text, I like that.  I'm quite
surprised they followed tidy, sensible IETF text protocol traditions
in the framing and then used an XML envelope in one of the essential
protocol elements.

> Not sure why XML parsing is an issue,

Many devices which parse HTTP right now don't have XML parsers.  Think
HTTP proxies, HTTP firewalls, and client devices with 16kB of memory.
XML parsing is not hard, but it's not trivial either.  Having to skip
the optional DTD whose syntax is not precisely defined in the XML spec
does not help.  Having to accept multiple charsets in a basic ASCII
protocol does not help either.

It would be nice if new messaging protocol were understandable by such
devices too, and it's probably quite important that the basic framing
is understandable to proxies and firewalls.

> but we can always replace it in other means to carry metadata - for
example, the meta-data getting exchanged is representable in HTTP
header form too

I agree that HTTP header form or similar is better than an XML
envelope, for the basic protocol anyway.  Of course specific
applications are welcome to use XML, JSON etc. inside the protocol
envelope.
 
-- Jamie

From jamie@shareable.org  Sun Apr 12 14:31:51 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4A0713A680D for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 14:31:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.099
X-Spam-Level: 
X-Spam-Status: No, score=-5.099 tagged_above=-999 required=5 tests=[AWL=-2.500, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Tm9mMwgNNefW for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 14:31:50 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 597AF3A6803 for <hybi@ietf.org>; Sun, 12 Apr 2009 14:31:50 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lt7IP-00029S-7K; Sun, 12 Apr 2009 22:32:57 +0100
Date: Sun, 12 Apr 2009 22:32:57 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090412213257.GL4394@shareable.org>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com> <49E1518E.7050402@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49E1518E.7050402@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 21:31:51 -0000

Greg Wilkins wrote:
> Generally agreeing with all your points.  Yet the attributes you
> describe do sound pretty much like BEEP protocol - or at least
> beep framing.
> 
> BEEP is very much a HTTP style protocol with addition of
> bidirectionality, multiple response requests and multiple
> channels over a single connection.

Yes, although BEEPs channels are unnecessarily expensive (in latency)
to create on demand.

> If BEEP was extended with a few default assumptions (eg
> UTF-8 text body if no content type specified), then it would
> approach the byte efficiency of websocket, while preserving
> the extensibility and flexibility of HTTP.
> 
> If the security and authentication aspects of BEEP were
> stripped, then a pretty simply protocol would remain.
> 
> IF I was forced to pick a winner.... then I'd certainly
> look very closely at BEEP or a BEEP derivative.

I've looked closely at BEEP for this, and I agree, BEEP has some very
good things which should be copied.

Unfortunately its channel setup is expensive (latency, not bandwidth),
and often you'd want a channel per message in flight.  That forces you
to either accept the channel setup latency, or queue messages behind
each other in the same way that HTTP pipelining forces messages to
block behind each other.  Neither is ideal.

There is a BEEP extension to create sub-channels on demand without
having to request one, which is closer to ideal behaviour when you
have a new message to send.  But it unnecessarily constrains the
ordering of messages.  (I asked the author to relax this, but he
didn't see any point for his applications and suggested there are
other protocols to choose from if I need that.  I don't see the point
in the constraint.)

What I think would be ideal, and it's been mentioned already as
"asynchronous HTTP", is:

    HTTP or simplified-HTTP message headers and body.
    Bidirectional - either side can send requests and responses.
    Channel id attached to each message.
    Chunked encoding, with channel id attached to each chunk.
    Per channel flow control exactly like BEEP over TCP profile.
    Request/response id attached to each message.
    Request/response id matches responses out of order to requests.
    Pure messages with no response allowed, using "no-response" r/r id.
    All the above features are optional.

When you make those features optional, it degrades gracefully to
HTTP pipelining, then to HTTP persistent connections.

Also by making features optional, it is much easier to implement just
the subset of features you need, or plug in a feature-complete
implementation when you need all of them.

That provides the message/event protocol some people are asking for,
and also provides the request/response protocol some people are asking
for, does both bidirectionally, and over a single socket, and with
multiple applications sharing the same socket if they want.

It's efficient too, if you regard HTTP as efficient.  They're about
the same overhead.  Latency is sometimes better than HTTP, because it
doesn't have pipeline blocking like HTTP.  Proxy buffering
requirements are about the same as HTTP, in particular they are
bounded similarly.

-- Jamie

From jamie@shareable.org  Sun Apr 12 14:42:34 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 89E8F3A698C for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 14:42:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.821
X-Spam-Level: 
X-Spam-Status: No, score=-4.821 tagged_above=-999 required=5 tests=[AWL=-2.222, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cLOSDH-fNfgA for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 14:42:33 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 9BE403A680D for <hybi@ietf.org>; Sun, 12 Apr 2009 14:42:33 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lt7Sp-0002Db-3x; Sun, 12 Apr 2009 22:43:43 +0100
Date: Sun, 12 Apr 2009 22:43:43 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mark Lentczner <markl@lindenlab.com>
Message-ID: <20090412214343.GM4394@shareable.org>
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 21:42:34 -0000

Mark Lentczner wrote:
> Then I mis-replied. I should have said that we have a need for full- 
> duplex request-response -- either side can spontaneously send a  
> request message, and either side can, when ready, reply to a prior  
> request.

I agree.  HTTP requests and responses are easily distinguished, so
it's quite simple to accept either.

Recovery behaviour when the connection disappears is tricky if you
have HTTP-like requests.

Right now, clients tend to send non-idempotent requests (i.e. POST) on
a new connection, because if you send to a persistent connection, and
the socket reports an error, you don't know if the request was
received and processed and then the network died, or if the server
timed out the persistent connection and closed it while you were
sending the request.  You can't tell the difference, and it's unsafe
to repeat the request, so you use a new connection.

With a bidirectional connection, typically the initiating client sets
everything up, then the server is able to send a request - but this is
analogous to sending a request over a persistent HTTP connection.

In general with this sort of network protocol, you define a recovery
process - or assume the system will recover naturally - or don't have
non-idempotent requests coming from the server end.  But if we're
claiming to support HTTP in both directions, then we need to define if
the server is allowed to send POST requests and such to the client,
and what to do realistically on seeing network errors.

> Only the establishment in WebSocket is addressed via URL.  In rHTTP  
> each request can be addressed by URL.  This is useful as it is a very  
> common piece of meta-data that applications want. For example, using  
> WebSockets, for any complex page, an application is almost certainly  
> going to want a way to address which widget on the page is the  
> intended recipient of the event. Hence, each application is going to  
> have to devise an addressing scheme, and more importantly, a method of  
> encoding that address in the simple string event.

I agree that the URL address space is a very common useful piece of
metadata.  There is so much built around it now; it makes sense to
provide applications with that.

I also agree that HTTP message frame parsing is readily available, and
easy to implement when it isn't.

-- Jamie

From mjs@apple.com  Sun Apr 12 15:03:38 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7F0A13A6C3F for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:03:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BtkoyhboNA8v for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:03:37 -0700 (PDT)
Received: from mail-out4.apple.com (mail-out4.apple.com [17.254.13.23]) by core3.amsl.com (Postfix) with ESMTP id D2EE53A6C0A for <hybi@ietf.org>; Sun, 12 Apr 2009 15:03:37 -0700 (PDT)
Received: from relay13.apple.com (relay13.apple.com [17.128.113.29]) by mail-out4.apple.com (Postfix) with ESMTP id 30B6F5EC5E28 for <hybi@ietf.org>; Sun, 12 Apr 2009 15:04:48 -0700 (PDT)
Received: from relay13.apple.com (unknown [127.0.0.1]) by relay13.apple.com (Symantec Brightmail Gateway) with ESMTP id 1C75B28085 for <hybi@ietf.org>; Sun, 12 Apr 2009 15:04:48 -0700 (PDT)
X-AuditID: 1180711d-a96f6bb000000259-63-49e2657f2e6d
Received: from gertie.apple.com (gertie.apple.com [17.151.62.15]) by relay13.apple.com (Apple SCV relay) with ESMTP id DEF3F2807D for <hybi@ietf.org>; Sun, 12 Apr 2009 15:04:47 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by gertie.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI000HY9DBZ4L20@gertie.apple.com> for hybi@ietf.org; Sun, 12 Apr 2009 15:04:47 -0700 (PDT)
Message-id: <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Rob Sayre <rsayre@mozilla.com>
In-reply-to: <49DEF171.4080506@mozilla.com>
Date: Sun, 12 Apr 2009 15:04:46 -0700
References: <49DEF171.4080506@mozilla.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 22:03:38 -0000

On Apr 10, 2009, at 12:12 AM, Rob Sayre wrote:

> Ian Hickson wrote:
>> Personally I am interested in "Minimal, Event", but I don't see any
>> reason why we wouldn't also have the other models.
>
> I am interested in "Minimal, Event" as well.

I like Minimal/Event too, although I am starting to think the  
WebSocket protocol may be a bit too minimal. I think adding an  
optional type field to the packet or to the initial negotiation would  
help independently developed client and server code coordinate, much  
in the way MIME types help HTTP clients and servers without the  
necessity to define every possible type up front or to couple a single  
client to a single server. Ethernet and IP include a protocol number  
in the packet, TCP includes a port number (associated with protocols  
by convention), and HTTP as mentioned has a Content-Type. Perhaps  
WebSocket could make use of the HTTP Content-Type to identify the type  
of WebSocket service being connected to, or else there could be a per- 
packet identifier, which would be leaning towards the "Message, Event"  
model.

Regards,
Maciej


From ian@hixie.ch  Sun Apr 12 15:33:58 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5A39F3A69B1 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:33:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.504
X-Spam-Level: 
X-Spam-Status: No, score=-3.504 tagged_above=-999 required=5 tests=[AWL=-0.905, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MuY6maNTnUZC for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:33:57 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 778753A68CD for <hybi@ietf.org>; Sun, 12 Apr 2009 15:33:57 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 5B822122541; Sun, 12 Apr 2009 15:35:06 -0700 (PDT)
Date: Sun, 12 Apr 2009 22:35:06 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Maciej Stachowiak <mjs@apple.com>
In-Reply-To: <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>
Message-ID: <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 22:33:58 -0000

On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
> 
> I like Minimal/Event too, although I am starting to think the WebSocket 
> protocol may be a bit too minimal. I think adding an optional type field 
> to the packet or to the initial negotiation would help independently 
> developed client and server code coordinate, much in the way MIME types 
> help HTTP clients and servers without the necessity to define every 
> possible type up front or to couple a single client to a single server. 
> [...] TCP includes a port number (associated with protocols by 
> convention)

WebSocket includes a path (from the ws:// URL), which is intended to be 
analogous to the TCP port number.

Does that address this particular point satisfactorily?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From mjs@apple.com  Sun Apr 12 15:38:50 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 009973A68A3 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:38:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0JYqCN+EDRfL for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:38:49 -0700 (PDT)
Received: from mail-out4.apple.com (mail-out4.apple.com [17.254.13.23]) by core3.amsl.com (Postfix) with ESMTP id 4B61F3A682A for <hybi@ietf.org>; Sun, 12 Apr 2009 15:38:49 -0700 (PDT)
Received: from relay14.apple.com (relay14.apple.com [17.128.113.52]) by mail-out4.apple.com (Postfix) with ESMTP id C47C15EC6E80 for <hybi@ietf.org>; Sun, 12 Apr 2009 15:39:59 -0700 (PDT)
Received: from relay14.apple.com (unknown [127.0.0.1]) by relay14.apple.com (Symantec Brightmail Gateway) with ESMTP id ADF4428083 for <hybi@ietf.org>; Sun, 12 Apr 2009 15:39:59 -0700 (PDT)
X-AuditID: 11807134-a5855bb000000ff0-66-49e26dbf6c38
Received: from et.apple.com (et.apple.com [17.151.62.12]) by relay14.apple.com (Apple SCV relay) with ESMTP id 8A9E12802F for <hybi@ietf.org>; Sun, 12 Apr 2009 15:39:59 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by et.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI000H0DEYML030@et.apple.com> for hybi@ietf.org; Sun, 12 Apr 2009 15:39:59 -0700 (PDT)
Message-id: <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Ian Hickson <ian@hixie.ch>
In-reply-to: <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>
Date: Sun, 12 Apr 2009 15:39:58 -0700
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 22:38:50 -0000

On Apr 12, 2009, at 3:35 PM, Ian Hickson wrote:

> On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
>>
>> I like Minimal/Event too, although I am starting to think the  
>> WebSocket
>> protocol may be a bit too minimal. I think adding an optional type  
>> field
>> to the packet or to the initial negotiation would help independently
>> developed client and server code coordinate, much in the way MIME  
>> types
>> help HTTP clients and servers without the necessity to define every
>> possible type up front or to couple a single client to a single  
>> server.
>> [...] TCP includes a port number (associated with protocols by
>> convention)
>
> WebSocket includes a path (from the ws:// URL), which is intended to  
> be
> analogous to the TCP port number.
>
> Does that address this particular point satisfactorily?

I'm not sure the path is a good choice for identifying the type of  
service available. HTTP has both a path and a type. The path is sent  
by the client and identifies the specific resource. The type is sent  
by the server and directs how the resource is to be interpreted. I do  
not think we can expect, for example, multiple chat services to use  
the same path format, but there may be multiple chat services which  
choose to use the same protocol-by-convention over WebSocket. Right  
now, there doesn't seem to be a good way to give this information to  
the client, in the same way that any arbitrary URL can tell the client  
that its contents are to be interpreted as XML or PDF or plain text.

Regards,
Maciej


From ian@hixie.ch  Sun Apr 12 15:59:20 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A0A983A6C16 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:59:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.483
X-Spam-Level: 
X-Spam-Status: No, score=-3.483 tagged_above=-999 required=5 tests=[AWL=-0.884, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZIxAUK-AOMvO for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 15:59:19 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id B4C9D3A682A for <hybi@ietf.org>; Sun, 12 Apr 2009 15:59:19 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 48448125C37; Sun, 12 Apr 2009 16:00:30 -0700 (PDT)
Date: Sun, 12 Apr 2009 23:00:30 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Maciej Stachowiak <mjs@apple.com>
In-Reply-To: <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com>
Message-ID: <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 22:59:20 -0000

On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
> > 
> > WebSocket includes a path (from the ws:// URL), which is intended to 
> > be analogous to the TCP port number.
> 
> I'm not sure the path is a good choice for identifying the type of 
> service available. HTTP has both a path and a type. The path is sent by 
> the client and identifies the specific resource. The type is sent by the 
> server and directs how the resource is to be interpreted. I do not think 
> we can expect, for example, multiple chat services to use the same path 
> format, but there may be multiple chat services which choose to use the 
> same protocol-by-convention over WebSocket. Right now, there doesn't 
> seem to be a good way to give this information to the client, in the 
> same way that any arbitrary URL can tell the client that its contents 
> are to be interpreted as XML or PDF or plain text.

I don't understand why a 16bit integer solves the problem for TCP, but a 
string doesn't solve the problem for WebSocket. They seem exactly 
equivalent to me -- they identify the target of the connection, and by 
convention, the protocol that that target supports.

Why couldn't we exect multiple chat services to use the same path for 
their chat server, in the same way that multiple chat services written on 
TCP use the same port?

There's also plenty of precedent on the Web of including type information 
in the URL -- so much so, in fact, that some people find the path 
(specifically the "extension" part of the "filename" component of the 
path) more reliable than the explicit Content-Type data. There's also 
"well-known paths" like robots.txt, which, while suboptimal from HTTP's 
point of view, seem like they'd fit in well with WebSocket.

It's not really clear to me how we would expose the information in the API 
if we were to use something other than the path. It's not like most 
authors are going to care, or check to see what the remote end thinks it 
is speaking -- they'll connect to their own servers, or to well-known 
remote third-party servers, and just talk the protocol that that endpoint 
supports. The path is convenient in that we already have it and it seems 
like it's exactly where authors would put the distinction (just as they do 
today on the Web -- content negotiation is used far less than just 
different URLs for different types, after all).

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Sun Apr 12 16:25:30 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AF0233A6811 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 16:25:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.508
X-Spam-Level: 
X-Spam-Status: No, score=-2.508 tagged_above=-999 required=5 tests=[AWL=0.091,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id c4Cxm5ipRskK for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 16:25:30 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.189]) by core3.amsl.com (Postfix) with ESMTP id 8E0D93A6BA7 for <hybi@ietf.org>; Sun, 12 Apr 2009 16:25:29 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 24so1783927tim.14 for <hybi@ietf.org>; Sun, 12 Apr 2009 16:26:38 -0700 (PDT)
Received: by 10.110.33.15 with SMTP id g15mr8135721tig.1.1239578798542; Sun, 12 Apr 2009 16:26:38 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id j5sm3357999tid.21.2009.04.12.16.26.35 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sun, 12 Apr 2009 16:26:37 -0700 (PDT)
Message-ID: <49E278A6.20709@webtide.com>
Date: Mon, 13 Apr 2009 09:26:30 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49DEF171.4080506@mozilla.com>	<A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>	<Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>	<71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 23:25:30 -0000

Ian Hickson wrote:

> I don't understand why a 16bit integer solves the problem for TCP, but a 
> string doesn't solve the problem for WebSocket. They seem exactly 
> equivalent to me -- they identify the target of the connection, and by 
> convention, the protocol that that target supports.

The big difference is that for the web you only have 2 strings
while TCP/IP has an infinite number of 16 bit integers.

ie The TCP/IP model is that if you want a ssh and 2 ftp conversations
then you open one connection to port 22 and 2 connections to port 21

On the web, we are limited (or should be limited) to 2 connections,
so you can't open an arbitrary number of websockets.

cheers

From mjs@apple.com  Sun Apr 12 16:35:07 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id ED21E3A6811 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 16:35:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CzkvBVWYBup3 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 16:35:07 -0700 (PDT)
Received: from mail-out3.apple.com (mail-out3.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id EF5FF3A691C for <hybi@ietf.org>; Sun, 12 Apr 2009 16:35:06 -0700 (PDT)
Received: from relay10.apple.com (relay10.apple.com [17.128.113.47]) by mail-out3.apple.com (Postfix) with ESMTP id 4D3C75BFE9CD for <hybi@ietf.org>; Sun, 12 Apr 2009 16:36:17 -0700 (PDT)
Received: from relay10.apple.com (unknown [127.0.0.1]) by relay10.apple.com (Symantec Brightmail Gateway) with ESMTP id 35F7928062 for <hybi@ietf.org>; Sun, 12 Apr 2009 16:36:17 -0700 (PDT)
X-AuditID: 1180712f-a4962bb0000012d3-00-49e27af0f961
Received: from et.apple.com (et.apple.com [17.151.62.12]) by relay10.apple.com (Apple SCV relay) with ESMTP id F08C32806C for <hybi@ietf.org>; Sun, 12 Apr 2009 16:36:16 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by et.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI000HBFHKGL030@et.apple.com> for hybi@ietf.org; Sun, 12 Apr 2009 16:36:16 -0700 (PDT)
Message-id: <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Ian Hickson <ian@hixie.ch>
In-reply-to: <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>
Date: Sun, 12 Apr 2009 16:36:15 -0700
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 12 Apr 2009 23:35:08 -0000

On Apr 12, 2009, at 4:00 PM, Ian Hickson wrote:

> On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
>>>
>>> WebSocket includes a path (from the ws:// URL), which is intended to
>>> be analogous to the TCP port number.
>>
>> I'm not sure the path is a good choice for identifying the type of
>> service available. HTTP has both a path and a type. The path is  
>> sent by
>> the client and identifies the specific resource. The type is sent  
>> by the
>> server and directs how the resource is to be interpreted. I do not  
>> think
>> we can expect, for example, multiple chat services to use the same  
>> path
>> format, but there may be multiple chat services which choose to use  
>> the
>> same protocol-by-convention over WebSocket. Right now, there doesn't
>> seem to be a good way to give this information to the client, in the
>> same way that any arbitrary URL can tell the client that its contents
>> are to be interpreted as XML or PDF or plain text.
>
> I don't understand why a 16bit integer solves the problem for TCP,  
> but a
> string doesn't solve the problem for WebSocket. They seem exactly
> equivalent to me -- they identify the target of the connection, and by
> convention, the protocol that that target supports.

HTTP is a better analogy here than TCP in my opinion. The fact that  
TCP only maps ports to protocols by convention is not so great - we  
have to go to extra design effort to prevent cross-protocol attacks as  
a result.

> Why couldn't we exect multiple chat services to use the same path for
> their chat server, in the same way that multiple chat services  
> written on
> TCP use the same port?

I think that would be a poor design in the same way as relying on the  
exact path of "/favicon.ico" or "/crossdomain.xml" is a poor design.  
For one thing, it would mean you can't host multiple chat services on  
the same host and port. That seems like an arbitrary limitation. Also,  
if your WebSocket service is running on the same host and port as your  
web service, you may not have a free choice of what path to use, since  
some may be taken.

> There's also plenty of precedent on the Web of including type  
> information
> in the URL -- so much so, in fact, that some people find the path
> (specifically the "extension" part of the "filename" component of the
> path) more reliable than the explicit Content-Type data. There's also
> "well-known paths" like robots.txt, which, while suboptimal from  
> HTTP's
> point of view, seem like they'd fit in well with WebSocket.

I think well-known paths are problematic and we should not design  
WebSockets around using it for everything. I cited two examples above  
where I believe you've stated on the record that well-known paths are  
a design problem. What is different about WebSocket that makes relying  
on well-known paths for everything a good idea?

It seems like extremely poor design to conflate the path (which  
identifies a specific resource) with the type (which identifies the  
*kind* of resource).

> It's not really clear to me how we would expose the information in  
> the API
> if we were to use something other than the path.

Exposing the information in the API would be trivial. And additional  
callback, or a property on the WebSocket object guaranteed to be set  
by the time the first message is received, would both work OK. By  
comparison, XMLHttpRequest provides for type checking by exposing the  
HTTP response headers.

> It's not like most authors are going to care, or check to see what  
> the remote end thinks it
> is speaking -- they'll connect to their own servers, or to well-known
> remote third-party servers, and just talk the protocol that that  
> endpoint
> supports. The path is convenient in that we already have it and it  
> seems
> like it's exactly where authors would put the distinction (just as  
> they do
> today on the Web -- content negotiation is used far less than just
> different URLs for different types, after all).

The rest of your argument is that most authors won't bother to check.  
It is true that most Web content authors don't currently check the  
type of XMLHttpRequest responses, except perhaps implicitly by using  
responseXML. However, until very recently there was no ability to  
connect to an unknown service via XMLHttpRequest since cross-domain  
support is a recent innovation that is still being deployed. Browsers  
check types all the time however (even though in various cases that  
information is not considered completely reliable). Browsers do not  
rely on hardcoded well-known paths to identify the type of the  
resource, and it's hard to imagine the Web working as well as it does  
if they did. One use case that we can foresee is a JavaScript library  
using WebSocket to connect to a chat service using a well-known  
protocol (perhaps someone will specify a protocol for Jabber over  
WebSocket). In such a case, it is much more likely that the protocol  
type will be checked than with hand-coded client code, since the  
library will want to report errors to the content author in a  
reasonable way rather than return garbage data.

Adding a protocol type would be a modest extension that is trivially  
added to the protocol and client-side API. It would add very little  
complexity to the case of connecting to a known service, but creates  
for the possibility of a sane way of connecting to unknown services.

Regards,
Maciej


From jamie@shareable.org  Sun Apr 12 17:10:18 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E7C2D3A6C75 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:10:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.599
X-Spam-Level: 
X-Spam-Status: No, score=-4.599 tagged_above=-999 required=5 tests=[AWL=-2.000, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BW4UQh7gmpeG for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:10:18 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 15B0B3A6C3E for <hybi@ietf.org>; Sun, 12 Apr 2009 17:10:18 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lt9li-00033V-1e; Mon, 13 Apr 2009 01:11:22 +0100
Date: Mon, 13 Apr 2009 01:11:22 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090413001121.GB10391@shareable.org>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com> <49E278A6.20709@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49E278A6.20709@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 00:10:19 -0000

Greg Wilkins wrote:
> >I don't understand why a 16bit integer solves the problem for TCP, but a 
> >string doesn't solve the problem for WebSocket. They seem exactly 
> >equivalent to me -- they identify the target of the connection, and by 
> >convention, the protocol that that target supports.
> 
> The big difference is that for the web you only have 2 strings
> while TCP/IP has an infinite number of 16 bit integers.
> 
> ie The TCP/IP model is that if you want a ssh and 2 ftp conversations
> then you open one connection to port 22 and 2 connections to port 21

That's separate connections, to the same service, not to separate
services.

If want to connect to different services on the same host, over TCP/IP
you typically need something like separate IP addresses (bad idea) or
a distinguishing "virtual host" header like HTTP's Host:.  Or
different ports, but then it's not a well known service and often
doesn't work over firewalls.

The analogous thing in WebSockets would be a path for the resource
being connected to, and a string/path to identify the service type.

However, I don't think there is really a need for the second string.

The first one is enough to identify both the resource being connected
to and the service type, in the usual ways that people assign
application-specific meanings to URLs.

-- Jamie

From ian@hixie.ch  Sun Apr 12 17:11:00 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DC36C3A6BAF for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:11:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.462
X-Spam-Level: 
X-Spam-Status: No, score=-3.462 tagged_above=-999 required=5 tests=[AWL=-0.863, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id P0PbXNI9A7VI for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:10:59 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id F3B8B3A6A2F for <hybi@ietf.org>; Sun, 12 Apr 2009 17:10:34 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id B0540122541; Sun, 12 Apr 2009 17:11:44 -0700 (PDT)
Date: Mon, 13 Apr 2009 00:11:43 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Maciej Stachowiak <mjs@apple.com>
In-Reply-To: <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com>
Message-ID: <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com> <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 00:11:00 -0000

On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
> > 
> > I don't understand why a 16bit integer solves the problem for TCP, but 
> > a string doesn't solve the problem for WebSocket. They seem exactly 
> > equivalent to me -- they identify the target of the connection, and by 
> > convention, the protocol that that target supports.
> 
> HTTP is a better analogy here than TCP in my opinion. The fact that TCP 
> only maps ports to protocols by convention is not so great - we have to 
> go to extra design effort to prevent cross-protocol attacks as a result.

WebSocket has built-in support for preventing cross-protocol attacks -- 
only trusted peers can send arbitrary data, everyone else is just 
disconnected before the server receives the data.

Why would HTTP be a better analogy? WebSocket really is just trying to be 
TCP for the Web. It doesn't really have anything in common with HTTP other 
than the handshake, and that's really a red herring.


> > Why couldn't we exect multiple chat services to use the same path for 
> > their chat server, in the same way that multiple chat services written 
> > on TCP use the same port?
> 
> I think that would be a poor design in the same way as relying on the 
> exact path of "/favicon.ico" or "/crossdomain.xml" is a poor design.

Fixed URLs are bad in HTTP because it causes people who don't know that 
the resources exist to try to obtain them. They're also a bad idea because 
HTTP is split across multiple users, but the fixed URLs are per-host.

With WebSocket the first point doesn't apply -- we aren't talking about 
people connecting to arbitrary hosts, we're talking about a convention by 
which someone can guess what protocol a server speaks.

The second point need not apply either, since we don't have to lock the 
entire path down, we can just say that the filename, or the extension, or 
whatever, is where you put the protocol you're willing to speak. So if you 
want to write a chat server that speaks WebSocketJabber, you can make it 
available at:

   ws://example.com/~jsmith/chat.wsj

...or some such.


> For one thing, it would mean you can't host multiple chat services on 
> the same host and port. That seems like an arbitrary limitation.

As shown above, this is not necessarily the case.


> Also, if your WebSocket service is running on the same host and port as 
> your web service, you may not have a free choice of what path to use, 
> since some may be taken.

That seems fine with just using the extension or filename part also.


> I think well-known paths are problematic and we should not design 
> WebSockets around using it for everything. I cited two examples above 
> where I believe you've stated on the record that well-known paths are a 
> design problem. What is different about WebSocket that makes relying on 
> well-known paths for everything a good idea?

What I'm advocating for is more like extensions that fixed paths, but even 
for fixed paths I don't believe the disadvantages they have in the realm 
of HTTP actually apply in the realm of WebSockets.


> It seems like extremely poor design to conflate the path (which 
> identifies a specific resource) with the type (which identifies the 
> *kind* of resource).

It's what pretty much every widely-used file system does. It's what 
authors understand. Content-Type has been pretty solidly shown to be a 
failed experiment, IMHO.


> > It's not really clear to me how we would expose the information in the 
> > API if we were to use something other than the path.
> 
> Exposing the information in the API would be trivial. And additional 
> callback, or a property on the WebSocket object guaranteed to be set by 
> the time the first message is received, would both work OK.

Current API:

   socket = new WebSocket('ws://example.com/chat.dcp');
   socket.onmessage = function (e) {
     if (e.data != 'DCP') {
       error('The server mysteriously stopped speaking DCP!');
       socket.disconnect();
     }
     socket.onmesage = chatHandler;
   }
   socket.postMessage('DCP');

Explicit API:

   socket = new WebSocket('ws://example.com/chat', 'DCP');
   socket.onprotocol = function (e) {
     if (e.data != 'DCP') {
       error('The server mysteriously stopped speaking DCP!');
       socket.disconnect();
     }
   }
   socket.onmesage = chatHandler;

It doesn't seem to gain us much.


> By comparison, XMLHttpRequest provides for type checking by exposing the 
> HTTP response headers. [...]
>
> It is true that most Web content authors don't currently check the type 
> of XMLHttpRequest responses, except perhaps implicitly by using 
> responseXML.

So why should we bother?


> However, until very recently there was no ability to 
> connect to an unknown service via XMLHttpRequest since cross-domain 
> support is a recent innovation that is still being deployed.

It's not like the people connecting to remote JSON resources using 
<script> ever considered that the remote script might not be JSON. They 
say "json" in the URL they connected to, and they assumed it would work.


> Browsers check types all the time however (even though in various cases 
> that information is not considered completely reliable). Browsers do not 
> rely on hardcoded well-known paths to identify the type of the resource, 
> and it's hard to imagine the Web working as well as it does if they did.

Browsers check types by sniffing more than anything, and rely on 
well-known paths quite a lot. I think it's a stretch to say that the check 
types "all the time". But this is academic, because the browsers are 
acting on the users' behalf, whereas these scripts are acting on the 
authors' behalf. As the subject line of this thread suggests, it's an 
apples and oranges comparison.


> One use case that we can foresee is a JavaScript library using WebSocket 
> to connect to a chat service using a well-known protocol (perhaps 
> someone will specify a protocol for Jabber over WebSocket). In such a 
> case, it is much more likely that the protocol type will be checked than 
> with hand-coded client code, since the library will want to report 
> errors to the content author in a reasonable way rather than return 
> garbage data.

Why is a trivial handshake not enough for this?


> Adding a protocol type would be a modest extension that is trivially 
> added to the protocol and client-side API. It would add very little 
> complexity to the case of connecting to a known service, but creates for 
> the possibility of a sane way of connecting to unknown services.

I don't understand what value it adds.

For custom protocols, authors aren't going to want to worry about the 
details of an additional handshake after the one we require now. Why add 
anything at all, when you can do it all -- if you want it -- over the 
regular WebSocket channel?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Sun Apr 12 17:22:31 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7D4C13A6812 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:22:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.514
X-Spam-Level: 
X-Spam-Status: No, score=-2.514 tagged_above=-999 required=5 tests=[AWL=0.085,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HcBK9c6w1yiO for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:22:30 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id 94F663A6C2C for <hybi@ietf.org>; Sun, 12 Apr 2009 17:22:30 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 24so1791431tim.14 for <hybi@ietf.org>; Sun, 12 Apr 2009 17:23:40 -0700 (PDT)
Received: by 10.110.5.14 with SMTP id 14mr8138289tie.40.1239582220236; Sun, 12 Apr 2009 17:23:40 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm8321281tib.28.2009.04.12.17.23.37 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sun, 12 Apr 2009 17:23:39 -0700 (PDT)
Message-ID: <49E28603.8030407@webtide.com>
Date: Mon, 13 Apr 2009 10:23:31 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: hybi@ietf.org
References: <8A829FEA-0EF2-46D6-974D-0EB237FF2728@lindenlab.com> <Pine.LNX.4.62.0904090030580.19453@hixie.dreamhostps.com> <A66ED417-B7AB-4138-B8DE-31323A4DC0C1@lindenlab.com> <49E1518E.7050402@webtide.com> <20090412213257.GL4394@shareable.org>
In-Reply-To: <20090412213257.GL4394@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 00:22:31 -0000

Jamie,

excellent information and insight in this and your other messages.

I think it is very valuable to look at existing protocols like
BEEP and to identify/enumerate their deficiencies. Straw men
are useful (and others probably appreciate not having me
continue to use websocket for that purpose :)

I've tried to capture a lot of this recent conversation on the
wiki: http://trac.tools.ietf.org/bof/trac/wiki/HyBi

I've added a new section called "Candidates Requirements" where
I list/describe the attributes of a protocol that people have
identified as good (although there is no consensus if all are
indeed good and/or needed).

For each proposal I've listed some pros and cons, mostly
against candidate requirements.    It would be good if others
could sanitize those for any biases coming from me.


I've also added a "Asynchronous HTTP" proposal based on the
description you gave.   You allude to existing proposals for
such a thing - do you have references to those?



cheers

From gregw@webtide.com  Sun Apr 12 17:27:36 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CE9F63A6C25 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:27:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.518
X-Spam-Level: 
X-Spam-Status: No, score=-2.518 tagged_above=-999 required=5 tests=[AWL=0.081,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id vLM2EgpH-jt5 for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:27:36 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id DD5AC3A6812 for <hybi@ietf.org>; Sun, 12 Apr 2009 17:27:35 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 24so1792197tim.14 for <hybi@ietf.org>; Sun, 12 Apr 2009 17:28:45 -0700 (PDT)
Received: by 10.110.90.9 with SMTP id n9mr8185357tib.6.1239582525831; Sun, 12 Apr 2009 17:28:45 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id i9sm7751092tid.13.2009.04.12.17.28.42 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sun, 12 Apr 2009 17:28:44 -0700 (PDT)
Message-ID: <49E28734.4060706@webtide.com>
Date: Mon, 13 Apr 2009 10:28:36 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49DEF171.4080506@mozilla.com>	<A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>	<Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>	<71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com>	<Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com> <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com>
In-Reply-To: <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 00:27:36 -0000

Maciej Stachowiak wrote:

> HTTP is a better analogy here than TCP in my opinion. The fact that TCP 
> only maps ports to protocols by convention is not so great - we have to 
> go to extra design effort to prevent cross-protocol attacks as a result.


Good point.

I think the limitation for websocket is that the URL
is associated with the connection, while for HTTP it is associated
with the request.

cheers


From mjs@apple.com  Sun Apr 12 17:39:11 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A23E43A6C7E for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:39:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id z4nLX7g1mTok for <hybi@core3.amsl.com>; Sun, 12 Apr 2009 17:39:09 -0700 (PDT)
Received: from mail-out3.apple.com (mail-out3.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id B500B3A6A6E for <hybi@ietf.org>; Sun, 12 Apr 2009 17:39:09 -0700 (PDT)
Received: from relay11.apple.com (relay11.apple.com [17.128.113.48]) by mail-out3.apple.com (Postfix) with ESMTP id 4BFCD5C006B6 for <hybi@ietf.org>; Sun, 12 Apr 2009 17:40:20 -0700 (PDT)
Received: from relay11.apple.com (unknown [127.0.0.1]) by relay11.apple.com (Symantec Brightmail Gateway) with ESMTP id 349C728091 for <hybi@ietf.org>; Sun, 12 Apr 2009 17:40:20 -0700 (PDT)
X-AuditID: 11807130-a708ebb000000fcd-7d-49e289f3b8b9
Received: from gertie.apple.com (gertie.apple.com [17.151.62.15]) by relay11.apple.com (Apple SCV relay) with ESMTP id EBC822808D for <hybi@ietf.org>; Sun, 12 Apr 2009 17:40:19 -0700 (PDT)
MIME-version: 1.0
Content-type: multipart/alternative; boundary="Boundary_(ID_+yjeiVaPsanU/BluJcc86Q)"
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by gertie.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI000HSEKJ74L30@gertie.apple.com> for hybi@ietf.org; Sun, 12 Apr 2009 17:40:19 -0700 (PDT)
Message-id: <C0524F14-6C1C-4B33-89D8-4C558B66B5CE@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Ian Hickson <ian@hixie.ch>
In-reply-to: <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>
Date: Sun, 12 Apr 2009 17:40:18 -0700
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com> <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com> <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 00:39:11 -0000

--Boundary_(ID_+yjeiVaPsanU/BluJcc86Q)
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-transfer-encoding: 7BIT


On Apr 12, 2009, at 5:11 PM, Ian Hickson wrote:

> On Sun, 12 Apr 2009, Maciej Stachowiak wrote:
>>>
>>> I don't understand why a 16bit integer solves the problem for TCP,  
>>> but
>>> a string doesn't solve the problem for WebSocket. They seem exactly
>>> equivalent to me -- they identify the target of the connection,  
>>> and by
>>> convention, the protocol that that target supports.
>>
>> HTTP is a better analogy here than TCP in my opinion. The fact that  
>> TCP
>> only maps ports to protocols by convention is not so great - we  
>> have to
>> go to extra design effort to prevent cross-protocol attacks as a  
>> result.
>
> WebSocket has built-in support for preventing cross-protocol attacks  
> --
> only trusted peers can send arbitrary data, everyone else is just
> disconnected before the server receives the data.
>
> Why would HTTP be a better analogy? WebSocket really is just trying  
> to be
> TCP for the Web. It doesn't really have anything in common with HTTP  
> other
> than the handshake, and that's really a red herring.

WebSocket is quite different from TCP, since it has defined message  
boundaries whereas TCP is just a two-way byte stream. TCP is much  
lower level than HTTP or WebSocket, as it imposes no structure on the  
data stream at all. It is expected that protocols are built on top of  
TCP which provide the needed structure. I think HTTP is a good  
protocol to use as inspiration, identifying resources by URL and the  
server reporting their type has proven to be a very effective model  
for data resources and for request-response services on the Web, and I  
think it would be similarly effective for bidirectional messaging  
services. There is little reason to think that bidirectional messaging  
services are fundamentally different.


>>> Why couldn't we exect multiple chat services to use the same path  
>>> for
>>> their chat server, in the same way that multiple chat services  
>>> written
>>> on TCP use the same port?
>>
>> I think that would be a poor design in the same way as relying on the
>> exact path of "/favicon.ico" or "/crossdomain.xml" is a poor design.
>
> Fixed URLs are bad in HTTP because it causes people who don't know  
> that
> the resources exist to try to obtain them. They're also a bad idea  
> because
> HTTP is split across multiple users, but the fixed URLs are per-host.
>
> With WebSocket the first point doesn't apply -- we aren't talking  
> about
> people connecting to arbitrary hosts, we're talking about a  
> convention by
> which someone can guess what protocol a server speaks.

Or more importantly, verify that it is the protocol they expect rather  
than interpret the response as garbage.

> The second point need not apply either, since we don't have to lock  
> the
> entire path down, we can just say that the filename, or the  
> extension, or
> whatever, is where you put the protocol you're willing to speak. So  
> if you
> want to write a chat server that speaks WebSocketJabber, you can  
> make it
> available at:
>
>   ws://example.com/~jsmith/chat.wsj
>
> ...or some such.

Using part of the path to identify the type sounds acceptable to me,  
as long as we define what part of the path holds type information as  
part of the WebSocket protocol instead of leaving it completely ad-hoc.

>>
>> Exposing the information in the API would be trivial. And additional
>> callback, or a property on the WebSocket object guaranteed to be  
>> set by
>> the time the first message is received, would both work OK.
>
> Current API:
>
>   socket = new WebSocket('ws://example.com/chat.dcp');
>   socket.onmessage = function (e) {
>     if (e.data != 'DCP') {
>       error('The server mysteriously stopped speaking DCP!');
>       socket.disconnect();
>     }
>     socket.onmesage = chatHandler;
>   }
>   socket.postMessage('DCP');
>
> Explicit API:
>
>   socket = new WebSocket('ws://example.com/chat', 'DCP');
>   socket.onprotocol = function (e) {
>     if (e.data != 'DCP') {
>       error('The server mysteriously stopped speaking DCP!');
>       socket.disconnect();
>     }
>   }
>   socket.onmesage = chatHandler;
>
> It doesn't seem to gain us much.

Putting type information in-bad is not very reliable. How can you tell  
a server that speaks DCP from one that happens to send that string DCP  
as the first message and doesn't have a convention of sending the  
protocol type as the first message. This also seems to be completely  
different from your path convention suggestion. Which do you think is  
the right way for WebSocket subprotocol identification, path or in- 
band messages? How will subprotocol identifiers be registered?

>> By comparison, XMLHttpRequest provides for type checking by  
>> exposing the
>> HTTP response headers. [...]
>>
>> It is true that most Web content authors don't currently check the  
>> type
>> of XMLHttpRequest responses, except perhaps implicitly by using
>> responseXML.
>
> So why should we bother?

I explained below - we are offering the ability to connect to unknown  
services, something that XHR hasn't offered at all. Consider a  
WebSocket-based syndication service - it initally send you the current  
copy of the feed, then incremental updates as they come in. This would  
clearly be better in many cases than, say Atom over http + periodic  
polling. And it's the kind of service you would likely want to open to  
cross-domain access. But it seems pretty useful for Web-based feed  
readers that used the hypothetical Atom-over-WebSocket to tell that  
this is indeed what they are talking to, and to be able to distinguish  
it from Jabber-over-WebSocket or CalDAV-over-WebSocket instead of  
possibly displaying garbage data to the user. After all, the user  
typed in a random supposedly feed URL, and the reader can't just trust  
that to be right.

>> However, until very recently there was no ability to
>> connect to an unknown service via XMLHttpRequest since cross-domain
>> support is a recent innovation that is still being deployed.
>
> It's not like the people connecting to remote JSON resources using
> <script> ever considered that the remote script might not be JSON.  
> They
> say "json" in the URL they connected to, and they assumed it would  
> work.

Remotely embedding remote scripts in the hope that they are JSON is an  
example of a very bad way to do things that should not be replicated.

>
>> Browsers check types all the time however (even though in various  
>> cases
>> that information is not considered completely reliable). Browsers  
>> do not
>> rely on hardcoded well-known paths to identify the type of the  
>> resource,
>> and it's hard to imagine the Web working as well as it does if they  
>> did.
>
> Browsers check types by sniffing more than anything, and rely on
> well-known paths quite a lot. I think it's a stretch to say that the  
> check
> types "all the time". But this is academic, because the browsers are
> acting on the users' behalf, whereas these scripts are acting on the
> authors' behalf. As the subject line of this thread suggests, it's an
> apples and oranges comparison.

I can only think of one well-known path relied on by the browser I  
work on, that being favicon.ico. While in some cases sniffing is  
applied, in other cases the type information is authoritative. Most  
importantly, other than a few types considered to effectively mean  
"unknown", the type of the main resource is generally trusted.

>
>> One use case that we can foresee is a JavaScript library using  
>> WebSocket
>> to connect to a chat service using a well-known protocol (perhaps
>> someone will specify a protocol for Jabber over WebSocket). In such a
>> case, it is much more likely that the protocol type will be checked  
>> than
>> with hand-coded client code, since the library will want to report
>> errors to the content author in a reasonable way rather than return
>> garbage data.
>
> Why is a trivial handshake not enough for this?

It provides a positive network effect if everyone does the handshake  
the same way, thus it is preferable to have subprotocol negotiation  
built into the protocol. The cost of doing so is minimal. Even using  
part of the path would be adequate, although I think it is preferable  
for the server to state the protocol than for the client to request  
it. The latter allows for the possibility that the same service may  
change protocols, just as on the Web the same resource may change types.

>> Adding a protocol type would be a modest extension that is trivially
>> added to the protocol and client-side API. It would add very little
>> complexity to the case of connecting to a known service, but  
>> creates for
>> the possibility of a sane way of connecting to unknown services.
>
> I don't understand what value it adds.
>
> For custom protocols, authors aren't going to want to worry about the
> details of an additional handshake after the one we require now. Why  
> add
> anything at all, when you can do it all -- if you want it -- over the
> regular WebSocket channel?

Doing protocol negotiation in-band with everyone doing it differently  
is more error-prone than having a simple out-of-band mechanism. It  
could be as simple as an additional HTTP header in the WebSocket  
handshake response. I would consider in-band subprotocol handshakes to  
be a worse alternative than a defined path convention. Now, WebSocket  
could require the first messages exchanged to be subprotocol  
identifiers, and require something distinct from any subprotocol  
identifier (such as the empty string). But that seems worse in just  
about every way to defining out-of-band subprotocol identification.

Regards,
Maciej




--Boundary_(ID_+yjeiVaPsanU/BluJcc86Q)
Content-type: text/html; charset=US-ASCII
Content-transfer-encoding: quoted-printable

<html><body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; =
-webkit-line-break: after-white-space; "><br><div><div>On Apr 12, 2009, =
at 5:11 PM, Ian Hickson wrote:</div><br =
class=3D"Apple-interchange-newline"><blockquote type=3D"cite"><div>On =
Sun, 12 Apr 2009, Maciej Stachowiak wrote:<br><blockquote =
type=3D"cite"><blockquote =
type=3D"cite"><br></blockquote></blockquote><blockquote =
type=3D"cite"><blockquote type=3D"cite">I don't understand why a 16bit =
integer solves the problem for TCP, but =
<br></blockquote></blockquote><blockquote type=3D"cite"><blockquote =
type=3D"cite">a string doesn't solve the problem for WebSocket. They =
seem exactly <br></blockquote></blockquote><blockquote =
type=3D"cite"><blockquote type=3D"cite">equivalent to me -- they =
identify the target of the connection, and by =
<br></blockquote></blockquote><blockquote type=3D"cite"><blockquote =
type=3D"cite">convention, the protocol that that target =
supports.<br></blockquote></blockquote><blockquote =
type=3D"cite"><br></blockquote><blockquote type=3D"cite">HTTP is a =
better analogy here than TCP in my opinion. The fact that TCP =
<br></blockquote><blockquote type=3D"cite">only maps ports to protocols =
by convention is not so great - we have to <br></blockquote><blockquote =
type=3D"cite">go to extra design effort to prevent cross-protocol =
attacks as a result.<br></blockquote><br>WebSocket has built-in support =
for preventing cross-protocol attacks -- <br>only trusted peers can send =
arbitrary data, everyone else is just <br>disconnected before the server =
receives the data.<br><br>Why would HTTP be a better analogy? WebSocket =
really is just trying to be <br>TCP for the Web. It doesn't really have =
anything in common with HTTP other <br>than the handshake, and that's =
really a red =
herring.<br></div></blockquote><div><br></div><div>WebSocket is quite =
different from TCP, since it has defined message boundaries whereas TCP =
is just a two-way byte stream. TCP is much lower level than HTTP or =
WebSocket, as it imposes no structure on the data stream at all. It is =
expected that protocols are built on top of TCP which provide the needed =
structure.&nbsp;I think HTTP is a good protocol to use as inspiration, =
identifying resources by URL and the server reporting their type has =
proven to be a very effective model for data resources and for =
request-response services on the Web, and I think it would be similarly =
effective for bidirectional messaging services. There is little reason =
to think that bidirectional messaging services are fundamentally =
different.&nbsp;</div><br><br><blockquote type=3D"cite"><div><blockquote =
type=3D"cite"><blockquote type=3D"cite">Why couldn't we exect multiple =
chat services to use the same path for =
<br></blockquote></blockquote><blockquote type=3D"cite"><blockquote =
type=3D"cite">their chat server, in the same way that multiple chat =
services written <br></blockquote></blockquote><blockquote =
type=3D"cite"><blockquote type=3D"cite">on TCP use the same =
port?<br></blockquote></blockquote><blockquote =
type=3D"cite"><br></blockquote><blockquote type=3D"cite">I think that =
would be a poor design in the same way as relying on the =
<br></blockquote><blockquote type=3D"cite">exact path of "/favicon.ico" =
or "/crossdomain.xml" is a poor design.<br></blockquote><br>Fixed URLs =
are bad in HTTP because it causes people who don't know that <br>the =
resources exist to try to obtain them. They're also a bad idea because =
<br>HTTP is split across multiple users, but the fixed URLs are =
per-host.<br><br>With WebSocket the first point doesn't apply -- we =
aren't talking about <br>people connecting to arbitrary hosts, we're =
talking about a convention by <br>which someone can guess what protocol =
a server speaks.<br></div></blockquote><div><br></div><div>Or more =
importantly, verify that it is the protocol they expect rather than =
interpret the response as garbage.</div><br><blockquote =
type=3D"cite"><div>The second point need not apply either, since we =
don't have to lock the <br>entire path down, we can just say that the =
filename, or the extension, or <br>whatever, is where you put the =
protocol you're willing to speak. So if you <br>want to write a chat =
server that speaks WebSocketJabber, you can make it <br>available =
at:<br><br> &nbsp;&nbsp;ws://example.com/~jsmith/chat.wsj<br><br>...or =
some such.<br></div></blockquote><div><br></div><div>Using part of the =
path to identify the type sounds acceptable to me, as long as we define =
what part of the path holds type information as part of the WebSocket =
protocol instead of leaving it completely =
ad-hoc.</div><div><br></div><blockquote type=3D"cite"><div><blockquote =
type=3D"cite"><font class=3D"Apple-style-span" =
color=3D"#000000"><br></font></blockquote><blockquote =
type=3D"cite">Exposing the information in the API would be trivial. And =
additional <br></blockquote><blockquote type=3D"cite">callback, or a =
property on the WebSocket object guaranteed to be set by =
<br></blockquote><blockquote type=3D"cite">the time the first message is =
received, would both work OK.<br></blockquote><br>Current API:<br><br> =
&nbsp;&nbsp;socket =3D new WebSocket('ws://example.com/chat.dcp');<br> =
&nbsp;&nbsp;socket.onmessage =3D function (e) {<br> =
&nbsp;&nbsp;&nbsp;&nbsp;if (e.data !=3D 'DCP') {<br> =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;error('The server mysteriously =
stopped speaking DCP!');<br> =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;socket.disconnect();<br> =
&nbsp;&nbsp;&nbsp;&nbsp;}<br> &nbsp;&nbsp;&nbsp;&nbsp;socket.onmesage =3D =
chatHandler;<br> &nbsp;&nbsp;}<br> =
&nbsp;&nbsp;socket.postMessage('DCP');<br><br>Explicit API:<br><br> =
&nbsp;&nbsp;socket =3D new WebSocket('ws://example.com/chat', =
'DCP');<br> &nbsp;&nbsp;socket.onprotocol =3D function (e) {<br> =
&nbsp;&nbsp;&nbsp;&nbsp;if (e.data !=3D 'DCP') {<br> =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;error('The server mysteriously =
stopped speaking DCP!');<br> =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;socket.disconnect();<br> =
&nbsp;&nbsp;&nbsp;&nbsp;}<br> &nbsp;&nbsp;}<br> =
&nbsp;&nbsp;socket.onmesage =3D chatHandler;<br><br>It doesn't seem to =
gain us much.<br></div></blockquote><div><br></div><div>Putting type =
information in-bad is not very reliable. How can you tell a server that =
speaks DCP from one that happens to send that string DCP as the first =
message and doesn't have a convention of sending the protocol type as =
the first message. This also seems to be completely different from your =
path convention suggestion. Which do you think is the right way for =
WebSocket subprotocol identification, path or in-band messages? How will =
subprotocol identifiers be registered?</div><br><blockquote =
type=3D"cite"><div><blockquote type=3D"cite">By comparison, =
XMLHttpRequest provides for type checking by exposing the =
<br></blockquote><blockquote type=3D"cite">HTTP response headers. =
[...]<br></blockquote><blockquote =
type=3D"cite"><br></blockquote><blockquote type=3D"cite">It is true that =
most Web content authors don't currently check the type =
<br></blockquote><blockquote type=3D"cite">of XMLHttpRequest responses, =
except perhaps implicitly by using <br></blockquote><blockquote =
type=3D"cite">responseXML.<br></blockquote><br>So why should we =
bother?<br></div></blockquote><div><br></div><div>I explained below - we =
are offering the ability to connect to unknown services, something that =
XHR hasn't offered at all. Consider a WebSocket-based syndication =
service - it initally send you the current copy of the feed, then =
incremental updates as they come in. This would clearly be better in =
many cases than, say Atom over http + periodic polling. And it's the =
kind of service you would likely want to open to cross-domain access. =
But it seems pretty useful for Web-based feed readers that used the =
hypothetical Atom-over-WebSocket to tell that this is indeed what they =
are talking to, and to be able to distinguish it from =
Jabber-over-WebSocket or CalDAV-over-WebSocket instead of possibly =
displaying garbage data to the user. After all, the user typed in a =
random supposedly feed URL, and the reader can't just trust that to be =
right.&nbsp;</div><br><blockquote type=3D"cite"><div><blockquote =
type=3D"cite">However, until very recently there was no ability to =
<br></blockquote><blockquote type=3D"cite">connect to an unknown service =
via XMLHttpRequest since cross-domain <br></blockquote><blockquote =
type=3D"cite">support is a recent innovation that is still being =
deployed.<br></blockquote><br>It's not like the people connecting to =
remote JSON resources using <br>&lt;script&gt; ever considered that the =
remote script might not be JSON. They <br>say "json" in the URL they =
connected to, and they assumed it would =
work.<br></div></blockquote><div><br></div><div>Remotely embedding =
remote scripts in the hope that they are JSON is an example of a very =
bad way to do things that should not be replicated.</div><br><blockquote =
type=3D"cite"><div><br><blockquote type=3D"cite">Browsers check types =
all the time however (even though in various cases =
<br></blockquote><blockquote type=3D"cite">that information is not =
considered completely reliable). Browsers do not =
<br></blockquote><blockquote type=3D"cite">rely on hardcoded well-known =
paths to identify the type of the resource, <br></blockquote><blockquote =
type=3D"cite">and it's hard to imagine the Web working as well as it =
does if they did.<br></blockquote><br>Browsers check types by sniffing =
more than anything, and rely on <br>well-known paths quite a lot. I =
think it's a stretch to say that the check <br>types "all the time". But =
this is academic, because the browsers are <br>acting on the users' =
behalf, whereas these scripts are acting on the <br>authors' behalf. As =
the subject line of this thread suggests, it's an <br>apples and oranges =
comparison.<br></div></blockquote><div><br></div><div>I can only think =
of one well-known path relied on by the browser I work on, that being =
favicon.ico. While in some cases sniffing is applied, in other cases the =
type information is authoritative. Most importantly, other than a few =
types considered to effectively mean "unknown", the type of the main =
resource is generally trusted.</div><br><blockquote =
type=3D"cite"><div><br><blockquote type=3D"cite">One use case that we =
can foresee is a JavaScript library using WebSocket =
<br></blockquote><blockquote type=3D"cite">to connect to a chat service =
using a well-known protocol (perhaps <br></blockquote><blockquote =
type=3D"cite">someone will specify a protocol for Jabber over =
WebSocket). In such a <br></blockquote><blockquote type=3D"cite">case, =
it is much more likely that the protocol type will be checked than =
<br></blockquote><blockquote type=3D"cite">with hand-coded client code, =
since the library will want to report <br></blockquote><blockquote =
type=3D"cite">errors to the content author in a reasonable way rather =
than return <br></blockquote><blockquote type=3D"cite">garbage =
data.<br></blockquote><br>Why is a trivial handshake not enough for =
this?<br></div></blockquote><div><br></div><div>It provides a positive =
network effect if everyone does the handshake the same way, thus it is =
preferable to have subprotocol negotiation built into the protocol. The =
cost of doing so is minimal. Even using part of the path would be =
adequate, although I think it is preferable for the server to state the =
protocol than for the client to request it. The latter allows for the =
possibility that the same service may change protocols, just as on the =
Web the same resource may change types.</div><br><blockquote =
type=3D"cite"><div><blockquote type=3D"cite">Adding a protocol type =
would be a modest extension that is trivially =
<br></blockquote><blockquote type=3D"cite">added to the protocol and =
client-side API. It would add very little <br></blockquote><blockquote =
type=3D"cite">complexity to the case of connecting to a known service, =
but creates for <br></blockquote><blockquote type=3D"cite">the =
possibility of a sane way of connecting to unknown =
services.<br></blockquote><br>I don't understand what value it =
adds.<br><br>For custom protocols, authors aren't going to want to worry =
about the <br>details of an additional handshake after the one we =
require now. Why add <br>anything at all, when you can do it all -- if =
you want it -- over the <br>regular WebSocket =
channel?<br></div></blockquote></div><br><div>Doing protocol negotiation =
in-band with everyone doing it differently is more error-prone than =
having a simple out-of-band mechanism. It could be as simple as an =
additional HTTP header in the WebSocket handshake response. I would =
consider in-band subprotocol handshakes to be a worse alternative than a =
defined path convention. Now, WebSocket could require the first messages =
exchanged to be subprotocol identifiers, and require something distinct =
from any subprotocol identifier (such as the empty string). But that =
seems worse in just about every way to defining out-of-band subprotocol =
identification.</div><div><br></div><div>Regards,</div><div>Maciej</div><d=
iv><br></div><div><br></div><div><br></div></body></html>=

--Boundary_(ID_+yjeiVaPsanU/BluJcc86Q)--

From markl@lindenlab.com  Mon Apr 13 09:45:00 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1F6A13A68B8 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 09:45:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.549
X-Spam-Level: 
X-Spam-Status: No, score=-3.549 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZKo6WASvh85p for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 09:44:59 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 778BF3A6DB0 for <hybi@ietf.org>; Mon, 13 Apr 2009 09:44:59 -0700 (PDT)
Received: from nil.lindenlab.com (nil.lindenlab.com [10.1.16.4]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 3124D3DBC06C for <hybi@ietf.org>; Mon, 13 Apr 2009 09:46:10 -0700 (PDT)
Message-Id: <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Mon, 13 Apr 2009 09:46:09 -0700
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>
X-Mailer: Apple Mail (2.930.3)
Subject: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 16:45:00 -0000

Ian Hickson wrote:
>  Personally I am interested in "Minimal, Event", ...


Rob Sayre wrote:
> I am interested in "Minimal, Event" as well.


John Fallows wrote:
> I am also primarily interested in "Minimal, Event".

Anne van Kesteren wrote:
> Me too.

....

Okay, I see that there is quite a bit of interest in the Minimal,  
Event space. I'm see how this comes from the view of what most Comet/ 
AJAX techniques are doing today on the web: Getting some piece of  
"update" information into a web-page that a user is viewing.

What I'm wondering is: How does one, given a Minimal, Event style  
protocol, handle when there are two or more widgets on the page that  
need updates? The home page view on Facebook, for example, has at  
least four separate areas being live-updated.

	- Mark

Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From jamie@shareable.org  Mon Apr 13 10:23:06 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C4B2D3A6D36 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 10:23:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.417
X-Spam-Level: 
X-Spam-Status: No, score=-4.417 tagged_above=-999 required=5 tests=[AWL=-1.818, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tqMIukNOvmwL for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 10:23:06 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 71BEB3A69AE for <hybi@ietf.org>; Mon, 13 Apr 2009 10:23:03 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LtPtE-0008MZ-Vz; Mon, 13 Apr 2009 18:24:12 +0100
Date: Mon, 13 Apr 2009 18:24:12 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mark Lentczner <markl@lindenlab.com>
Message-ID: <20090413172412.GA31677@shareable.org>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 17:23:06 -0000

Mark Lentczner wrote:
> What I'm wondering is: How does one, given a Minimal, Event style  
> protocol, handle when there are two or more widgets on the page that  
> need updates? The home page view on Facebook, for example, has at  
> least four separate areas being live-updated.

I imagine most systems use a multiple-subscriber pattern:

    The web client sends 4 comet-style requests for updates (one per
    widget), sometimes combining those into a single HTTP request
    carrying 4 comet-style encoded requests, sometimes using 4
    separate HTTP requests.  At most one of the HTTP requests will be
    a "long GET", waiting for updates to any of the widgets, or the
    long GET may be a separate request.

And I imagine a few use a single-subscriber pattern:

    The web client sends 1 request for updates for all the widgets,
    with the web server knowing which 4 widgets are present on that
    page.  That request is probably the same "long GET" (or long POST)
    which waits for updates to any of the widgets.

In both cases, events come back over a single long GET/POST (or
polling request) and are dispatched to the appropriate widget, by
whatever ad hoc addressing scheme.

For my own pages I'm combining multiple "group subscriptions" into a
hierarchy of address/etag pairs.  One compact request is enough for
the whole page if the server has that state in its group-subscription
cache, but it drops to encoding multiple requests when the server's
cache no longer has that information, or when the page is dynamically
modified to contain a different set of widgets.

My scheme is efficient when you have hundreds/thousands of widgets,
such as a table of information where every cell is individually
tracking some information, a screen full of sliders, and line/bar
graphs where you might update just a subset of the data from time to
time.  The etags allow the data to be present in the initially fetched
page and updated only when needed.

-- Jamie

From markl@lindenlab.com  Mon Apr 13 11:12:14 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6CCA93A68F3 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 11:12:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.552
X-Spam-Level: 
X-Spam-Status: No, score=-3.552 tagged_above=-999 required=5 tests=[AWL=0.047,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id dFnvKjl8jupi for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 11:12:13 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 8CD4C3A68B6 for <hybi@ietf.org>; Mon, 13 Apr 2009 11:12:13 -0700 (PDT)
Received: from nil.lindenlab.com (nil.lindenlab.com [10.1.16.4]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id BDB9B3DBC06C for <hybi@ietf.org>; Mon, 13 Apr 2009 11:13:24 -0700 (PDT)
Message-Id: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Mon, 13 Apr 2009 11:13:24 -0700
X-Mailer: Apple Mail (2.930.3)
Subject: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 18:12:14 -0000

On Apr 11, 2009, at 7:27 PM, Greg Wilkins wrote:
> BEEP!
> ...
> BEEP!
> ...
> BEEP!
> ...
> BEEP!

As pointed out, BEEP has many attributes that make is seem a likely  
candidate for this space.

But, to summarize (and add to) the points against it:

Seems that each request/response would need its own BEEP channel. This  
is because of the synchronous nature of BEEP channels(*). Channel  
creation is mid-weight, but at least requires a round-trip before the  
actual request could even begin. It might be possible to encode the  
request within the channel creation request, but there is a 4k payload  
limit for such messages.

BEEP requires both XML and MIME parsers. Note that BEEP messages are  
not equivalent to HTTP messages.  BEEP has no address (URL) or method  
for the messages, and the headers of BEEP are MIME, whereas the  
headers of HTTP are not. One could imagine using the message/http MIME  
type to an HTTP message as the payload within a BEEP message. But then  
you'd need XML, MIME and HTTP parsers... This personally doesn't  
bother me, but it is getting rather heavyweight for something small.

There is no addressability in BEEP for channels or messages. I see two  
ways two add one to a profile: 1) Add the address (and perhaps method)  
as fields in the initialization message on channel create. This would  
require a new channel for every request. 2) Encode the address in each  
message (see above for using message/http)

Some unclear aspects:

The overhead of BEEP seems like it would greater, though I admit I  
haven't calculated it. I doubt it would be outrageous.

I don't know enough to know if flow control internally would be  
needed. If so, BEEP has it.  If not, it is a little scary to implement!

I think one could shoe-horn into BEEP via Upgrade with a little  
careful draftsmanship. Defining a fall-back mechanism might be  
considerably more difficult, as BEEP is considerably more general.

Lastly, we did consider BEEP about a year ago. The biggest thing  
against it was that libraries didn't seem widely available. I realize  
there is a chicken and egg thing here, but since BEEP didn't appear to  
really hit our use case dead on, we didn't feel like we'd be the best  
candidates to encourage its library development.

	- Mark

(*) There is a current draft addressing this: http://tools.ietf.org/html/draft-thomson-beep-async-02


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From markl@lindenlab.com  Mon Apr 13 11:35:44 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EB1F53A6AF7 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 11:35:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.555
X-Spam-Level: 
X-Spam-Status: No, score=-3.555 tagged_above=-999 required=5 tests=[AWL=0.044,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Lj5KsS6aqMM7 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 11:35:44 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 285503A67DF for <hybi@ietf.org>; Mon, 13 Apr 2009 11:35:44 -0700 (PDT)
Received: from nil.lindenlab.com (nil.lindenlab.com [10.1.16.4]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 6BA803DBC06C for <hybi@ietf.org>; Mon, 13 Apr 2009 11:36:55 -0700 (PDT)
Message-Id: <C893F610-2839-4D5F-B13C-4CDC1D1293FF@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <20090413172412.GA31677@shareable.org>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Mon, 13 Apr 2009 11:36:55 -0700
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <20090413172412.GA31677@shareable.org>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 18:35:45 -0000

On Apr 13, 2009, at 10:24 AM, Jamie Lokier wrote:
> ... and are dispatched to the appropriate widget, by whatever ad hoc  
> addressing scheme.
>
> For my own pages I'm combining multiple "group subscriptions" into a  
> hierarchy of address/etag pairs.
> ... My scheme is efficient when you have hundreds/thousands of  
> widgets,


How would such a think work over the model of WebSockets, where all  
you get is a single string per event?

	- Mark


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From ian@hixie.ch  Mon Apr 13 12:40:02 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4B4063A6CD7 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 12:40:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.443
X-Spam-Level: 
X-Spam-Status: No, score=-3.443 tagged_above=-999 required=5 tests=[AWL=-0.844, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id JA87Hqb1mcZU for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 12:40:01 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 6B9B93A6819 for <hybi@ietf.org>; Mon, 13 Apr 2009 12:40:01 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 23C3827B96; Mon, 13 Apr 2009 12:41:11 -0700 (PDT)
Date: Mon, 13 Apr 2009 19:41:11 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Mark Lentczner <markl@lindenlab.com>
In-Reply-To: <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
Message-ID: <Pine.LNX.4.62.0904131938150.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 19:40:02 -0000

On Mon, 13 Apr 2009, Mark Lentczner wrote:
> 
> What I'm wondering is: How does one, given a Minimal, Event style 
> protocol, handle when there are two or more widgets on the page that 
> need updates? The home page view on Facebook, for example, has at least 
> four separate areas being live-updated.

The simplest solution is probably to start off a shared worker [1] and 
then have that worker open the WebSocket connection back to the server and 
simply have it broadcast updates to all the scripts that have connected to 
the shared worker.

[1] http://www.whatwg.org/specs/web-workers/current-work/

This would have the benefit of not increasing the number of connections 
even when the user opens more tabs, too.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From mridulm80@yahoo.com  Mon Apr 13 12:47:49 2009
Return-Path: <mridulm80@yahoo.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BEA6D3A6A06 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 12:47:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.25
X-Spam-Level: 
X-Spam-Status: No, score=-2.25 tagged_above=-999 required=5 tests=[AWL=0.349,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pOoJpA9mniz3 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 12:47:48 -0700 (PDT)
Received: from n5-vm1.bullet.mail.in.yahoo.com (n5-vm1.bullet.mail.in.yahoo.com [202.86.4.130]) by core3.amsl.com (Postfix) with SMTP id D96B73A67D8 for <hybi@ietf.org>; Mon, 13 Apr 2009 12:47:47 -0700 (PDT)
Received: from [202.86.4.171] by n5.bullet.mail.in.yahoo.com with NNFMP; 13 Apr 2009 19:48:57 -0000
Received: from [203.104.18.52] by t2.bullet.in.yahoo.com with NNFMP; 13 Apr 2009 19:48:57 -0000
Received: from [127.0.0.1] by omp113.mail.in2.yahoo.com with NNFMP; 13 Apr 2009 19:48:57 -0000
X-Yahoo-Newman-Property: ymail-3
X-Yahoo-Newman-Id: 352872.8940.bm@omp113.mail.in2.yahoo.com
Received: (qmail 1444 invoked by uid 60001); 13 Apr 2009 19:48:57 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1239652137; bh=FA7VWcVOSOAelK152brB35sOuFzEzRxA48D0C9zdDQQ=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=ryyFDMql3u/TqNcn0lthJOrhqkhgH4Kp1FpPUvXhd9c634HWClMHdeSAHsKPFTwRuCpCmZa8W54zpT/2P5hMhm4juaX8tEKEvwKK7w1pliIaSxBMpw2Q/fXvqlamZyF5GRPyQrammlp4LS3EIowU63ZZHWT1LdRudjndbjoxF1I=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=OYLGlrXqruKwifsmw8dXrdhTVb20972een7tqlb6JC1PW5UnVQLHcI9pNem2o54XmqAb3Xj6N5ZLldpD6wEPrdOWxBPx1L0w5sAc/S5mTvEBsOh1mIs8igfX9yHXj5n5NEFm942BqVo2Cwaaz4qRAGp6KTuRAaC0+UVFAN2vPPA=;
Message-ID: <251941.1265.qm@web95411.mail.in2.yahoo.com>
X-YMail-OSG: v3SZB60VM1n1.ZEt8vqon0hztVEqRAfDZtTvzdnM69LmwHjunWzLL0V4BCzXYd1KJKaepQSkT2r0Nz4mNcFwTdKT9sRv051vB6cqNHSz_YV6ReUGSGs00MyDBZJCBqIHm.eD8SNMmrOg7ouvYEMfYPo.d0edtcrLwz14DgCixrdDxsPrRorBub0r51Gh4bkB07UJK0v7lTpnUA--
Received: from [122.167.176.126] by web95411.mail.in2.yahoo.com via HTTP; Tue, 14 Apr 2009 01:18:57 IST
X-Mailer: YahooMailClassic/5.2.15 YahooMailWebService/0.7.289.1
Date: Tue, 14 Apr 2009 01:18:57 +0530 (IST)
From: Mridul Muralidharan <mridulm80@yahoo.com>
To: Greg Wilkins <gregw@webtide.com>, Jamie Lokier <jamie@shareable.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Random Responses
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 19:47:49 -0000

=0A=0A--- On Mon, 13/4/09, Jamie Lokier <jamie@shareable.org> wrote:=0A=0A>=
 From: Jamie Lokier <jamie@shareable.org>=0A> Subject: Re: [hybi] Random Re=
sponses=0A> To: "Greg Wilkins" <gregw@webtide.com>=0A> Cc: hybi@ietf.org=0A=
> Date: Monday, 13 April, 2009, 3:02 AM=0A> Greg Wilkins wrote:=0A> > Gener=
ally agreeing with all your points.=A0 Yet the=0A> attributes you=0A> > des=
cribe do sound pretty much like BEEP protocol - or=0A> at least=0A> > beep =
framing.=0A> > =0A> > BEEP is very much a HTTP style protocol with addition=
=0A> of=0A> > bidirectionality, multiple response requests and=0A> multiple=
=0A> > channels over a single connection.=0A> =0A> Yes, although BEEPs chan=
nels are unnecessarily expensive=0A> (in latency)=0A> to create on demand.=
=0A=0APlease see below.=0A=0A=0A> =0A> > If BEEP was extended with a few de=
fault assumptions=0A> (eg=0A> > UTF-8 text body if no content type specifie=
d), then it=0A> would=0A> > approach the byte efficiency of websocket, whil=
e=0A> preserving=0A> > the extensibility and flexibility of HTTP.=0A> > =0A=
> > If the security and authentication aspects of BEEP=0A> were=0A> > strip=
ped, then a pretty simply protocol would remain.=0A> > =0A> > IF I was forc=
ed to pick a winner.... then I'd=0A> certainly=0A> > look very closely at B=
EEP or a BEEP derivative.=0A> =0A> I've looked closely at BEEP for this, an=
d I agree, BEEP has=0A> some very=0A> good things which should be copied.=
=0A> =0A> Unfortunately its channel setup is expensive (latency, not=0A> ba=
ndwidth),=0A> and often you'd want a channel per message in flight.=A0=0A> =
That forces you=0A> to either accept the channel setup latency, or queue=0A=
> messages behind=0A> each other in the same way that HTTP pipelining force=
s=0A> messages to=0A> block behind each other.=A0 Neither is ideal.=0A> =0A=
> There is a BEEP extension to create sub-channels on demand=0A> without=0A=
> having to request one, which is closer to ideal behaviour=0A> when you=0A=
> have a new message to send.=A0 But it unnecessarily=0A> constrains the=0A=
> ordering of messages.=A0 (I asked the author to relax=0A> this, but he=0A=
> didn't see any point for his applications and suggested=0A> there are=0A>=
 other protocols to choose from if I need that.=A0 I=0A> don't see the poin=
t=0A> in the constraint.)=0A> =0A> What I think would be ideal, and it's be=
en mentioned=0A> already as=0A> "asynchronous HTTP", is:=0A> =0A> =A0 =A0 H=
TTP or simplified-HTTP message headers and=0A> body.=0A> =A0 =A0 Bidirectio=
nal - either side can send requests=0A> and responses.=0A> =A0 =A0 Channel =
id attached to each message.=0A> =A0 =A0 Chunked encoding, with channel id =
attached to=0A> each chunk.=0A> =A0 =A0 Per channel flow control exactly li=
ke BEEP=0A> over TCP profile.=0A> =A0 =A0 Request/response id attached to e=
ach=0A> message.=0A> =A0 =A0 Request/response id matches responses out of=
=0A> order to requests.=0A> =A0 =A0 Pure messages with no response allowed,=
 using=0A> "no-response" r/r id.=0A> =A0 =A0 All the above features are opt=
ional.=0A> =0A> When you make those features optional, it degrades=0A> grac=
efully to=0A> HTTP pipelining, then to HTTP persistent connections.=0A> =0A=
> Also by making features optional, it is much easier to=0A> implement just=
=0A> the subset of features you need, or plug in a=0A> feature-complete=0A>=
 implementation when you need all of them.=0A> =0A> That provides the messa=
ge/event protocol some people are=0A> asking for,=0A> and also provides the=
 request/response protocol some people=0A> are asking=0A> for, does both bi=
directionally, and over a single socket,=0A> and with=0A> multiple applicat=
ions sharing the same socket if they=0A> want.=0A> =0A> It's efficient too,=
 if you regard HTTP as efficient.=A0=0A> They're about=0A> the same overhea=
d.=A0 Latency is sometimes better than=0A> HTTP, because it=0A> doesn't hav=
e pipeline blocking like HTTP.=A0 Proxy=0A> buffering=0A> requirements are =
about the same as HTTP, in particular they=0A> are=0A> bounded similarly.=
=0A=0A=0A=0AWhat I seem to be missing here is the need to create a BEEP cha=
nnel for each request/response. That would definitely be expensive (probabl=
y unusably so).=0A=0AWhat I am thinking of is, if the channel creation beco=
mes more 'application' oriented - as in, one channel per widget/app/other g=
ranularity on top of the beep channel - then wont we not have a better usag=
e ?=0AKind of like, various widgets on a page sharing the same underlying B=
EEP session each using its own channel to communicate to server (I think DO=
JO's comet protocol did have some notion of channels when I had glanced at =
it a couple of years back).=0A=0A=0ANon-async request/response within a bee=
p channel is similar to pipelining in http - unless previous request is pro=
cessed and response sent, the next is not handled - or did I miss something=
 here ?=0AThe sync behavior is per channel - it is async across channels.=
=0A=0A=0A=0ASo I guess it depends on how the channel usage is modeled - as =
you and others mentioned, if each request/response is going to be a new cha=
nnel, the costs of setup and the latency could be very high - but if we 'pr=
escribe' some rules or suggestions of channel usage - then the cost becomes=
 pretty low for long running channels.=0A=0A=0ANote that I am not actively =
advocating BEEP (though it might look that way :-) ), it is just that BEEP =
seems to me to be designed for and already handles a bunch of the usecases =
mentioned here.=0AA few things we might drop could be - tls, if already ove=
r http, SASL - if already authenticated http session, etc - the resulting p=
rofile could be pretty lightweight (and might address the latency issues).=
=0A=0A=0AIf the client requirement issue is a problem (need for xml, etc), =
it might be worthwhile to think about alternative framing rules.=0A=0A=0ARe=
gards,=0AMridul=0A=0A=0A=0A=0A=0A=0A> =0A> -- Jamie=0A> ___________________=
____________________________=0A> hybi mailing list=0A> hybi@ietf.org=0A> ht=
tps://www.ietf.org/mailman/listinfo/hybi=0A> =0A=0A=0A      Cricket on your=
 mind? Visit the ultimate cricket website. Enter http://beta.cricket.yahoo.=
com


From gregw@webtide.com  Mon Apr 13 16:02:20 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4E6653A6EDB for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:02:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.523
X-Spam-Level: 
X-Spam-Status: No, score=-2.523 tagged_above=-999 required=5 tests=[AWL=0.076,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5A+W6rQrbakd for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:02:19 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.187]) by core3.amsl.com (Postfix) with ESMTP id A087528C18E for <hybi@ietf.org>; Mon, 13 Apr 2009 16:01:14 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so15829tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 16:02:25 -0700 (PDT)
Received: by 10.110.46.3 with SMTP id t3mr9477136tit.45.1239663745067; Mon, 13 Apr 2009 16:02:25 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id j5sm2962101tid.21.2009.04.13.16.02.22 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 16:02:23 -0700 (PDT)
Message-ID: <49E3C476.7090005@webtide.com>
Date: Tue, 14 Apr 2009 09:02:14 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090319)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49DEF171.4080506@mozilla.com>	<c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>	<op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
In-Reply-To: <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 23:02:20 -0000

Mark Lentczner wrote:

> What I'm wondering is: How does one, given a Minimal, Event style 
> protocol, handle when there are two or more widgets on the page that 
> need updates? The home page view on Facebook, for example, has at least 
> four separate areas being live-updated.

I see the "many widgets on one page" as just the simplest case of
a need for multiplexing events over a single transport.

Very few sites like facebook actually put multiple widgets on
the same page, since those widgets potentially come from different
untrusted sources (and in the igoogle example, you don't want your
hot-babe-of-the-day widget sneaking a peek at your stock-ticker
widget).

So normal practise is to put widgets into iframes that are
used as sandboxes to isolate widgets.

So the first challenge is to get events in and out of these
sandboxes before they can be multiplexed onto a common channel.

This has to be coordinated by the parent page, because by
definition the sandboxed widgets can't know if there are
other widgets around with whom they should be multiplexing.

View like this, multi-widgets on a page really becomes
multiple frames with communication needs to the same
domain.     Which in turn is just a specific case of
the problem of multiple frames, tabs and/or windows
on the same domain.

It may be that as well as multiplexing events for
different widgets on the same page, we need to multiplex
and/or distribute events for different instances of the
same widget on different pages.

These are some of the issues the open ajax alliance
is trying to deal with - but I'm out of touch with their
latest thinking.

Anyway, my point is, that when considering multiplexing
events onto the same transport, we could consider the
multi frame/tab/window issues as well.


cheers


























From Martin.Thomson@andrew.com  Mon Apr 13 16:47:43 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0FAD83A685D for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:47:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.531
X-Spam-Level: 
X-Spam-Status: No, score=-2.531 tagged_above=-999 required=5 tests=[AWL=0.068,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fHt75CWr0CIZ for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:47:42 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 0D51E3A67B1 for <hybi@ietf.org>; Mon, 13 Apr 2009 16:47:41 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_13_19_09_22
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from aopexbh2.andrew.com [10.86.20.25] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Mon, 13 Apr 2009 19:09:21 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by aopexbh2.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Mon, 13 Apr 2009 18:48:52 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Mon, 13 Apr 2009 18:48:50 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>
In-Reply-To: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] Is there a traffic jam?
Thread-Index: Acm8Y4trGdLIpwlQSCiXXO2P4f6S3wAK1dbA
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Mark Lentczner" <markl@lindenlab.com>, <hybi@ietf.org>
X-OriginalArrivalTime: 13 Apr 2009 23:48:52.0729 (UTC) FILETIME=[667A0690:01C9BC92]
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 23:47:43 -0000
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=


From ian@hixie.ch  Mon Apr 13 16:52:21 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BAF573A63EB for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:52:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.424
X-Spam-Level: 
X-Spam-Status: No, score=-3.424 tagged_above=-999 required=5 tests=[AWL=-0.825, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id SZyduLQ9iBT0 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:52:20 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id DAD5C3A67F7 for <hybi@ietf.org>; Mon, 13 Apr 2009 16:52:20 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 0A94A121FB0; Mon, 13 Apr 2009 16:53:32 -0700 (PDT)
Date: Mon, 13 Apr 2009 23:53:31 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>
Message-ID: <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 23:52:21 -0000

On Mon, 13 Apr 2009, Thomson, Martin wrote:
>
> The "XML" parsers I've seen in implementations of BEEP are rarely as 
> complex as you make out.

Does this mean BEEP implementations don't reuse actual XML parsers?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From Martin.Thomson@andrew.com  Mon Apr 13 16:59:51 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id F2A2A3A6D0F for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:59:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.532
X-Spam-Level: 
X-Spam-Status: No, score=-2.532 tagged_above=-999 required=5 tests=[AWL=0.067,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gV4d+S50geJr for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 16:59:51 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 152F43A6A68 for <hybi@ietf.org>; Mon, 13 Apr 2009 16:59:51 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_13_19_21_30
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from acdcexbh1.andrew.com [10.86.20.91] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Mon, 13 Apr 2009 19:21:30 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by acdcexbh1.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Mon, 13 Apr 2009 19:01:01 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Mon, 13 Apr 2009 19:01:00 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>
In-Reply-To: <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] Is there a traffic jam?
Thread-Index: Acm8kw5RphP5Eus6QJe0rKBle2JRUAAAOeiA
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Ian Hickson" <ian@hixie.ch>
X-OriginalArrivalTime: 14 Apr 2009 00:01:01.0505 (UTC) FILETIME=[18DC6B10:01C9BC94]
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2009 23:59:52 -0000
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From ian@hixie.ch  Mon Apr 13 17:03:08 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2A5B528C11C for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:03:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.406
X-Spam-Level: 
X-Spam-Status: No, score=-3.406 tagged_above=-999 required=5 tests=[AWL=-0.807, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id aHXrJYzC6IdO for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:03:07 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 7304528C0FA for <hybi@ietf.org>; Mon, 13 Apr 2009 17:03:07 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id CD273121FB0; Mon, 13 Apr 2009 17:04:18 -0700 (PDT)
Date: Tue, 14 Apr 2009 00:04:18 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>
Message-ID: <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:03:08 -0000

On Mon, 13 Apr 2009, Thomson, Martin wrote:
>
> Some do the right thing.  Others do what you might fear.

This is roughly why I believe one of our requirements should be that 
implementations should be possible in few lines of code without the use of 
any libraries. Sure, HTTP libraries are a dime a dozen, and we can rely on 
XML parsers and so on, but in practice, some will do the right thing, and 
others won't.

I'd rather have a protocol that's so simple that nobody will feel the need 
to write their own HTTP parser, say.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From rsayre@mozilla.com  Mon Apr 13 17:04:48 2009
Return-Path: <rsayre@mozilla.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2F6F728C0F8 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:04:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.001
X-Spam-Level: 
X-Spam-Status: No, score=0.001 tagged_above=-999 required=5 tests=[BAYES_50=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BIYD0zskNbts for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:04:47 -0700 (PDT)
Received: from dm-mail02.mozilla.org (dm-mail02.mozilla.org [63.245.208.176]) by core3.amsl.com (Postfix) with ESMTP id 8C1AF28C0FA for <hybi@ietf.org>; Mon, 13 Apr 2009 17:04:47 -0700 (PDT)
Received: from robert-sayres-macbook-pro.local (user-160vobr.cable.mindspring.com [76.15.225.123]) (Authenticated sender: rsayre@mozilla.com) by dm-mail02.mozilla.org (Postfix) with ESMTP id BB44A8240EA for <hybi@ietf.org>; Mon, 13 Apr 2009 17:05:56 -0700 (PDT)
Message-ID: <49E3D363.8030803@mozilla.com>
Date: Mon, 13 Apr 2009 20:05:55 -0400
From: Rob Sayre <rsayre@mozilla.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b3pre) Gecko/20090223 Thunderbird/3.0b2
MIME-Version: 1.0
To: hybi@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:04:48 -0000

Most of the BEEP RFCs are 8 years old now, and there hasn't been a lot 
of adoption. There might be good ideas there, but something is clearly 
amiss. It might be the XML dependency, it might be something else.

- Rob

From prescod@gmail.com  Mon Apr 13 17:11:05 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CB1C13A6918 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:11:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.655
X-Spam-Level: 
X-Spam-Status: No, score=-1.655 tagged_above=-999 required=5 tests=[AWL=0.322,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eUXVY8ge64pE for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:11:05 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.229]) by core3.amsl.com (Postfix) with ESMTP id 5D4F03A67F7 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:10:52 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id b25so1881690rvf.5 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:12:03 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=Q6swZCZMw5C8BKTe/B8N9UO2DrtuxMk29/DtyrLVmoQ=; b=P3cz2Zh1kO3O+m3STcQqTHNx6BMZNpJUl5TCFbygrjjYhn6CPOTIXkxGCZh8LVmZzh DXRhm15Sw04DCobhECySbsflhA1jervLdbXab3WzV084h1wR+TyqFLJfM9wkqCcu0LWP dKsFyiIyWeM/tEIikHfRru1adydiusNOCxFtI=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=DISmJwHRgo60H8rbQOo6/E2ld1aJ+fB4cypGEYzR0GSD0Ps3zcK4X0MFmHpvtqaYQ7 kgSjkVsRBgQcNN5JSAF1PH9LgoAvVMckq12WAJTwdiE2n5wD60NuJjU8wHV81Bdd1BNp 72fdxRpoAjOeTLb5IH6tt+G1v76FZzOjWboCw=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.140.166.16 with SMTP id o16mr2959421rve.25.1239667923425; Mon,  13 Apr 2009 17:12:03 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>
Date: Mon, 13 Apr 2009 17:12:03 -0700
X-Google-Sender-Auth: e92505eff3a04a16
Message-ID: <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org, "Thomson, Martin" <Martin.Thomson@andrew.com>
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:11:05 -0000

On Mon, Apr 13, 2009 at 5:04 PM, Ian Hickson <ian@hixie.ch> wrote:
> ...
>
> I'd rather have a protocol that's so simple that nobody will feel the need
> to write their own HTTP parser, say.

Do you have any evidence that people write their own HTTP parsers when
they use other protocols that run on top of or extend HTTP? Atompub,
WebDAV, SOAP, XML-RPC?

Why would web sockets programmers be more likely to write their own
HTTP parsers than other programmers?

 Paul Prescod

From gregw@webtide.com  Mon Apr 13 17:17:48 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 58DC83A67A1 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:17:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.527
X-Spam-Level: 
X-Spam-Status: No, score=-2.527 tagged_above=-999 required=5 tests=[AWL=0.072,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BUd7nNwk-T1g for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:17:47 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id 5462A3A67F3 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:17:47 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so27399tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:18:57 -0700 (PDT)
Received: by 10.110.8.5 with SMTP id 5mr9595400tih.10.1239668337744; Mon, 13 Apr 2009 17:18:57 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id u8sm2455652tia.10.2009.04.13.17.18.55 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 17:18:56 -0700 (PDT)
Message-ID: <49E3D66C.5060002@webtide.com>
Date: Tue, 14 Apr 2009 10:18:52 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>	<Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:17:48 -0000

Ian Hickson wrote:
> On Mon, 13 Apr 2009, Thomson, Martin wrote:
>> Some do the right thing.  Others do what you might fear.
> 
> This is roughly why I believe one of our requirements should be that 
> implementations should be possible in few lines of code without the use of 
> any libraries. Sure, HTTP libraries are a dime a dozen, and we can rely on 
> XML parsers and so on, but in practice, some will do the right thing, and 
> others won't.
> 
> I'd rather have a protocol that's so simple that nobody will feel the need 
> to write their own HTTP parser, say.

Ian,

the problem with a really simple protocol is that encourages wheels to
be reinvented further up the stack.

Not supporting features in the base protocol does not
make these things not a problem - it just makes them somebody
else's problem.

eg just because the base transport does not support multi-channel
interleaved messages with flow control, does not mean that
such requirements are not needed.

If for example we picked Websocket as *THE* protocol and
all the firewalls and gateways were changed to support upgrade
to websocket - then I would still expect to see many different
protocols being tunnelled over websocket datagrams.

These may be application specific protocols based on UTF-8
datagrams or they may even be existing protocols such as BEEP
tunnelled over websocket binary datagrams.

There would still need to be conventions or standards so
that we could work out that the websocket datagram really
does contain the protocol we are expecting, and then that
protocol will need to be implemented with all it's parsers
and libraries.


The question is - is there a reasonable subset of features
that could be supported at the base protocol level that
would reduce the number of reinvented wheels.

I really liked Jamie Lokiers suggestion of a set of optional
extension to HTTP, so that minimal implementations need only
implement the features that they need.   Thus support of
multiplexed interleaved messages with flow control would
not be manditory, but also would not have to be reimplemented
by every stack that needed it.


cheers




























From rsayre@mozilla.com  Mon Apr 13 17:21:05 2009
Return-Path: <rsayre@mozilla.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C16B63A6AD1 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:21:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.299
X-Spam-Level: 
X-Spam-Status: No, score=-1.299 tagged_above=-999 required=5 tests=[AWL=1.300,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7kEyEcQAMHul for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:21:05 -0700 (PDT)
Received: from dm-mail02.mozilla.org (dm-mail02.mozilla.org [63.245.208.176]) by core3.amsl.com (Postfix) with ESMTP id 30DF03A6EA1 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:20:59 -0700 (PDT)
Received: from robert-sayres-macbook-pro.local (user-160vobr.cable.mindspring.com [76.15.225.123]) (Authenticated sender: rsayre@mozilla.com) by dm-mail02.mozilla.org (Postfix) with ESMTP id 5576E8240E0; Mon, 13 Apr 2009 17:22:10 -0700 (PDT)
Message-ID: <49E3D731.30305@mozilla.com>
Date: Mon, 13 Apr 2009 20:22:09 -0400
From: Rob Sayre <rsayre@mozilla.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b3pre) Gecko/20090223 Thunderbird/3.0b2
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>	<Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>	<Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com>
In-Reply-To: <49E3D66C.5060002@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:21:05 -0000

On 4/13/09 8:18 PM, Greg Wilkins wrote:
> the problem with a really simple protocol is that encourages wheels to
> be reinvented further up the stack.

"The good thing about reinventing the wheel is that you can get a round 
one."

  - Rob

From mario.balibrera@gmail.com  Mon Apr 13 17:26:05 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 950443A6800 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:26:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ragpJeyINGwf for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:26:04 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.179]) by core3.amsl.com (Postfix) with ESMTP id 68B183A67F4 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:26:04 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so1160052waf.5 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:27:15 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=Mz7gntTfbuf4n6cHIem5qS+JpJ93LEm5Jr5vQ7260gk=; b=KqOPK+SfcDf7ogArH9olwK2FQLsGTQtCQke9Fild+SLEiaNcQAbcT0v0sN/5wRd1GT M7y1NB+I6OBPum3uldmauwJI4hL1qqKveApJ+2ejK7PsyBsnJXfljy9mgzDDCCnjMORx 0s/uByts4kocMtzCZdTFK8MXqzG7clhEiZf+M=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=C4uLRbzQuM0FkW1qHeyk8flMA8nze2E9nIekp1iToWLxvVqwxgQVVnjtoHP4xGWDAG Zyt8QdoCy0nh0J55kDY0JWgMaIhiEJ2h2kG/oSt1I7LvGSGWIbDTyofwMY9GnCporxfK fNQN5xDv+cLN2dRe5jjLwFhxbqIJLD5xjdB7g=
MIME-Version: 1.0
Received: by 10.114.95.1 with SMTP id s1mr3418565wab.50.1239668835609; Mon, 13  Apr 2009 17:27:15 -0700 (PDT)
In-Reply-To: <49E3D731.30305@mozilla.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com>
Date: Mon, 13 Apr 2009 17:27:15 -0700
Message-ID: <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Rob Sayre <rsayre@mozilla.com>
Content-Type: multipart/alternative; boundary=00163645924695ef38046778e4e3
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:26:05 -0000

--00163645924695ef38046778e4e3
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

If for example we picked Websocket as *THE* protocol and
all the firewalls and gateways were changed to support upgrade
to websocket - then I would still expect to see many different
protocols being tunnelled over websocket datagrams.

Doesn't that sound like exactly what everyone wants? Then we could all
implement our protocols, and expect them to work properly through the
dreaded proxies, all with a total overhead of one 0 byte and one 255 byte
per frame.

Mario Balibrera

On Mon, Apr 13, 2009 at 5:22 PM, Rob Sayre <rsayre@mozilla.com> wrote:

> On 4/13/09 8:18 PM, Greg Wilkins wrote:
>
>> the problem with a really simple protocol is that encourages wheels to
>> be reinvented further up the stack.
>>
>
> "The good thing about reinventing the wheel is that you can get a round
> one."
>
>  - Rob
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--00163645924695ef38046778e4e3
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

If for example we picked Websocket as *THE* protocol and<br>
all the firewalls and gateways were changed to support upgrade<br>
to websocket - then I would still expect to see many different<br>
protocols being tunnelled over websocket datagrams.<br><br>Doesn&#39;t that=
 sound like exactly what everyone wants? Then we could all implement our pr=
otocols, and expect them to work properly through the dreaded proxies, all =
with a total overhead of one 0 byte and one 255 byte per frame.<br>
<br>Mario Balibrera<br><br><div class=3D"gmail_quote">On Mon, Apr 13, 2009 =
at 5:22 PM, Rob Sayre <span dir=3D"ltr">&lt;<a href=3D"mailto:rsayre@mozill=
a.com">rsayre@mozilla.com</a>&gt;</span> wrote:<br><blockquote class=3D"gma=
il_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0=
pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">On 4/13/09 8:18 PM, Greg Wilkins wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
the problem with a really simple protocol is that encourages wheels to<br>
be reinvented further up the stack.<br>
</blockquote>
<br></div>
&quot;The good thing about reinventing the wheel is that you can get a roun=
d one.&quot;<br><font color=3D"#888888">
<br>
=A0- Rob</font><div><div></div><div class=3D"h5"><br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org" target=3D"_blank">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--00163645924695ef38046778e4e3--

From gregw@webtide.com  Mon Apr 13 17:29:43 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B706A3A6AF7 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:29:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.53
X-Spam-Level: 
X-Spam-Status: No, score=-2.53 tagged_above=-999 required=5 tests=[AWL=0.069,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Y41dmQF1kz3l for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:29:43 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.185]) by core3.amsl.com (Postfix) with ESMTP id C14E13A6A86 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:29:42 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so29433tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:30:53 -0700 (PDT)
Received: by 10.110.50.19 with SMTP id x19mr9583560tix.31.1239669053073; Mon, 13 Apr 2009 17:30:53 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id u12sm2494894tia.26.2009.04.13.17.30.50 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 17:30:52 -0700 (PDT)
Message-ID: <49E3D937.5020009@webtide.com>
Date: Tue, 14 Apr 2009 10:30:47 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Rob Sayre <rsayre@mozilla.com>, hybi@ietf.org
References: <49E3D363.8030803@mozilla.com>
In-Reply-To: <49E3D363.8030803@mozilla.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:29:43 -0000

Rob Sayre wrote:
> Most of the BEEP RFCs are 8 years old now, and there hasn't been a lot 
> of adoption. There might be good ideas there, but something is clearly 
> amiss. It might be the XML dependency, it might be something else.

I think BEEP suffers from the same problem that all alternative protocols
suffer from: firewalls, gateways and browsers don't let them through.

For example, XMPP has great market share - except when it comes to
webchat - where there are now a plethora of alternative protocols
for chat - all trying to work out there own way to tunnel through
firewalls/gateways and get into the browser DOM.


Thus I think these detail protocol discussions are very interesting
and a good exercise to share requirements, view points and solutions.

But fundamentally the issue we face is not *what* protocol to use,
but how to get it through the internet and into the browser DOM.

Websocket has done a great job of considering this full journey
solution. It realizes that both a wire protocol supported by gateways
and an API supported by browsers is needed.

But I think there will only be 1 opportunity this decade to make
significant changes to the firewalls/gateways of the internet.
If that is true, then is websocket the best change to make?

We could do a lot worse and the binary websocket datagrams could
at least serve as the basis for building other protocols that
can transit firewalls/gateways.   But I think we could also do
somewhat better.


cheers

PS. I love your round wheel quote!









From rsayre@mozilla.com  Mon Apr 13 17:36:02 2009
Return-Path: <rsayre@mozilla.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5B7223A63EB for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:36:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.949
X-Spam-Level: 
X-Spam-Status: No, score=-1.949 tagged_above=-999 required=5 tests=[AWL=0.650,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5P1DCidkkQNG for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:36:01 -0700 (PDT)
Received: from dm-mail01.mozilla.org (dm-mail01.mozilla.org [63.245.208.150]) by core3.amsl.com (Postfix) with ESMTP id 1293F28C0FA for <hybi@ietf.org>; Mon, 13 Apr 2009 17:35:56 -0700 (PDT)
Received: from robert-sayres-macbook-pro.local (user-160vobr.cable.mindspring.com [76.15.225.123]) (Authenticated sender: rsayre@mozilla.com) by dm-mail01.mozilla.org (Postfix) with ESMTP id 1893EB8036; Mon, 13 Apr 2009 17:37:07 -0700 (PDT)
Message-ID: <49E3DAB2.9010900@mozilla.com>
Date: Mon, 13 Apr 2009 20:37:06 -0400
From: Rob Sayre <rsayre@mozilla.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b3pre) Gecko/20090223 Thunderbird/3.0b2
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com>
In-Reply-To: <49E3D937.5020009@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:36:02 -0000

On 4/13/09 8:30 PM, Greg Wilkins wrote:
> I think BEEP suffers from the same problem that all alternative protocols
> suffer from: firewalls, gateways and browsers don't let them through.

Maybe protocols are considered "alternative" because they suck for some 
reason. Some of these alternative protocols might even be an 
improvement, but not enough of one to justify making all of the changes 
necessary to enable them.

- Rob

From ian@hixie.ch  Mon Apr 13 17:37:00 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EEACF3A6A86 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:37:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.389
X-Spam-Level: 
X-Spam-Status: No, score=-3.389 tagged_above=-999 required=5 tests=[AWL=-0.790, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jBzwIFJOkYYM for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:37:00 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 2534D3A69FB for <hybi@ietf.org>; Mon, 13 Apr 2009 17:37:00 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 501FA27B5A; Mon, 13 Apr 2009 17:38:11 -0700 (PDT)
Date: Tue, 14 Apr 2009 00:38:11 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:37:01 -0000

On Mon, 13 Apr 2009, Paul Prescod wrote:
> On Mon, Apr 13, 2009 at 5:04 PM, Ian Hickson <ian@hixie.ch> wrote:
> >
> > I'd rather have a protocol that's so simple that nobody will feel the 
> > need to write their own HTTP parser, say.
> 
> Do you have any evidence that people write their own HTTP parsers when 
> they use other protocols that run on top of or extend HTTP? Atompub, 
> WebDAV, SOAP, XML-RPC?

These are all either highly complex protocols, which generally means that 
the implementors are more likely to be experienced and competent 
programmers, or they run "on top" of HTTP in a way that makes it easier to 
just write CGI scripts rather than work at the socket layer, and thus 
there's no HTTP parsing to do.

Authors _do_ write their own handlers for the parts of HTTP that aren't 
handled by Apache, like parsing query parameters or multipart/form-data, 
despite libraries for those existing too.


> Why would web sockets programmers be more likely to write their own HTTP 
> parsers than other programmers?

Because such servers are more likely to be standalone and written by 
inexperienced programmers.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Mon Apr 13 17:38:33 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 307E13A6A13 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:38:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.533
X-Spam-Level: 
X-Spam-Status: No, score=-2.533 tagged_above=-999 required=5 tests=[AWL=0.066,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 1WCzFmiJ7wlH for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:38:32 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.188]) by core3.amsl.com (Postfix) with ESMTP id 37CF73A69FB for <hybi@ietf.org>; Mon, 13 Apr 2009 17:38:31 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so31182tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:39:41 -0700 (PDT)
Received: by 10.110.42.17 with SMTP id p17mr8427991tip.16.1239669581572; Mon, 13 Apr 2009 17:39:41 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id i9sm3162927tid.13.2009.04.13.17.39.38 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 17:39:40 -0700 (PDT)
Message-ID: <49E3DB47.5060801@webtide.com>
Date: Tue, 14 Apr 2009 10:39:35 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Mario Balibrera <mario.balibrera@gmail.com>, hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>	 <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>	 <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>	 <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>	 <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>	 <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com>
In-Reply-To: <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:38:33 -0000

Mario Balibrera wrote:
> If for example we picked Websocket as *THE* protocol and
> all the firewalls and gateways were changed to support upgrade
> to websocket - then I would still expect to see many different
> protocols being tunnelled over websocket datagrams.
> 
> Doesn't that sound like exactly what everyone wants? Then we could all 
> implement our protocols, and expect them to work properly through the 
> dreaded proxies, all with a total overhead of one 0 byte and one 255 
> byte per frame.


Mario,

it does come close (although I'd expect the binary datagram with
lengths to be used more than the 0/255 sentinels around UTF-8)

But websocket does lack support for meta data so that you can
know what protocols are being tunneled over it.

The URL of the connection can be used to indicate what protocol
is being used, but that limits us to 1 protocol per connection.

In the multi widget/frame/tab/window scenario, picking one
protocol to run over websocket could be difficult/impossible.
The result would more likely be a new connection for each protocol
and that could easily devolve into one connection per
widget/frame/tab/window.

If there was just a smidge of meta data available per
datagram - then the ability to share a single connection
would be hugely enhanced.

If upgrading the network infrastructure is a once a
decade exercise, then I really think we should consider
that a minimal solution with limited extensibility
might not be the best next step.


I'd much rather see recommendations that would allow
arbitrary protocols to be supported by upgrades to
HTTP connections.   These would support websocket, but
also allow other protocols to try out for the
one-protocol-to-carry-them-all job.

cheers














From ian@hixie.ch  Mon Apr 13 17:41:27 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B0AF83A6986 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:41:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.372
X-Spam-Level: 
X-Spam-Status: No, score=-3.372 tagged_above=-999 required=5 tests=[AWL=-0.773, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9Vs-Ktj8lwvG for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:41:26 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id DFD0F3A688F for <hybi@ietf.org>; Mon, 13 Apr 2009 17:41:26 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 4DB8727B77; Mon, 13 Apr 2009 17:42:38 -0700 (PDT)
Date: Tue, 14 Apr 2009 00:42:38 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49E3D937.5020009@webtide.com>
Message-ID: <Pine.LNX.4.62.0904140040260.10339@hixie.dreamhostps.com>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:41:27 -0000

On Tue, 14 Apr 2009, Greg Wilkins wrote:
>
> But I think there will only be 1 opportunity this decade to make 
> significant changes to the firewalls/gateways of the internet.

If significant changes can be made at all, I see no reason to believe 
there would be a ten year cool-down period between changes. I think this 
is an artificial concern.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Mon Apr 13 17:51:22 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D3C3B3A68FF for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:51:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.678
X-Spam-Level: 
X-Spam-Status: No, score=-1.678 tagged_above=-999 required=5 tests=[AWL=0.299,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FT1IfIEjrx17 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:51:22 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.227]) by core3.amsl.com (Postfix) with ESMTP id 370223A68F3 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:51:21 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1984147rvb.49 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:52:33 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=SgaMHaJFwMg8ryHFjY168VQTmzD9IaQvLsNbEcJNzLg=; b=MuSM6zUIsGvWuxch4FZPeJI6Zqh7s9AnV8BeVmlWo0Dm34tFeTlraQ0VuTHlkn6TWE f/XEp9gWMZN5qsfNg70YX8Gzid+WzktB8DGvpWjh5Aymi2WT0+7s9AUUUk4rlIhCt8hG G20lyQq4CnunXspTi9NrrPP18GzBpS5np0x6U=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=Eb5AXKUo+1Hy7qlMbnVeUUczzUMLghvqGqhaUP6OKCeMLET7tFCIj8CsGkF3x2IpJT QxUQwTWIL+hI/9pZKHyeeGPeTaEocc2SmISvjWPndi58Z9LbNBKWiVB0a+sGDZ9G0RIX iESKhGx7Us2BiNMxPgv7mJDNKlcPNVrSG0Zas=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.140.171.18 with SMTP id t18mr2977385rve.10.1239670353153; Mon,  13 Apr 2009 17:52:33 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>
Date: Mon, 13 Apr 2009 17:52:33 -0700
X-Google-Sender-Auth: 45408e9bdc5e0dda
Message-ID: <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:51:22 -0000

On Mon, Apr 13, 2009 at 5:38 PM, Ian Hickson <ian@hixie.ch> wrote:
>...
>
> Because such servers are more likely to be standalone and written by
> inexperienced programmers.

Programmers who are comfortable writing reasonably scalable TCP
servers for binary protocols are unlikely to be "inexperienced." I am
quite surprised that you think that inexperienced programmers would
want to  write a protocol handler for a binary connection-oriented
protocol. This is likely to require multi-threading in most languages.
I think it far more likely that they would expect such a thing to be
implemented by the Apache foundation or the Python core team or
Microsoft. If the HTTP protocol is hard for these implementers, wait
until they learn about multi-threading!

 Paul Prescod

From ian@hixie.ch  Mon Apr 13 17:59:23 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 89C5A3A6A13 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:59:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.357
X-Spam-Level: 
X-Spam-Status: No, score=-3.357 tagged_above=-999 required=5 tests=[AWL=-0.758, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0UGxLyXTvZeC for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 17:59:22 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id AED413A6800 for <hybi@ietf.org>; Mon, 13 Apr 2009 17:59:22 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 1EE1B121FB0; Mon, 13 Apr 2009 18:00:31 -0700 (PDT)
Date: Tue, 14 Apr 2009 01:00:31 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 00:59:23 -0000

On Mon, 13 Apr 2009, Paul Prescod wrote:
> On Mon, Apr 13, 2009 at 5:38 PM, Ian Hickson <ian@hixie.ch> wrote:
> >
> > Because such servers are more likely to be standalone and written by 
> > inexperienced programmers.
> 
> Programmers who are comfortable writing reasonably scalable TCP servers 
> for binary protocols are unlikely to be "inexperienced."

I agree. But who mentioned scalable TCP servers for binary protocols? I 
want to let people like my dad write little scripts that do insignificant 
little things, just like today people write CGI scripts for all kinds of 
stuff. I want to let 12 year olds write multiplayer games. I want to let 
someone who picked up HTML on Monday and CSS on Wednesday try their hand 
at WebSockets on Friday. I want to support the intern whose boss told him 
to put one of those interactive chat widgets on the intranet.


> This is likely to require multi-threading in most languages.

It's easy to write simple TCP servers using a select() loop without 
threads.


To reiterate: I have no objection to there being many protocols 
standardised here. But I do want to make sure that at least one of them 
enables an inexperienced programmer to bang out an interactive toy in a 
few days without having to use libraries, just like they can bang out a 
CGI script today and get interaction in forms. Currently, of the various 
proposals, only one of them comes even close to this, IMHO.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Mon Apr 13 18:00:14 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C7EF23A6800 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:00:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.536
X-Spam-Level: 
X-Spam-Status: No, score=-2.536 tagged_above=-999 required=5 tests=[AWL=0.063,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id c0sSsUhB8t8Z for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:00:14 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id D86D53A688F for <hybi@ietf.org>; Mon, 13 Apr 2009 18:00:13 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so35821tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:01:24 -0700 (PDT)
Received: by 10.110.21.17 with SMTP id 17mr9648470tiu.5.1239670884023; Mon, 13 Apr 2009 18:01:24 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id y5sm971520tia.9.2009.04.13.18.01.21 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 18:01:22 -0700 (PDT)
Message-ID: <49E3E05D.6070800@webtide.com>
Date: Tue, 14 Apr 2009 11:01:17 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>, hybi@ietf.org
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com> <Pine.LNX.4.62.0904140040260.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140040260.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:00:14 -0000

Ian Hickson wrote:
> On Tue, 14 Apr 2009, Greg Wilkins wrote:
>> But I think there will only be 1 opportunity this decade to make 
>> significant changes to the firewalls/gateways of the internet.
> 
> If significant changes can be made at all, I see no reason to believe 
> there would be a ten year cool-down period between changes. I think this 
> is an artificial concern.


Ian,

RFC2616 is dated June 1999 and governs the basic
behaviour of firewalls and proxies that are the impediment
to any of the protocols we are talking about.

Even if changes can be made faster than that, we should
not be proposing a change to the network infrastructure that is
not extensible to a reasonable set of anticipated requirements.

RFC2616 has lasted a decade because it is flexible and extensible
enough to adapt to changing demands.  It's not that other protocols
have not been better, it just that HTTP has been good enough to
prevent any other protocol establishing dominance.

This is a good thing and we should strive to emulate it!

cheers




From gregw@webtide.com  Mon Apr 13 18:07:59 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4111B3A677C for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:07:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.539
X-Spam-Level: 
X-Spam-Status: No, score=-2.539 tagged_above=-999 required=5 tests=[AWL=0.060,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id praCGFjZgrsY for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:07:58 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.187]) by core3.amsl.com (Postfix) with ESMTP id A16A43A6765 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:07:54 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so37566tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:09:05 -0700 (PDT)
Received: by 10.110.53.19 with SMTP id b19mr9620457tia.36.1239671344937; Mon, 13 Apr 2009 18:09:04 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id i9sm2563784tid.33.2009.04.13.18.09.01 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 18:09:03 -0700 (PDT)
Message-ID: <49E3E229.2060907@webtide.com>
Date: Tue, 14 Apr 2009 11:08:57 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>, hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>	<1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:07:59 -0000

Ian,

I think there is a big difference between protocol and API.

I understand that Websocket has addressed both wire protocol
and API in the same effort - and that I think is because those
involved have correctly identified that both need to be improved
to fix the general problem.

Having a simple approachable API for use by 12 year olds and
70 years old is great!     But there is not need for that
simplicity to be reflected in the protocol, or for the
protocol to not support other more complex APIs (or indeed
multiplexing of many instances of the simple API).


cheers







Ian Hickson wrote:
> On Mon, 13 Apr 2009, Paul Prescod wrote:
>> On Mon, Apr 13, 2009 at 5:38 PM, Ian Hickson <ian@hixie.ch> wrote:
>>> Because such servers are more likely to be standalone and written by 
>>> inexperienced programmers.
>> Programmers who are comfortable writing reasonably scalable TCP servers 
>> for binary protocols are unlikely to be "inexperienced."
> 
> I agree. But who mentioned scalable TCP servers for binary protocols? I 
> want to let people like my dad write little scripts that do insignificant 
> little things, just like today people write CGI scripts for all kinds of 
> stuff. I want to let 12 year olds write multiplayer games. I want to let 
> someone who picked up HTML on Monday and CSS on Wednesday try their hand 
> at WebSockets on Friday. I want to support the intern whose boss told him 
> to put one of those interactive chat widgets on the intranet.
> 
> 
>> This is likely to require multi-threading in most languages.
> 
> It's easy to write simple TCP servers using a select() loop without 
> threads.
> 
> 
> To reiterate: I have no objection to there being many protocols 
> standardised here. But I do want to make sure that at least one of them 
> enables an inexperienced programmer to bang out an interactive toy in a 
> few days without having to use libraries, just like they can bang out a 
> CGI script today and get interaction in forms. Currently, of the various 
> proposals, only one of them comes even close to this, IMHO.
> 


From ian@hixie.ch  Mon Apr 13 18:08:32 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B834E3A68E5 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:08:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.341
X-Spam-Level: 
X-Spam-Status: No, score=-3.341 tagged_above=-999 required=5 tests=[AWL=-0.742, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PAJhLgyxe2AK for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:08:31 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id C2A5A3A6765 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:08:31 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id D49D627B5A; Mon, 13 Apr 2009 18:09:42 -0700 (PDT)
Date: Tue, 14 Apr 2009 01:09:42 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49E3E05D.6070800@webtide.com>
Message-ID: <Pine.LNX.4.62.0904140104220.10339@hixie.dreamhostps.com>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com> <Pine.LNX.4.62.0904140040260.10339@hixie.dreamhostps.com> <49E3E05D.6070800@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:08:32 -0000

On Tue, 14 Apr 2009, Greg Wilkins wrote:
> Ian Hickson wrote:
> > On Tue, 14 Apr 2009, Greg Wilkins wrote:
> > > But I think there will only be 1 opportunity this decade to make
> > > significant changes to the firewalls/gateways of the internet.
> > 
> > If significant changes can be made at all, I see no reason to believe there
> > would be a ten year cool-down period between changes. I think this is an
> > artificial concern.
> 
> RFC2616 is dated June 1999 and governs the basic behaviour of firewalls 
> and proxies that are the impediment to any of the protocols we are 
> talking about.
> 
> Even if changes can be made faster than that, we should not be proposing 
> a change to the network infrastructure that is not extensible to a 
> reasonable set of anticipated requirements.
> 
> RFC2616 has lasted a decade because it is flexible and extensible enough 
> to adapt to changing demands.  It's not that other protocols have not 
> been better, it just that HTTP has been good enough to prevent any other 
> protocol establishing dominance.
> 
> This is a good thing and we should strive to emulate it!

There's a big difference between "I don't think we should be making 
changes more than once per decade" and "I don't think it's possible to 
make a change more than once a decade".

I think that "we should not be making changes more than once per decade" 
is a terrible reason to overengineer whatever protocol we come up with. I 
would much rather have incremental updates over a ten year period, leading 
to a massive protocol, than start with a massive protocol and hope we got 
it right without testing it as it is developed.

(I actually think the main reason that nothing has changed in the past 
decade is the same reason that nothing much has changed in HTML, CSS, the 
DOM, etc, in the past ten years. The browser vendors were out to lunch.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Mon Apr 13 18:13:35 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A6B343A6800 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:13:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.698
X-Spam-Level: 
X-Spam-Status: No, score=-1.698 tagged_above=-999 required=5 tests=[AWL=0.279,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id o2S8hSmbLDfl for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:13:34 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.229]) by core3.amsl.com (Postfix) with ESMTP id D67443A6765 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:13:34 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so1991169rvb.49 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:14:45 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=iR1gAAcXelY4h5XyQfBR3c3DR7Sg+Wr28Y1Q4z2w26U=; b=TbgWpsp0yBGwm0FFslLZ0Bje8jXF5X/VLOcIcu2A2WQz9hvOzSJ+skI9C6Edl/Then M9jZmxNG1SZAnShftzS2zqYPney99Ger3JGl9eFg/0M7ibO1LUDdvz1LwYZFcXcJua2D eFIz3nfy6vj7NxG+dSUPNE75HHJ6Lj5dRasPA=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=D4/g0oXEQ1uh+zW7d7bWX3UnuwB1JnAOiEZwfHOD0NYpJd32kZxQK7RSAh46ysP/sB ggrtEWU5qkSGocCetIAtJt6GSXKiFe9zahnuc8eINGDNHb3ah5UcZ8leIhaKR7aVB2v3 KGB4Vqer92mortSI1a1wlKVtsigqovmPT9cl4=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.141.50.11 with SMTP id c11mr2981568rvk.45.1239671685900; Mon,  13 Apr 2009 18:14:45 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
Date: Mon, 13 Apr 2009 18:14:45 -0700
X-Google-Sender-Auth: 34cffb7d90a8d7d7
Message-ID: <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:13:35 -0000

On Mon, Apr 13, 2009 at 6:00 PM, Ian Hickson <ian@hixie.ch> wrote:
> ...
> To reiterate: I have no objection to there being many protocols
> standardised here. But I do want to make sure that at least one of them
> enables an inexperienced programmer to bang out an interactive toy in a
> few days without having to use libraries, just like they can bang out a
> CGI script today and get interaction in forms. Currently, of the various
> proposals, only one of them comes even close to this, IMHO.

You are moving the goal posts in a very unfair way.

You use the example of CGI scripts. Do you know anyone who runs CGI
scripts without Apache in front? The very definition of CGI is that it
is a tool that allows a front-end to handle the protocol and have a
back end handle the logic. If CGI is your example of a successful
technology, then HTTP is by analogy a protocol that has the right
level of complexity/power for its task.

Yes, you want your dad to write "WebSocket" scripts. I want my 12 year
old neighbour to be able to as well. Please do not assume that we have
different goals. But people like my twelve year old neighbour know CGI
and PHP. They do not know how to parse binary data coming in from a
socket. The humane thing to do is to PROVIDE a CGI or PHP or Rails
interface to the protocol. In substantially less time than it will
take to standardize a protocol, we could make Apache modules to
implement it. Surely this is better than telling them: "here's a
binary protocol. Good luck with that!"

You're constructing a completely unrealistic scenario, like when a
reality show tells a bunch of chefs that "for today's contest you may
not use knives!" Why would we ignore 30 years of computer engineering
experience that tells us that people can and should cooperate rather
than develop everything from scratch themselves? Do you expect your
Dad to write his own UTF-8 parser?

> I want to let 12 year olds write multiplayer games.

Please be very specific about HOW a 12-year old who knows PHP would
write a multiplayer game using PHP and WebSockets. Given how easy you
say it is, code samples would be appropriate. Tic Tac toe in
PHP/WebSockets would be a very helpful demonstration of the technology
and proposed programming model.

 Paul Prescod

From ian@hixie.ch  Mon Apr 13 18:16:33 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DCDAC3A689A for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:16:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.327
X-Spam-Level: 
X-Spam-Status: No, score=-3.327 tagged_above=-999 required=5 tests=[AWL=-0.728, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Qx51x+b-ttBB for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:16:33 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id F35E43A6765 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:16:32 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 7201427B77; Mon, 13 Apr 2009 18:17:44 -0700 (PDT)
Date: Tue, 14 Apr 2009 01:17:44 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49E3E229.2060907@webtide.com>
Message-ID: <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:16:34 -0000

On Tue, 14 Apr 2009, Greg Wilkins wrote:
>
> Having a simple approachable API for use by 12 year olds and 70 years 
> old is great!  But there is not need for that simplicity to be reflected 
> in the protocol, or for the protocol to not support other more complex 
> APIs (or indeed multiplexing of many instances of the simple API).

The 12 year olds and 70 year olds in the scenarios I described are the 
people who would be writing the server-side components also. So yes, there 
is a _critical_ need for the simplicitly to be reflected in the protocol.

A simple protocol doesn't preclude more complex APIs, or multiplexing. 

For example, instead of having frames like this:

 +-FRAME-------------+-DATA----------------+
 | +-----+---------+ | +-----------------+ |
 | | ... | Channel | | | Message         | |
 | +-----+---------+ | +-----------------+ |
 +-------------------+---------------------+

...where the Channel part of multiplexing is a core part of the protocol 
that _everyone_ has to deal with, we can just as easily have:

 +-FRAME---+-DATA--------------------------+
 | +-----+ | +---------+-----------------+ |
 | | ... | | | Channel | Message         | |
 | +-----+ | +---------+-----------------+ |
 +---------+-------------------------------+

...where the Channel part is part of the application-level protocol. 
Instead of putting multiplexing into the core API and wire protocol, 
people can implement it using custom APIs and custom formats over the 
basic wire protocol.

This is just like the way that TCP has no support for individual blocks 
over a connection -- it's a streaming protocol. But we can layer blocks on 
top of TCP. That's not a failing of TCP.

(In these diagrams the "..." bit represents whatever other per-frame 
metadata we might have; in WebSockets right now this is just a frame type 
byte and a length for binary frames.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From mark@coactus.com  Mon Apr 13 18:19:27 2009
Return-Path: <mark@coactus.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C54493A689A for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:19:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.228
X-Spam-Level: 
X-Spam-Status: No, score=-4.228 tagged_above=-999 required=5 tests=[AWL=-2.251, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cWer2oPbO-mZ for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:19:27 -0700 (PDT)
Received: from qw-out-2122.google.com (qw-out-2122.google.com [74.125.92.24]) by core3.amsl.com (Postfix) with ESMTP id 07AF93A6765 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:19:26 -0700 (PDT)
Received: by qw-out-2122.google.com with SMTP id 9so1997875qwb.31 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:20:37 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.229.73.194 with SMTP id r2mr1863117qcj.29.1239672037763; Mon,  13 Apr 2009 18:20:37 -0700 (PDT)
In-Reply-To: <49E3D363.8030803@mozilla.com>
References: <49E3D363.8030803@mozilla.com>
Date: Mon, 13 Apr 2009 21:20:37 -0400
Message-ID: <e9dffd640904131820jae774a1x8e7a3c71d9bf89c3@mail.gmail.com>
From: Mark Baker <mark@coactus.com>
To: Rob Sayre <rsayre@mozilla.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] BEEP
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:19:27 -0000

On Mon, Apr 13, 2009 at 8:05 PM, Rob Sayre <rsayre@mozilla.com> wrote:
> Most of the BEEP RFCs are 8 years old now, and there hasn't been a lot of
> adoption. There might be good ideas there, but something is clearly amiss.
> It might be the XML dependency, it might be something else.

The problem with any standardized protocol at that layer is that there
are essentially no network effects; multiple parties using a BEEP
based protocol cannot benefit from each other's deployed software in
the same way that Web publishers benefit from each new deployed
browser (and vice-versa).  Network effects of that kind come only from
agreement on an application protocol, not a framework for developing
application protocols.

Mark.

From jonas@sicking.cc  Mon Apr 13 18:28:25 2009
Return-Path: <jonas@sicking.cc>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 600A63A68AA for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:28:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Fif+-0uGtb7n for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 18:28:24 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.30]) by core3.amsl.com (Postfix) with ESMTP id 60CF23A677C for <hybi@ietf.org>; Mon, 13 Apr 2009 18:28:24 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 5so3240755ywh.49 for <hybi@ietf.org>; Mon, 13 Apr 2009 18:29:35 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.151.48.20 with SMTP id a20mr14069225ybk.7.1239672575271; Mon,  13 Apr 2009 18:29:35 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>  <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>  <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>  <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com> <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>
From: Jonas Sicking <jonas@sicking.cc>
Date: Mon, 13 Apr 2009 18:29:20 -0700
Message-ID: <63df84f0904131829j5bc1b85amf8d24fdd30644f1d@mail.gmail.com>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 01:28:25 -0000

On Sun, Apr 12, 2009 at 5:11 PM, Ian Hickson <ian@hixie.ch> wrote:
> Current API:
>
> =A0 socket =3D new WebSocket('ws://example.com/chat.dcp');
> =A0 socket.onmessage =3D function (e) {
> =A0 =A0 if (e.data !=3D 'DCP') {
> =A0 =A0 =A0 error('The server mysteriously stopped speaking DCP!');
> =A0 =A0 =A0 socket.disconnect();
> =A0 =A0 }
> =A0 =A0 socket.onmesage =3D chatHandler;
> =A0 }
> =A0 socket.postMessage('DCP');
>
> Explicit API:
>
> =A0 socket =3D new WebSocket('ws://example.com/chat', 'DCP');
> =A0 socket.onprotocol =3D function (e) {
> =A0 =A0 if (e.data !=3D 'DCP') {
> =A0 =A0 =A0 error('The server mysteriously stopped speaking DCP!');
> =A0 =A0 =A0 socket.disconnect();
> =A0 =A0 }
> =A0 }
> =A0 socket.onmesage =3D chatHandler;

I agree with Maciej here. I think having explicit metadata information
regarding which protocol is being transferred is of value. Allowing
people to do the "RightThing" while still letting them do the
"EasyThing that doesn't get in the way of the RightThing" is of even
higher value.

If the API could be something like

socket =3D new WebSocket('ws://example.com/chat', 'DCP');
socket.onprotocol =3D chat_handler;

for the people that want checking. And

socket =3D new WebSocket('ws://example.com/chat');
socket.onprotocol =3D iKnowWhoImTalkingToAndDontNeedNoStinkingChecking_hand=
ler;

for the people that are using their own home-brewed protocol. This
could use some magic value as the type (a'la
"application/octet-stream") or leave out the protocol indicator
entirely.

I don't have an opinion on if this type should be transferred to the
server which would be responsible for the checking, or if the type is
transferred to the client which would do the checking, or if its
transferred both ways for added goodness.

But the point is that the type is transferred at a specified location.
And IMHO that specified location should be separated from other fields
with other purposes. My understanding is that the path is intended to
allow for several websocket services from the same origin, thus I
would prefer if the protocol is separate from the path.

/ Jonas

From gregw@webtide.com  Mon Apr 13 19:16:05 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 58BEC3A698D for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 19:16:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.541
X-Spam-Level: 
X-Spam-Status: No, score=-2.541 tagged_above=-999 required=5 tests=[AWL=0.058,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XbdoNjX3RuCA for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 19:16:04 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.189]) by core3.amsl.com (Postfix) with ESMTP id 15D5D3A685D for <hybi@ietf.org>; Mon, 13 Apr 2009 19:15:50 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so56244tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 19:17:01 -0700 (PDT)
Received: by 10.110.14.12 with SMTP id 12mr9712982tin.17.1239675421618; Mon, 13 Apr 2009 19:17:01 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id y5sm1231767tia.29.2009.04.13.19.16.58 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 19:17:00 -0700 (PDT)
Message-ID: <49E3F218.4080209@webtide.com>
Date: Tue, 14 Apr 2009 12:16:56 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com> <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 02:16:05 -0000

Ian Hickson wrote:
 > A simple protocol doesn't preclude more complex APIs, or multiplexing.


Ian,

Of course arbitrarily complex protocols can be tunnelled over
websocket.

But we can already tunnel arbitrary protocols over HTTP, so if
the ability to transport arbitrary protocols is sufficient, then
we are already there and nothing new is needed.

The question is - can we do something better than tunnelling
over HTTP.

To be frank, I'd actually rather use HTTP has a basis
than websocket.    The websocket wire protocol appears to
have focused on byte squeezing at the expense of meta data
and extensibility.

Byte-squeezing is simply not a pain point that I'm
experiencing while deploying bidirection webapplications.
Even the latency benefits of websocket are more perceived
benefits than any measurable improvement in average latency.

The pain points I'm having deploying bidirectional webapplications
are mostly about sharing connections:

  1) How to stop a browser sharing a connection by queuing
     normal requests behind long polled requests.
  2) How to share limited connections between
     widgets/frames/tabs/windows.
  3) How to limit the growth of connections on the server.


Websocket potentially helps with #1, but only if upgraded connections
are excluded from connection accounting (obvious to do, but not
sure if that is specified anywhere?).

Websocket doesn't help with #2 and #3 at all.  In fact it probably
makes it worse.  While individual connections are bidirectional, each
widgets will be able to each initiate their own websocket connection to
their own backend handling.  So the end result is likely to be a lot
more unshared connection.


cheers



From ian@hixie.ch  Mon Apr 13 19:54:15 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 85B583A6A26 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 19:54:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.313
X-Spam-Level: 
X-Spam-Status: No, score=-3.313 tagged_above=-999 required=5 tests=[AWL=-0.714, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZJCFBP1Ggh7P for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 19:54:14 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 76FD53A6863 for <hybi@ietf.org>; Mon, 13 Apr 2009 19:54:14 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 66AC527B5A; Mon, 13 Apr 2009 19:55:25 -0700 (PDT)
Date: Tue, 14 Apr 2009 02:55:25 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49E3F218.4080209@webtide.com>
Message-ID: <Pine.LNX.4.62.0904140248050.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com> <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com> <49E3F218.4080209@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 02:54:15 -0000

On Tue, 14 Apr 2009, Greg Wilkins wrote:
> 
> But we can already tunnel arbitrary protocols over HTTP, so if the 
> ability to transport arbitrary protocols is sufficient, then we are 
> already there and nothing new is needed.

Personally I think that having a single underlying TCP connection and the 
ability to do this across domains are pretty important new aspects.


> To be frank, I'd actually rather use HTTP has a basis than websocket.

As I've said before, I have nothing against people using HTTP, either 
straight up as is done today, or in some other protocol like rHTTP.


> The pain points I'm having deploying bidirectional webapplications are 
> mostly about sharing connections:
> 
>  1) How to stop a browser sharing a connection by queuing
>     normal requests behind long polled requests.

WebSocket addresses this by not sharing TCP connections with regular HTTP 
traffic.


>  2) How to share limited connections between
>     widgets/frames/tabs/windows.

WebSocket and SharedWorkers can be used together to get the number of 
long-lived connections per origin (domain) down to one.


>  3) How to limit the growth of connections on the server.

I don't know how to reduce it to less than one TCP connection per origin 
per user. I'm certainly interested in proposals that achieve this without 
sacrificing the other requirements.


> Websocket potentially helps with #1, but only if upgraded connections 
> are excluded from connection accounting (obvious to do, but not sure if 
> that is specified anywhere?).

Since WebSocket isn't HTTP, the HTTP limit doesn't apply, as far as I can 
tell. However, I'm happy to edit the spec to make this clearer, or to 
introduce a limit if one should be imposed.


> Websocket doesn't help with #2 and #3 at all.  In fact it probably makes 
> it worse.  While individual connections are bidirectional, each widgets 
> will be able to each initiate their own websocket connection to their 
> own backend handling.  So the end result is likely to be a lot more 
> unshared connection.

I don't see any reason why this should be the case. Why would it be 
possible to use one pair of connections with HTTP today, but not use a 
single WebSocket connection in the future? Especially considering new 
mechanisms like SharedWorker, which allow resources from one origin to be 
shared across multiple tabs in a browser.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Mon Apr 13 21:12:44 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E12ED3A69AA for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 21:12:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.543
X-Spam-Level: 
X-Spam-Status: No, score=-2.543 tagged_above=-999 required=5 tests=[AWL=0.056,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hgxujXDqreN8 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 21:12:44 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id 023203A6829 for <hybi@ietf.org>; Mon, 13 Apr 2009 21:12:43 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so83666tim.25 for <hybi@ietf.org>; Mon, 13 Apr 2009 21:13:54 -0700 (PDT)
Received: by 10.110.28.15 with SMTP id b15mr9849743tib.4.1239682434405; Mon, 13 Apr 2009 21:13:54 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm103549tib.28.2009.04.13.21.13.51 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 21:13:53 -0700 (PDT)
Message-ID: <49E40D7B.7090409@webtide.com>
Date: Tue, 14 Apr 2009 14:13:47 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com> <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com> <49E3F218.4080209@webtide.com> <Pine.LNX.4.62.0904140248050.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140248050.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 04:12:45 -0000

Ian Hickson wrote:

> I don't see any reason why this should be the case. Why would it be 
> possible to use one pair of connections with HTTP today, but not use a 
> single WebSocket connection in the future? 

Of course it is possible to share websocket connections just as we
currently share HTTP connections.

But sharing HTTP connections is currently painful and difficult and
relies on a variety of javascript/dom hacks.   It also needs protocol
support on top of HTTP to support the multiplexing.

Websockets does not help with this.  It will still be painful to
share connections and protocols to multiplex will still need to
be written on top of it.   However, with the more approachable
API of websockets, it is likely that it will be used directly
by widget/application developers (and it is "sold" as such), so
the probability of sharing/multiplexing frameworks being
widely used is small.

 > Especially considering new
 > mechanisms like SharedWorker, which allow resources from one origin to be
 > shared across multiple tabs in a browser.

Perhaps SharedWorkers may indeed help with sharing a connection.
Indeed such new features are going to be needed in the
browser to solve this issue regardless of the protocol used and
would equally apply to HTTP.

Note also that if such features do better support sharing of connections,
then good support for multiplexing becomes even more important.

regards






From ian@hixie.ch  Mon Apr 13 22:26:35 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 16BB13A67F8 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 22:26:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.999
X-Spam-Level: 
X-Spam-Status: No, score=-2.999 tagged_above=-999 required=5 tests=[AWL=-1.000, BAYES_00=-2.599, J_CHICKENPOX_46=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id q8lOhmaVaxal for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 22:26:34 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 1FE8F3A67D1 for <hybi@ietf.org>; Mon, 13 Apr 2009 22:26:34 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 776CD27B77; Mon, 13 Apr 2009 22:27:26 -0700 (PDT)
Date: Tue, 14 Apr 2009 05:27:25 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49E40D7B.7090409@webtide.com>
Message-ID: <Pine.LNX.4.62.0904140510380.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com> <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com> <49E3F218.4080209@webtide.com> <Pine.LNX.4.62.0904140248050.10339@hixie.dreamhostps.com> <49E40D7B.7090409@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 05:26:35 -0000

On Tue, 14 Apr 2009, Greg Wilkins wrote:
> 
> But sharing HTTP connections is currently painful and difficult and 
> relies on a variety of javascript/dom hacks.  It also needs protocol 
> support on top of HTTP to support the multiplexing.
> 
> Websockets does not help with this.  It will still be painful to share 
> connections and protocols to multiplex will still need to be written on 
> top of it.  However, with the more approachable API of websockets, it is 
> likely that it will be used directly by widget/application developers 
> (and it is "sold" as such), so the probability of sharing/multiplexing 
> frameworks being widely used is small.

Come now. All you have to do to share a connection on the client is have a 
single object that does all the sending and receiving, with a 
corresponding piece of code on the server, which simply does three things:

1. To create a new "subconnection", sends a message, say "NEW" followed by 
a unique number followed by whatever information is needed to set up the 
remote end (e.g. the kind of widget), then hands back to the API caller a 
handle that is assocaited with the new unique number.

    "NEW 1 calendar-widget user=john@example.com"

2. When sending a message given one of these handles, sends the number 
followed by whatever the message is.

    "1 add-event,2009-01-01,Holiday"

3. When receiving a message, strips off the leading number, then hands the 
rest of the message to the corresponding listener.

Why would this be painful? You can do it in JavaScript in barely 20 lines 
of code:

  function MultiplexConnection(url) {
    this.socket = new WebSocket(url);
    this.connections = {};
    this.lastID = 0;
    this.newConnection = function(callback, s) {
      this.lastID = this.lastID + 1;
      this.connections[this.lastID] = callback;
      this.socket.postMessage('NEW ' + this.lastID + ' ' + s);
      return this.lastID;
    };
    this.postMessage = function (handle, s) {
      this.socket.postMessage(handle + ' ' + s);
    };
    this.socket.onmessage = function (e) {
      var handle = e.data.substr(0, e.data.indexOf(' '));
      this.connections[handle](e.data.substr(e.data.indexOf(' ')+1));
    };
  }

The server side is only midly more complicated because of having to handle 
the "NEW" connections.

If you want to use a multiplexing framework, then you can write one (and 
make it available to people, if you want others to use it too). Doing so 
on top of WebSockets is easy. Whether other people do it or not is neither 
your problem nor mine, as far as I can tell.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From sh@defuze.org  Mon Apr 13 23:09:31 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1D28D3A6886 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:09:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.165
X-Spam-Level: 
X-Spam-Status: No, score=-2.165 tagged_above=-999 required=5 tests=[AWL=0.100,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yFrBkZ3VF2Gd for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:09:30 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 44EF93A6A59 for <hybi@ietf.org>; Mon, 13 Apr 2009 23:09:30 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3E6Acdr015065; Tue, 14 Apr 2009 01:10:39 -0500
Message-ID: <49E428DD.3070803@defuze.org>
Date: Tue, 14 Apr 2009 08:10:37 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090318)
MIME-Version: 1.0
To: Greg Wilkins <gregw@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>		<49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com>	<79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com>
In-Reply-To: <49E3DB47.5060801@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 06:09:31 -0000

>
> If upgrading the network infrastructure is a once a
> decade exercise, then I really think we should consider
> that a minimal solution with limited extensibility
> might not be the best next step.
>
>
> I'd much rather see recommendations that would allow
> arbitrary protocols to be supported by upgrades to
> HTTP connections.   These would support websocket, but
> also allow other protocols to try out for the
> one-protocol-to-carry-them-all job.

Indeed. I actually wonder in fact if the problem hasn't been addressed 
the wrong way around. It seems to me that we've had to discuss so far of 
a protocol that could basically work out for everything. But could it? I 
guess most people here have at some point implemented various protocols 
and feel like WS will not be good at supporting any of them without 
major headaches and hacks.

The question I have now is, what does precisely stop browsers from 
offering a raw TCP socket API at the Javascript level?

I assume there are two main reasons:

* security risks
* problems to go through firewalls, proxies, etc.

The answer to those so far has been WebSocket but is it the right way? 
The endless nitpicking of this list should probably be warning enough in 
my opinion that it might be the wrong path.

Personally I'd rather have a direct TCP socket access from the browser 
and let organisations decide if they should allow specific ports to be 
opened. Afterall, tunelling through HTTP might end up with those 
organisation sniffing traffic and deny HTTP Upgrade anyway.

- Sylvain
http://www.defuze.org

From mike@dierken.com  Mon Apr 13 23:16:47 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A346528C113 for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:16:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.669
X-Spam-Level: 
X-Spam-Status: No, score=-1.669 tagged_above=-999 required=5 tests=[AWL=-0.929, BAYES_20=-0.74]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id o0O4Z0cqeaQN for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:16:47 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.231]) by core3.amsl.com (Postfix) with ESMTP id 062B028B56A for <hybi@ietf.org>; Mon, 13 Apr 2009 23:16:46 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id b25so1994637rvf.5 for <hybi@ietf.org>; Mon, 13 Apr 2009 23:17:58 -0700 (PDT)
Received: by 10.141.114.19 with SMTP id r19mr3083511rvm.24.1239689878328; Mon, 13 Apr 2009 23:17:58 -0700 (PDT)
Received: from mercury (c-76-121-195-84.hsd1.wa.comcast.net [76.121.195.84]) by mx.google.com with ESMTPS id k37sm15526505rvb.0.2009.04.13.23.17.56 (version=SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 23:17:57 -0700 (PDT)
From: "S. Mike Dierken" <mike@dierken.com>
To: "'Sylvain Hellegouarch'" <sh@defuze.org>, "'Greg Wilkins'" <gregw@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>		<49E3D66C.5060002@webtide.com><49E3D731.30305@mozilla.com>	<79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com><49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org>
Date: Mon, 13 Apr 2009 23:17:57 -0700
Message-ID: <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
Thread-Index: Acm8x79BWtDct7FRR+6KUicLzQkgtwAAOYTA
In-Reply-To: <49E428DD.3070803@defuze.org>
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 06:16:47 -0000

> The question I have now is, what does precisely stop browsers 
> from offering a raw TCP socket API at the Javascript level?
> 
I thought that was available, via object/applet style plug-ins. 


From mike@dierken.com  Mon Apr 13 23:16:49 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3A99C28C11A for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:16:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.36
X-Spam-Level: 
X-Spam-Status: No, score=-1.36 tagged_above=-999 required=5 tests=[AWL=-0.620,  BAYES_20=-0.74]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Hk-tBzWRmtwY for <hybi@core3.amsl.com>; Mon, 13 Apr 2009 23:16:48 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.238]) by core3.amsl.com (Postfix) with ESMTP id 78C5828C113 for <hybi@ietf.org>; Mon, 13 Apr 2009 23:16:48 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2086702rvb.49 for <hybi@ietf.org>; Mon, 13 Apr 2009 23:17:59 -0700 (PDT)
Received: by 10.140.170.12 with SMTP id s12mr3083246rve.53.1239689879661; Mon, 13 Apr 2009 23:17:59 -0700 (PDT)
Received: from mercury (c-76-121-195-84.hsd1.wa.comcast.net [76.121.195.84]) by mx.google.com with ESMTPS id k37sm15526505rvb.0.2009.04.13.23.17.58 (version=SSLv3 cipher=RC4-MD5); Mon, 13 Apr 2009 23:17:58 -0700 (PDT)
From: "S. Mike Dierken" <mike@dierken.com>
To: "'Ian Hickson'" <ian@hixie.ch>, "'Paul Prescod'" <paul@prescod.net>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com><1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
Date: Mon, 13 Apr 2009 23:17:57 -0700
Message-ID: <3BEA574FEF444A379E008FA59330CF77@mercury>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
Thread-Index: Acm8nHYnRx3wUu5lT/uStiAkQH+0hQAKEtwA
In-Reply-To: <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 06:16:49 -0000

I have the same wish - let most anybody write cool little event driven apps
in browsers and other web-enabled apps and devices.

However I do think that some of the heavy lifting and 'best practices'
should be done by the user-agents. I don't see that the raw low level byte
stream has to be presented to the casual programmer. But I also think the
low level stream can be simple enough for most slightly-more-than-casual
programmers to handle.

I don't understand what all the difficulty is around this discussion - the
rHTTP draft looks great & is pretty much was I was suggesting three years
ago. Let's put rHTTP into a browser and see what happens!




> I want to let people like my dad write little 
> scripts that do insignificant little things, just like today 
> people write CGI scripts for all kinds of stuff. I want to 
> let 12 year olds write multiplayer games. I want to let 
> someone who picked up HTML on Monday and CSS on Wednesday try 
> their hand at WebSockets on Friday. I want to support the 
> intern whose boss told him to put one of those interactive 
> chat widgets on the intranet.
 

> To reiterate: I have no objection to there being many protocols 
> standardised here. But I do want to make sure that at least 
> one of them 
> enables an inexperienced programmer to bang out an 
> interactive toy in a 
> few days without having to use libraries, just like they can 
> bang out a 
> CGI script today and get interaction in forms. Currently, of 
> the various 
> proposals, only one of them comes even close to this, IMHO.
> 


From sh@defuze.org  Tue Apr 14 00:05:54 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DAB633A69DD for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:05:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.19
X-Spam-Level: 
X-Spam-Status: No, score=-2.19 tagged_above=-999 required=5 tests=[AWL=0.075,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9cCY85xgMx6q for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:05:54 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 99B083A68D3 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:05:36 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3E76ius029384; Tue, 14 Apr 2009 02:06:44 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n3E76hfM029382; Tue, 14 Apr 2009 09:06:43 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Tue, 14 Apr 2009 09:06:43 +0200 (CEST)
Message-ID: <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com>
In-Reply-To: <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com><49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com><49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury>
Date: Tue, 14 Apr 2009 09:06:43 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: "S. Mike Dierken" <mike@dierken.com>
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 07:05:54 -0000

>
>> The question I have now is, what does precisely stop browsers
>> from offering a raw TCP socket API at the Javascript level?
>>
> I thought that was available, via object/applet style plug-ins.
>

But if that was the solution we wouldn't have this discussion. It's
because people want to/have to write their application in javascript that
there is a need for a better solution than what exists now.

Perhaps the reason BEEP has never taken off is because it's so versatile
that it ultimately creates a bigger mismatch. I'm afraid WS might follow
the same trend.

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From gregw@webtide.com  Tue Apr 14 00:09:51 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9F1F13A6A5A for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:09:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.545
X-Spam-Level: 
X-Spam-Status: No, score=-2.545 tagged_above=-999 required=5 tests=[AWL=0.054,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mPdYm4og6Jai for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:09:50 -0700 (PDT)
Received: from mail-gx0-f160.google.com (mail-gx0-f160.google.com [209.85.217.160]) by core3.amsl.com (Postfix) with ESMTP id 351AC3A6A97 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:09:50 -0700 (PDT)
Received: by gxk4 with SMTP id 4so5281185gxk.13 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:11:01 -0700 (PDT)
Received: by 10.90.89.18 with SMTP id m18mr1305403agb.0.1239693061244; Tue, 14 Apr 2009 00:11:01 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 8sm7575578agd.58.2009.04.14.00.10.59 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 14 Apr 2009 00:11:00 -0700 (PDT)
Message-ID: <49E436FE.1020503@webtide.com>
Date: Tue, 14 Apr 2009 17:10:54 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>, hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <49E3E229.2060907@webtide.com> <Pine.LNX.4.62.0904140110040.10339@hixie.dreamhostps.com> <49E3F218.4080209@webtide.com> <Pine.LNX.4.62.0904140248050.10339@hixie.dreamhostps.com> <49E40D7B.7090409@webtide.com> <Pine.LNX.4.62.0904140510380.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904140510380.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 07:09:51 -0000

Ian Hickson wrote:
> Come now. All you have to do to share a connection on the client is have a 
> single object that does all the sending and receiving, with a 
> corresponding piece of code on the server, which simply does three things:

Ian,

implementing multiplexing is not the pain point as there are a million
ways to do it, ranging from trivial to amazingly complex.   We
can multi-plex over HTTP, yet it is still a paint point.

The problem is getting everybody to agree to share and use
the same multiplexing.

Websocket presents a nice simple interface that is easy for web developers
to use directly.  So they will use it directly!    As well as that, there
will be dojo/jquery/prototype/mootools/yahoo/ext/etc wrappers.   The
situation on the server side is even more diverse with regards to servers
and frameworks.

There is ZERO chance that all these diverse communities will come together
and decide on how to share and multiplex connections.   Websockets will
simply encourage direct consumption of connections.  If you are really
lucky, some frameworks might do their own multiplexing between their
own widgets and frames.    But with websocket, it simply will not
happen that a common standard will easily/quickly emerge for sharing and
multiplexing.


regards





From gregw@webtide.com  Tue Apr 14 00:35:46 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CCABB3A6AA8 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:35:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.547
X-Spam-Level: 
X-Spam-Status: No, score=-2.547 tagged_above=-999 required=5 tests=[AWL=0.052,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id C2AmPyfeyzLo for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:35:46 -0700 (PDT)
Received: from yx-out-2324.google.com (yx-out-2324.google.com [74.125.44.30]) by core3.amsl.com (Postfix) with ESMTP id F0B283A69D3 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:35:45 -0700 (PDT)
Received: by yx-out-2324.google.com with SMTP id 8so3358324yxm.49 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:36:57 -0700 (PDT)
Received: by 10.100.248.16 with SMTP id v16mr5257786anh.60.1239694617059; Tue, 14 Apr 2009 00:36:57 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id b7sm19966538ana.12.2009.04.14.00.36.55 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 14 Apr 2009 00:36:56 -0700 (PDT)
Message-ID: <49E43D13.60308@webtide.com>
Date: Tue, 14 Apr 2009 17:36:51 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>		<E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>		<Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>		<49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com>	<79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org>
In-Reply-To: <49E428DD.3070803@defuze.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 07:35:46 -0000

Sylvain Hellegouarch wrote:
> The endless nitpicking of this list should probably be warning enough in 
> my opinion that it might be the wrong path.

Well as one of the prime nitpickers.... I don't actually think it is
endless.

So sorry if it seams that I'm endlessly picking on websocket.
Not my intention - but it has been a convenient straw man to
illustrate that a simple protocol is not going to resolve some
outstanding issues, as BEEP has been a straw man to show that
a more complex protocol is going to have overheads that some
find unacceptable.

The conversation has been worthwhile (at least for me).  I'm
moving away from my "let's not pick winners" advocacy towards
a base protocol that does support sharing and multiplexing.
The HTTP extensions that Jamie described sound very interesting!

cheers


From sh@defuze.org  Tue Apr 14 00:41:27 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7733A3A68AC for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:41:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.205
X-Spam-Level: 
X-Spam-Status: No, score=-2.205 tagged_above=-999 required=5 tests=[AWL=0.060,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AW5W+n3YnVMT for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 00:41:26 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 8609C3A6A03 for <hybi@ietf.org>; Tue, 14 Apr 2009 00:41:26 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3E7gb4a006147; Tue, 14 Apr 2009 02:42:37 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n3E7gbT5006145; Tue, 14 Apr 2009 09:42:37 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Tue, 14 Apr 2009 09:42:37 +0200 (CEST)
Message-ID: <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com>
In-Reply-To: <49E43D13.60308@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <49E43D13.60308@webtide.com>
Date: Tue, 14 Apr 2009 09:42:37 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: "Greg Wilkins" <gregw@webtide.com>
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 07:41:27 -0000

> Sylvain Hellegouarch wrote:
>> The endless nitpicking of this list should probably be warning enough in
>> my opinion that it might be the wrong path.
>
> Well as one of the prime nitpickers.... I don't actually think it is
> endless.
>

I apologise if you felt I was targetting you. It wasn't my intention at all.

> So sorry if it seams that I'm endlessly picking on websocket.
> Not my intention - but it has been a convenient straw man to
> illustrate that a simple protocol is not going to resolve some
> outstanding issues, as BEEP has been a straw man to show that
> a more complex protocol is going to have overheads that some
> find unacceptable.

Much agreed.

>
> The conversation has been worthwhile (at least for me).  I'm
> moving away from my "let's not pick winners" advocacy towards
> a base protocol that does support sharing and multiplexing.

I'm still not convinced about the idea of a base protocol myself.

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From annevk@opera.com  Tue Apr 14 01:02:45 2009
Return-Path: <annevk@opera.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6B7A63A6841 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 01:02:45 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Nccy4FL9oucu for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 01:02:44 -0700 (PDT)
Received: from smtp.opera.com (sam.opera.com [213.236.208.81]) by core3.amsl.com (Postfix) with ESMTP id 16F713A6873 for <hybi@ietf.org>; Tue, 14 Apr 2009 01:02:43 -0700 (PDT)
Received: from anne-van-kesterens-macbook.local ([89.130.83.193]) (authenticated bits=0) by smtp.opera.com (8.13.4/8.13.4/Debian-3sarge3) with ESMTP id n3E83WaA005001 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Tue, 14 Apr 2009 08:03:40 GMT
Date: Tue, 14 Apr 2009 10:03:31 +0200
To: "Sylvain Hellegouarch" <sh@defuze.org>, "Greg Wilkins" <gregw@webtide.com>
From: "Anne van Kesteren" <annevk@opera.com>
Organization: Opera Software ASA
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org>
Content-Transfer-Encoding: 7bit
Message-ID: <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local>
In-Reply-To: <49E428DD.3070803@defuze.org>
User-Agent: Opera Mail/9.63 (MacIntel)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 08:02:45 -0000

On Tue, 14 Apr 2009 08:10:37 +0200, Sylvain Hellegouarch <sh@defuze.org>  
wrote:
> Indeed. I actually wonder in fact if the problem hasn't been addressed  
> the wrong way around. It seems to me that we've had to discuss so far of  
> a protocol that could basically work out for everything. But could it? I  
> guess most people here have at some point implemented various protocols  
> and feel like WS will not be good at supporting any of them without  
> major headaches and hacks.

This hasn't really been demonstrated. Some people feel that what goes over  
the wire should be identifiable as an intrinsic part of the protocol, but  
I'm still out as to why this is important. (I.e. something more than just  
"string" vs "bytes" vs "...".)

One concern that was raised is that library authors might not work  
together to do things in common. However, library authors already work  
together on common things, such as supporting the same query language,  
ways of allowing multiple libraries to be used by a single page, et  
cetera. There's a whole "alliance" based around this.


> The question I have now is, what does precisely stop browsers from  
> offering a raw TCP socket API at the Javascript level?

The same-origin security policy.


> The answer to those so far has been WebSocket but is it the right way?  
> The endless nitpicking of this list should probably be warning enough in  
> my opinion that it might be the wrong path.

It seems to me that the metadata point is endlessly repeated. I'm not sure  
if that should be an indication of "wrong path" or "clearly entrenched".


> Personally I'd rather have a direct TCP socket access from the browser  
> and let organisations decide if they should allow specific ports to be  
> opened. Afterall, tunelling through HTTP might end up with those  
> organisation sniffing traffic and deny HTTP Upgrade anyway.

The closest to raw TCP socket access we can offer is something that looks  
like Web Sockets.


-- 
Anne van Kesteren
http://annevankesteren.nl/

From mjs@apple.com  Tue Apr 14 01:27:24 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B05103A6838 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 01:27:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.139
X-Spam-Level: 
X-Spam-Status: No, score=-106.139 tagged_above=-999 required=5 tests=[AWL=0.460, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id SRFQzz5jk8Hv for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 01:27:23 -0700 (PDT)
Received: from mail-out4.apple.com (mail-out4.apple.com [17.254.13.23]) by core3.amsl.com (Postfix) with ESMTP id E5B843A6814 for <hybi@ietf.org>; Tue, 14 Apr 2009 01:27:23 -0700 (PDT)
Received: from relay14.apple.com (relay14.apple.com [17.128.113.52]) by mail-out4.apple.com (Postfix) with ESMTP id 0A9D25F16B67 for <hybi@ietf.org>; Tue, 14 Apr 2009 01:28:33 -0700 (PDT)
Received: from relay14.apple.com (unknown [127.0.0.1]) by relay14.apple.com (Symantec Brightmail Gateway) with ESMTP id E53C72808B for <hybi@ietf.org>; Tue, 14 Apr 2009 01:28:32 -0700 (PDT)
X-AuditID: 11807134-a4853bb000000ff0-1d-49e449301ed9
Received: from et.apple.com (et.apple.com [17.151.62.12]) by relay14.apple.com (Apple SCV relay) with ESMTP id B947A28087 for <hybi@ietf.org>; Tue, 14 Apr 2009 01:28:32 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [17.151.95.111] by et.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI300FEJ0VK9W00@et.apple.com> for hybi@ietf.org; Tue, 14 Apr 2009 01:28:32 -0700 (PDT)
Message-id: <694727E2-C898-4119-A265-6351E167A8A8@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Anne van Kesteren <annevk@opera.com>
In-reply-to: <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local>
Date: Tue, 14 Apr 2009 01:28:31 -0700
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 08:27:24 -0000

On Apr 14, 2009, at 1:03 AM, Anne van Kesteren wrote:

>
>
>> Personally I'd rather have a direct TCP socket access from the  
>> browser and let organisations decide if they should allow specific  
>> ports to be opened. Afterall, tunelling through HTTP might end up  
>> with those organisation sniffing traffic and deny HTTP Upgrade  
>> anyway.
>
> The closest to raw TCP socket access we can offer is something that  
> looks like Web Sockets.

Actually, we could offer closer to raw TCP socket access by providing  
a raw undelimited octet stream without message boundaries or text  
encoding, after the initial HTTP-upgrade handshake.  I do not think  
that this would actually be more useful however. I am sympathetic to  
imposing the minimal structure needed to act as a good transport, both  
for trivial direct use and for layering more structured protocols on  
top. But I do think that minimal amount of structure is greater than  
0. Partitioning message boundaries and offering direct support for  
sending and receiving text and not just bytes are among the useful  
minimal structure to provide. In my opinion, optional but well-defined  
subprotocol identification is another way, since it enabled building  
more structured protocols on top in a clear way.

Here's one way to think about it: a protocol with too much structure  
can make it hard to build other structured protocols on top. It is too  
opinionated, and it becomes hard to go against the grain. But a  
protocol with 0 structure also makes things difficult, because then  
it's hard to coordinate between independently controlled endpoints.

I think multiplexing multiple message streams, though it may seem  
useful at the protocol level, is not a critical building block. First,  
if WebSocket allows for reasonable protocol negotiation, one can  
easily build a multi-WebSocket-over-WebSocket protocol with little  
additional overhead. Second, opening multiple WebSockets to the same  
server is not necessarily a disaster. Third, for the "many widgets in  
one page" scenario, it's likely the widgets will actually want to talk  
to different servers. And finally, for a fully custom protocol,  
multiplexing can be implemented at the client/library level for  
application-specific needs. And I think supporting multiplexing  
directly in the base protocol may actually make life harder for cases  
like Atom-over-WebSocket, CalDAV-over-WebSocket, Jabber-over-WebSocket  
and other such purpose-specific but widely applicable protocols that  
we may see invented. Such protocols would have to work around the  
extra complexity imposed by the protocol-level multiplexing.

Regards,
Maciej


From annevk@opera.com  Tue Apr 14 02:53:02 2009
Return-Path: <annevk@opera.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D7F5F3A69E2 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 02:53:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7vy+CMZWkRIK for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 02:53:02 -0700 (PDT)
Received: from smtp.opera.com (sam.opera.com [213.236.208.81]) by core3.amsl.com (Postfix) with ESMTP id A631E3A67A1 for <hybi@ietf.org>; Tue, 14 Apr 2009 02:53:01 -0700 (PDT)
Received: from anne-van-kesterens-macbook.local ([89.130.83.193]) (authenticated bits=0) by smtp.opera.com (8.13.4/8.13.4/Debian-3sarge3) with ESMTP id n3E9rvPE010753 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Tue, 14 Apr 2009 09:53:59 GMT
Date: Tue, 14 Apr 2009 11:53:55 +0200
To: "Maciej Stachowiak" <mjs@apple.com>
From: "Anne van Kesteren" <annevk@opera.com>
Organization: Opera Software ASA
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local> <694727E2-C898-4119-A265-6351E167A8A8@apple.com>
Content-Transfer-Encoding: 7bit
Message-ID: <op.usde35il64w2qv@anne-van-kesterens-macbook.local>
In-Reply-To: <694727E2-C898-4119-A265-6351E167A8A8@apple.com>
User-Agent: Opera Mail/9.63 (MacIntel)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 09:53:02 -0000

On Tue, 14 Apr 2009 10:28:31 +0200, Maciej Stachowiak <mjs@apple.com>  
wrote:
> In my opinion, optional but well-defined subprotocol identification is  
> another way, since it enabled building more structured protocols on top  
> in a clear way.

It seems to me that the path can be used for that, but that is not formal  
enough for your taste currently as I understand it (it's not  
well-defined). Any ideas as to what would work better? Or maybe as to how  
to make the path fit the criteria?


-- 
Anne van Kesteren
http://annevankesteren.nl/

From sh@defuze.org  Tue Apr 14 03:03:46 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6DC113A6B4D for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:03:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.215
X-Spam-Level: 
X-Spam-Status: No, score=-2.215 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id OcnIm-ljiMAl for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:03:45 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 9440A3A6B09 for <hybi@ietf.org>; Tue, 14 Apr 2009 03:03:45 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3EA4uqB014370; Tue, 14 Apr 2009 05:04:56 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n3EA4usS014368; Tue, 14 Apr 2009 12:04:56 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Tue, 14 Apr 2009 12:04:56 +0200 (CEST)
Message-ID: <54928.193.253.216.132.1239703496.squirrel@mail1.webfaction.com>
In-Reply-To: <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local>
Date: Tue, 14 Apr 2009 12:04:56 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: "Anne van Kesteren" <annevk@opera.com>
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 10:03:46 -0000

> On Tue, 14 Apr 2009 08:10:37 +0200, Sylvain Hellegouarch <sh@defuze.org>
> wrote:
>> Indeed. I actually wonder in fact if the problem hasn't been addressed
>> the wrong way around. It seems to me that we've had to discuss so far of
>> a protocol that could basically work out for everything. But could it? I
>> guess most people here have at some point implemented various protocols
>> and feel like WS will not be good at supporting any of them without
>> major headaches and hacks.
>
> This hasn't really been demonstrated. Some people feel that what goes over
> the wire should be identifiable as an intrinsic part of the protocol, but
> I'm still out as to why this is important. (I.e. something more than just
> "string" vs "bytes" vs "...".)

I would assume for multiplexing over one single WS connection. Mind you
the protocol could also impose connections can't be shared.

>
> One concern that was raised is that library authors might not work
> together to do things in common. However, library authors already work
> together on common things, such as supporting the same query language,
> ways of allowing multiple libraries to be used by a single page, et
> cetera. There's a whole "alliance" based around this.
>
>
>> The question I have now is, what does precisely stop browsers from
>> offering a raw TCP socket API at the Javascript level?
>
> The same-origin security policy.

I can appreciate this being a serious issue but somehow I'm fuzzy as to
why the response to that is a new base protocol other protocols should
interface with. From the browser's perspective, I can see why this is so
simple when I go through the WS specification but from a server's
perspective this means adding another layer which means:

* Having to understand for each protocol (XMPP, IMAP, etc.) the best way
to channel said protocols into the WS constraints.
* Having to gauge how to scale such infrastructure.

I find it odd to throw years of experience from the server side to fix one
security issue (even such a serious one).

> The closest to raw TCP socket access we can offer is something that looks
> like Web Sockets.
>

I find that seriously limitating.

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From annevk@opera.com  Tue Apr 14 03:18:27 2009
Return-Path: <annevk@opera.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7A7563A6AB2 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:18:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HHl2Gru12Ie8 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:18:26 -0700 (PDT)
Received: from smtp.opera.com (sam.opera.com [213.236.208.81]) by core3.amsl.com (Postfix) with ESMTP id 648D13A6AEC for <hybi@ietf.org>; Tue, 14 Apr 2009 03:18:26 -0700 (PDT)
Received: from anne-van-kesterens-macbook.local ([89.130.83.193]) (authenticated bits=0) by smtp.opera.com (8.13.4/8.13.4/Debian-3sarge3) with ESMTP id n3EAJQLO011816 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Tue, 14 Apr 2009 10:19:29 GMT
Date: Tue, 14 Apr 2009 12:19:25 +0200
To: sh@defuze.org
From: "Anne van Kesteren" <annevk@opera.com>
Organization: Opera Software ASA
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local> <54928.193.253.216.132.1239703496.squirrel@mail1.webfaction.com>
Content-Transfer-Encoding: 7bit
Message-ID: <op.usdgani764w2qv@anne-van-kesterens-macbook.local>
In-Reply-To: <54928.193.253.216.132.1239703496.squirrel@mail1.webfaction.com>
User-Agent: Opera Mail/9.63 (MacIntel)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 10:18:27 -0000

On Tue, 14 Apr 2009 12:04:56 +0200, Sylvain Hellegouarch <sh@defuze.org>  
wrote:
>> The same-origin security policy.
>
> I can appreciate this being a serious issue but somehow I'm fuzzy as to
> why the response to that is a new base protocol other protocols should
> interface with.

I'm not necessarily opposed to supporting other protocols natively in the  
browser.


> From the browser's perspective, I can see why this is so
> simple when I go through the WS specification but from a server's
> perspective this means adding another layer which means:
>
> * Having to understand for each protocol (XMPP, IMAP, etc.) the best way
> to channel said protocols into the WS constraints.
> * Having to gauge how to scale such infrastructure.

They would need to fit similar constraints at least if they were to be  
exposed to JavaScript.


> I find it odd to throw years of experience from the server side to fix  
> one security issue (even such a serious one).

Odd or not, it remains a design constraint.


>> The closest to raw TCP socket access we can offer is something that  
>> looks like Web Sockets.
>
> I find that seriously limitating.

It is, but there does not seem to be a better way either.


-- 
Anne van Kesteren
http://annevankesteren.nl/

From sh@defuze.org  Tue Apr 14 03:38:10 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C0C343A67AA for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:38:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.222
X-Spam-Level: 
X-Spam-Status: No, score=-2.222 tagged_above=-999 required=5 tests=[AWL=0.043,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qvi48UFw5xDX for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 03:38:10 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 03DFC3A694D for <hybi@ietf.org>; Tue, 14 Apr 2009 03:38:09 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3EAdJIt023839; Tue, 14 Apr 2009 05:39:20 -0500
Message-ID: <49E467D6.4060008@defuze.org>
Date: Tue, 14 Apr 2009 12:39:18 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090318)
MIME-Version: 1.0
To: Anne van Kesteren <annevk@opera.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local> <54928.193.253.216.132.1239703496.squirrel@mail1.webfaction.com> <op.usdgani764w2qv@anne-van-kesterens-macbook.local>
In-Reply-To: <op.usdgani764w2qv@anne-van-kesterens-macbook.local>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 10:38:10 -0000

>> From the browser's perspective, I can see why this is so
>> simple when I go through the WS specification but from a server's
>> perspective this means adding another layer which means:
>>
>> * Having to understand for each protocol (XMPP, IMAP, etc.) the best way
>> to channel said protocols into the WS constraints.
>> * Having to gauge how to scale such infrastructure.
>
> They would need to fit similar constraints at least if they were to be 
> exposed to JavaScript.

Would it? If there was direct access to raw TCP sockets there would be 
no requirements for a change in infrastructure, it would only mean 
client libs have to be written for each supported protocol, like in any 
other language.

>>> The closest to raw TCP socket access we can offer is something that 
>>> looks like Web Sockets.
>>
>> I find that seriously limitating.
>
> It is, but there does not seem to be a better way either.

Since I have no real knowledge of the browsers' design constraints, I 
can't really offer alternatives.

- Sylvain



From jamie@shareable.org  Tue Apr 14 07:16:37 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 74A6B3A6AF6 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 07:16:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.266
X-Spam-Level: 
X-Spam-Status: No, score=-4.266 tagged_above=-999 required=5 tests=[AWL=-1.667, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id W7wwNUQJMlRf for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 07:16:36 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id B2F0B3A697D for <hybi@ietf.org>; Tue, 14 Apr 2009 07:16:36 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LtjSK-00076d-IO; Tue, 14 Apr 2009 15:17:44 +0100
Date: Tue, 14 Apr 2009 15:17:44 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Sylvain Hellegouarch <sh@defuze.org>
Message-ID: <20090414141744.GA26621@shareable.org>
References: <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <49E43D13.60308@webtide.com> <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 14:16:37 -0000

Sylvain Hellegouarch wrote:
> > The conversation has been worthwhile (at least for me).  I'm
> > moving away from my "let's not pick winners" advocacy towards
> > a base protocol that does support sharing and multiplexing.
> 
> I'm still not convinced about the idea of a base protocol myself.

WebSockets is a base protocol.

Are you advocating raw TCP?

-- Jamie

From jamie@shareable.org  Tue Apr 14 07:42:55 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1AE5928C14B for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 07:42:55 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.137
X-Spam-Level: 
X-Spam-Status: No, score=-4.137 tagged_above=-999 required=5 tests=[AWL=-1.538, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id wNjN57tPCvR7 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 07:42:54 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 19A8328C143 for <hybi@ietf.org>; Tue, 14 Apr 2009 07:42:54 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Ltjrm-0007JP-GN; Tue, 14 Apr 2009 15:44:02 +0100
Date: Tue, 14 Apr 2009 15:44:02 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090414144402.GB26621@shareable.org>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 14:42:55 -0000

Ian Hickson wrote:
> I agree. But who mentioned scalable TCP servers for binary protocols? I 
> want to let people like my dad write little scripts that do insignificant 
> little things, just like today people write CGI scripts for all kinds of 
> stuff. I want to let 12 year olds write multiplayer games. I want to let 
> someone who picked up HTML on Monday and CSS on Wednesday try their hand 
> at WebSockets on Friday. I want to support the intern whose boss told him 
> to put one of those interactive chat widgets on the intranet.
> 
> > This is likely to require multi-threading in most languages.
> 
> It's easy to write simple TCP servers using a select() loop without 
> threads.

It's easy to write a simple HTTP server using a select() loop too, if
you know unix, or Perl, or Python.  There are lots of bad ones around
to prove it.  It's even easier with fork().

But mostly people use big HTTP servers written by other people and hang
applications off the back now, using CGI/FastCGI/WSGI/mod_*/ etc.

I expect your dad / 12 year old / person who picks up HTML on Monday
will use Apache or similar with a CGI or similar backend to their
favoured programming language.

They probably don't care much about HTTP, just that it calls their
application when a page is requested.  And they probably don't know
anything about unix system calls.

So on Friday they'd rather use Apache with a "hybi-CGI" backend to
their favoured programming language, than write their own select() loop.

The important thing for that scenario is that "hybi-CGI" must be easy
to understand and use.  Whether that's because it's so trivial anyone
can write TCP/WebSockets in good programming languages, or because the
protocol fits into widely used CGI-like frameworks in a way people
understand.

Right now, AJAX is widely used, and polling-AJAX is widely used for
server-driven updates to web pages because people find it easy to
understand, and it's easy to use existing frameworks - just another
tiny web page (fragment) to serve.

Indeed a lot of people think polling-AJAX is the most efficient way to
keep bits of a web page updated.  From a CGI-writing point of view,
that's true.  (From my point of view it's obviously false.)

For those people (imho the majority), what's most important is that
their preferred CGI-like framework supports the new protocols and
doesn't change the way they program too much.

Of course we'd like to support people who are happy writing their own
low-level server implementations too, and work well in that case.

> To reiterate: I have no objection to there being many protocols 
> standardised here. But I do want to make sure that at least one of them 
> enables an inexperienced programmer to bang out an interactive toy in a 
> few days without having to use libraries, just like they can bang out a 
> CGI script today and get interaction in forms. Currently, of the various 
> proposals, only one of them comes even close to this, IMHO.

Which one?

(I've designed my flavour of asynchronous HTTP to fit banging out
extended-CGI/etc. services like a glove, while also addressing
TCP-style streaming/messaging services.  I think WebSockets (without
metadata) does not fit the CGI-style usage as well, but it's
interesting to see the perspectives and differing requirements being
expressed.)

-- Jamie

From jamie@shareable.org  Tue Apr 14 08:31:35 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EE2A33A6DCC for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 08:31:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.028
X-Spam-Level: 
X-Spam-Status: No, score=-4.028 tagged_above=-999 required=5 tests=[AWL=-1.429, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MnVE92FRKFWW for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 08:31:35 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id F0DF23A6E28 for <hybi@ietf.org>; Tue, 14 Apr 2009 08:31:34 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Ltkct-0007of-Nm; Tue, 14 Apr 2009 16:32:43 +0100
Date: Tue, 14 Apr 2009 16:32:43 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090414153243.GD26621@shareable.org>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49E3D937.5020009@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 15:31:36 -0000

Greg Wilkins wrote:
> For example, XMPP has great market share - except when it comes to
> webchat - where there are now a plethora of alternative protocols
> for chat - all trying to work out there own way to tunnel through
> firewalls/gateways and get into the browser DOM.
>
> But fundamentally the issue we face is not *what* protocol to use,
> but how to get it through the internet and into the browser DOM.

I foresee WebSocket in its current form will hit the same problem.
How to get past firewalls and transparent HTTP proxies.

So will any other proposal, including rHTTP and mine for
bidirectional/async HTTP.

"The problem of firewalls and transparent proxies in the existing
infrastructure" is an important topic to address and I think it
deserves analysis _independent_ of API and low-level protocol.

Here are a few things to consider:

    Corporate firewalls sometimes block all TCP connections,
    except for port 80 and 443.

    Some firewalls sniff HTTP URLs, HTTP headers and even the content
    (e.g. checking for viruses), and block based on those.

    Transparent HTTP proxies and proxy-caches on port 80 are common
    both in corporate settings and home internet connections.  Maybe
    not the majority of connections (I have no figures), but enough to
    matter.

    Explicit HTTP proxies which pass HTTP CONNECT requests often
    limit that to port 443.

    I would be surprised if any widely deployed HTTP proxies, explicit
    or transparent, currently work with HTTP UPGRADE.

    The Bayeaux people seem to have decided that long-GET is more
    reliable than slow-streaming-GET, because some HTTP proxies don't
    forward any of the response until they have received the whole
    response.  (Is that right?)

    Any apparently-binary protocol which is opaque to intermediate
    firewalls/proxies is likely to be blocked at many sites who are
    currently able to use protocols which tunnel over HTTP.  I suspect
    blanket blocking is more likely when there's no distinguishing
    metadata in the protocol for a firewall to match on.

> But I think there will only be 1 opportunity this decade to make
> significant changes to the firewalls/gateways of the internet.
> If that is true, then is websocket the best change to make?

I think there is 0 opportunity to change the whole internet quickly.

We need to at least _talk about_ whether we care about a new protocol
interoperating with the existing firewalls/gateways, and if we want to
provide an upgrade path to help those work with new protocols better.

Issues for specific protocols to consider which might be of concern:

    - Can you make it tunnel over HTTP when needed?
    - How easily?
    - What are the problems and other issues when doing that?

It's always possible to tunnel over HTTP, because you can tunnel TCP
over HTTP, even without CONNECT or UPGRADE), but for some protocols
it's a poor fit (TCP!), and some fit tunnelling relatively well.  Some
are even reasonable candidates for firewalls to filter selectively and
proxies to intercept and cache.

-- Jamie

From sh@defuze.org  Tue Apr 14 09:46:54 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 245263A6D97 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 09:46:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.228
X-Spam-Level: 
X-Spam-Status: No, score=-2.228 tagged_above=-999 required=5 tests=[AWL=0.038,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id V-5oCVu0Qhwp for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 09:46:53 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id DFAEC3A6CDB for <hybi@ietf.org>; Tue, 14 Apr 2009 09:46:28 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3EGlcia014757; Tue, 14 Apr 2009 11:47:39 -0500
Message-ID: <49E4BE29.8030800@defuze.org>
Date: Tue, 14 Apr 2009 18:47:37 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090318)
MIME-Version: 1.0
To: Jamie Lokier <jamie@shareable.org>
References: <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <49E43D13.60308@webtide.com> <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com> <20090414141744.GA26621@shareable.org>
In-Reply-To: <20090414141744.GA26621@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 16:46:54 -0000

Jamie Lokier a écrit :
> Sylvain Hellegouarch wrote:
>   
>>> The conversation has been worthwhile (at least for me).  I'm
>>> moving away from my "let's not pick winners" advocacy towards
>>> a base protocol that does support sharing and multiplexing.
>>>       
>> I'm still not convinced about the idea of a base protocol myself.
>>     
>
> WebSockets is a base protocol.
>
> Are you advocating raw TCP?
>   
I am yes.

- Sylvain

From mario.balibrera@gmail.com  Tue Apr 14 09:56:39 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B347D28C0E1 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 09:56:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 1AznkU9iJhC0 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 09:56:38 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.225]) by core3.amsl.com (Postfix) with ESMTP id A2C633A6847 for <hybi@ietf.org>; Tue, 14 Apr 2009 09:56:38 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2309319rvb.49 for <hybi@ietf.org>; Tue, 14 Apr 2009 09:57:50 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=/MJdZUK0hFlO/rMWnNmkL0hS4DIQG0QPVhTpIdJrQYA=; b=gFQkhBXeK/tyB0AJRYIqMf09RjNw+2TdHwnKWZw5mhRcoIjMHziukfv1cLPhQXc7Gq 1n7GbJpvfub+ckshxytY9TG3rj5gS4lcc8hEsz6xn/y7RQRDDkRyTTYsQ5aFrDWS7Gpj U+R4NtHtiIowwR5WtNl2O+UwbCARGuG/+eoLs=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=Ti5q2D9IkXXJbl7J0rXKbo9fcHh8Nn6+GQiwqwk53f7F5MuN/h3qdjHOooJ7iPcwuj NiLyTTQdd4l0vIwTtTI0b5bx8s2xWFEozfHlf8BIkeYfl7OpUeSUmsTTXbagIMJfCAfI oGH5mcn0LPgPcYxYRGxaT/GijHTdn+pJgIT+c=
MIME-Version: 1.0
Received: by 10.114.94.12 with SMTP id r12mr3755467wab.229.1239728270304; Tue,  14 Apr 2009 09:57:50 -0700 (PDT)
In-Reply-To: <49E4BE29.8030800@defuze.org>
References: <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <49E43D13.60308@webtide.com> <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com> <20090414141744.GA26621@shareable.org> <49E4BE29.8030800@defuze.org>
Date: Tue, 14 Apr 2009 09:57:50 -0700
Message-ID: <79ea848f0904140957n670bf3f1v8e6f4051f6137a23@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Sylvain Hellegouarch <sh@defuze.org>
Content-Type: multipart/alternative; boundary=0016364587282b8873046786bb9b
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 16:56:39 -0000

--0016364587282b8873046786bb9b
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

* Having to understand for each protocol (XMPP, IMAP, etc.) the best way
to channel said protocols into the WS constraints.
* Having to gauge how to scale such infrastructure.

Lightweight WebSocket proxies exist (check out the dez project), and are
trivial to deploy.

For those people (imho the majority), what's most important is that
their preferred CGI-like framework supports the new protocols and
doesn't change the way they program too much.

The fact is, the people who want CGI-like interactions already have plenty
of tools that perfectly fit their use-case, and they don't need another one=
.
We're all here because the current toolset does not sufficiently address ou=
r
needs. We don't have to create something incompatible with existing
frameworks, but the very nature of CGI is in diametric opposition to the
concept of a streaming protocol. It's clear that people working in CGI-like
frameworks will _have_ to change the way they program, because they will be
writing applications that are fundamentally different.

Note that different !=3D harder. Anyone in the world can write a CGI script
because anyone in the world can type "cgi script" into google and get a
thousand code snippets to paste in. Writing actual network applications is =
a
big secret only because no one's writing the tutorials. (We should start).

Mario Balibrera

On Tue, Apr 14, 2009 at 9:47 AM, Sylvain Hellegouarch <sh@defuze.org> wrote=
:

> Jamie Lokier a =E9crit :
>
>> Sylvain Hellegouarch wrote:
>>
>>
>>> The conversation has been worthwhile (at least for me).  I'm
>>>> moving away from my "let's not pick winners" advocacy towards
>>>> a base protocol that does support sharing and multiplexing.
>>>>
>>>>
>>> I'm still not convinced about the idea of a base protocol myself.
>>>
>>>
>>
>> WebSockets is a base protocol.
>>
>> Are you advocating raw TCP?
>>
>>
> I am yes.
>
> - Sylvain
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--0016364587282b8873046786bb9b
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

* Having to understand for each protocol (XMPP, IMAP, etc.) the best way<br=
>
to channel said protocols into the WS constraints.<br>
* Having to gauge how to scale such infrastructure.<br><br>Lightweight WebS=
ocket proxies exist (check out the dez project), and are trivial to deploy.=
<br><br>For those people (imho the majority), what&#39;s most important is =
that<br>

their preferred CGI-like framework supports the new protocols and<br>
doesn&#39;t change the way they program too much.<br><br>The fact is, the p=
eople who want CGI-like interactions already have plenty of tools that perf=
ectly fit their use-case, and they don&#39;t need another one. We&#39;re al=
l here because the current toolset does not sufficiently address our needs.=
 We don&#39;t have to create something incompatible with existing framework=
s, but the very nature of CGI is in diametric opposition to the concept of =
a streaming protocol. It&#39;s clear that people working in CGI-like framew=
orks will _have_ to change the way they program, because they will be writi=
ng applications that are fundamentally different.<br>
<br>Note that different !=3D harder. Anyone in the world can write a CGI sc=
ript because anyone in the world can type &quot;cgi script&quot; into googl=
e and get a thousand code snippets to paste in. Writing actual network appl=
ications is a big secret only because no one&#39;s writing the tutorials. (=
We should start).<br>
<br>Mario Balibrera<br><br><div class=3D"gmail_quote">On Tue, Apr 14, 2009 =
at 9:47 AM, Sylvain Hellegouarch <span dir=3D"ltr">&lt;<a href=3D"mailto:sh=
@defuze.org">sh@defuze.org</a>&gt;</span> wrote:<br><blockquote class=3D"gm=
ail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt =
0pt 0pt 0.8ex; padding-left: 1ex;">
Jamie Lokier a =E9crit :<div class=3D"im"><br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Sylvain Hellegouarch wrote:<br>
 =A0<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><blockquote class=
=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin=
: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">

The conversation has been worthwhile (at least for me). =A0I&#39;m<br>
moving away from my &quot;let&#39;s not pick winners&quot; advocacy towards=
<br>
a base protocol that does support sharing and multiplexing.<br>
 =A0 =A0 =A0<br>
</blockquote>
I&#39;m still not convinced about the idea of a base protocol myself.<br>
 =A0 =A0<br>
</blockquote>
<br>
WebSockets is a base protocol.<br>
<br>
Are you advocating raw TCP?<br>
 =A0<br>
</blockquote></div>
I am yes.<br><font color=3D"#888888">
<br>
- Sylvain</font><div><div></div><div class=3D"h5"><br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org" target=3D"_blank">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--0016364587282b8873046786bb9b--

From jamie@shareable.org  Tue Apr 14 10:14:18 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 51A643A68B0 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:14:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.932
X-Spam-Level: 
X-Spam-Status: No, score=-3.932 tagged_above=-999 required=5 tests=[AWL=-1.333, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RAuwSX-FQPQX for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:14:17 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 249E13A6E33 for <hybi@ietf.org>; Tue, 14 Apr 2009 10:14:16 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LtmEE-0000Lk-Tr; Tue, 14 Apr 2009 18:15:22 +0100
Date: Tue, 14 Apr 2009 18:15:22 +0100
From: Jamie Lokier <jamie@shareable.org>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
Message-ID: <20090414171522.GA32311@shareable.org>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 17:14:18 -0000

Thomson, Martin wrote:
> > Seems that each request/response would need its own BEEP channel. This
> > is because of the synchronous nature of BEEP channels(*). Channel
> > creation is mid-weight, but at least requires a round-trip before the
> > actual request could even begin. It might be possible to encode the
> > request within the channel creation request, but there is a 4k payload
> > limit for such messages.
> 

> [[MT]] BEEP does pipelining, just like HTTP.

Yes, that's why it sucks :-)

HTTP pipelining is a problem.  Have you noticed that no sane client
sends pipelined POSTs?  There's a reason.

Have you noticed that proxies do not aggregate multiple clients into
fewer connections to the same origin server?  Pipelining not being
asychronous is the main reason why they can't.

For widgets in a web page, the BEEP synchronous restriction forces
application workarounds for "large" messages - such as a large PNG (or
unwanted interference as soon as you mix widgets on a page) - and it
forces each application to have its own addressing layer _above_ the
base protocol if they have multiple widgets on a page all subscribing
for updates which can come in any order.

Basically it prevents lots of widgets from operating independently of
each other over the same channel.  Use separate channels (or server
connections) for a few widgets?  That doesn't scale to 100 or more
widgets per page, which is easy to reach with a bit of imagination.

> Of course, the synchronous restriction is artificial, and there are
> use cases that demand a greater degree of asynchrony.  Hence, the
> draft you refer to.

Unfortunately the asynchrony draft still has an artifical restriction
which makes it useless for independent requests.

> [[MT]] The "XML" parsers I've seen in implementations of BEEP are
> rarely as complex as you make out.  The control channel doesn't demand
> much complexity of a parser.  Now you can take that as an indication
> of the maturity of BEEP implementation if you like, I can't disagree
> about _that_ point...

To be a "correct" BEEP implementation, it needs to correctly parse
XML, including skipping a potential embedded DTD (whose lexical syntax
wasn't even precisely defined in the XML spec last time I looked),
accepting UTF-16 both endians, and things like that.

No problem if you have a good XML library handy, only if you don't.
(Still a bit heavy even if you do, compared with HTTP's headers.)

_Normal_ XML is easy to parse in any language, the problem is a
protocol implementation is allowed to send unusual XML, and a protocol
parser which doesn't handle that is non-compliant.

Is there a new BEEP planning to come out which specifies that the
irrelevant parts of XML will not appear on the wire?  That would be
helpful.  I'm sure I've seen that done in some other IETF documents.

> > There is no addressability in BEEP for channels or messages. I see two
> > ways two add one to a profile: 1) Add the address (and perhaps method)
> > as fields in the initialization message on channel create. This would
> > require a new channel for every request. 2) Encode the address in each
> > message (see above for using message/http)
> 
> [[MT]] If you are looking to interact with the same resource,
> identifying the resource at channel creation time might make sense.
> Alternatively, the MIME headers can be used to identify a resource for
> each request.

I agree with both of those; they make sense to me.

My only objection is that BEEP channel creation needs a round trip -
you can't create them at the request side as needed.

> > The overhead of BEEP seems like it would greater, though I admit I
> > haven't calculated it. I doubt it would be outrageous.
> 
> [[MT]] Greater than websockets?  That seems quite likely.  Depends
> on how much you are using from of what BEEP offers.

Overhead: the initial setup looks similar for each.  After that, BEEP
uses more bytes, but about the same number of network packets.

> > I don't know enough to know if flow control internally would be
> > needed. If so, BEEP has it.  If not, it is a little scary to implement!
> 
> [[MT]] If you want to avoid implementation, it's not actually that
> hard to send an occasional SEQ frame.  It's quite easy to avoid flow
> control being a problem.

Indeed.  Per channel flow control is important - if you have multiple
independent channels, otherwise problems you don't want to care about
bite some applications.

> However, I'll note that flow control in current implementations
> isn't like that - a small fixed window is common, which severely
> limits throughput.  Of course, that's a further sign of stack maturity
> more than anything else, there is nothing in the protocol preventing
> bandwidth-delay product estimation.

Agreed.  Also SCTP instead of TCP is promising - let the operating
system do the estimation for you better.  (There's some discussion of
HTTP over SCTP around, but nothing concrete yet that I know of.)

-- Jamie

From mario.balibrera@gmail.com  Tue Apr 14 10:17:17 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E2B5E3A6E33 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:17:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.298
X-Spam-Level: 
X-Spam-Status: No, score=-2.298 tagged_above=-999 required=5 tests=[AWL=-0.300, BAYES_00=-2.599, HTML_MESSAGE=0.001, J_CHICKENPOX_22=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AP41sqRUGYFO for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:17:16 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.230]) by core3.amsl.com (Postfix) with ESMTP id 90C8B3A68B0 for <hybi@ietf.org>; Tue, 14 Apr 2009 10:17:16 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2318636rvb.49 for <hybi@ietf.org>; Tue, 14 Apr 2009 10:18:28 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=zDoqcId754JiQi9eW7wQPT1S6m8+ZpwvXCTyZ94TsHY=; b=uyWTWXZ84f3gaEcaslQ02rfCLPaekkz946upoSMdrJVAt0XQegS+HG9FQXKxPhyZg4 jkLwT2R1OrIgUIuYuFhGQwnwatb0TzSCoa+lLG0FelERC3ixVNNUakhPvNRl+wcuxyux 7XbHIPukk8o+z2uUeIxGS6OWIKUXzx6XxcCfo=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=m18WhZulHSZdPpSsJShPtqhNO2c00P7FLEQv3cH62uhYfWKrX7H3KyPi0g1Ao2AeSs uB8+phbMtp43+PhUrtz1qGF8ODj8EfggE0SWn4YdRPSIVqPhHQNZOlU90EBjEJTjINhP qalacqsinCMoteFg0EMaihQ0uuVle0AelRZOM=
MIME-Version: 1.0
Received: by 10.114.181.6 with SMTP id d6mr3793420waf.8.1239729505460; Tue, 14  Apr 2009 10:18:25 -0700 (PDT)
In-Reply-To: <20090414153243.GD26621@shareable.org>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com> <20090414153243.GD26621@shareable.org>
Date: Tue, 14 Apr 2009 10:18:25 -0700
Message-ID: <79ea848f0904141018j59659f21n82944f3723409991@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Jamie Lokier <jamie@shareable.org>
Content-Type: multipart/alternative; boundary=00163646d987ca62ae0467870499
Cc: hybi@ietf.org
Subject: Re: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 17:17:18 -0000

--00163646d987ca62ae0467870499
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

   - Can you make it tunnel over HTTP when needed?
   - How easily?
   - What are the problems and other issues when doing that?

Currently, js.io works that way with WebSocket. If there is a native browser
implementation available (none yet, obviousy), it uses that to connect
directly to the targeted WS server. Otherwise, it speaks WebSocket protocol
on top of a TCPConnection object (from Orbited, APE, or any other project
imported into the global scope that implements the API), which connects to
its own TCP proxy that in turn connects to the WebSocket server. Easy Peasy.

It would be trivial to add a graceful fallback mechanism, so that it tries
the native WebSocket (as soon as we get one...) and in case of timeout or
immediate connection closure, falls back to HTTP tunneling.

Mario Balibrera

On Tue, Apr 14, 2009 at 8:32 AM, Jamie Lokier <jamie@shareable.org> wrote:

> Greg Wilkins wrote:
> > For example, XMPP has great market share - except when it comes to
> > webchat - where there are now a plethora of alternative protocols
> > for chat - all trying to work out there own way to tunnel through
> > firewalls/gateways and get into the browser DOM.
> >
> > But fundamentally the issue we face is not *what* protocol to use,
> > but how to get it through the internet and into the browser DOM.
>
> I foresee WebSocket in its current form will hit the same problem.
> How to get past firewalls and transparent HTTP proxies.
>
> So will any other proposal, including rHTTP and mine for
> bidirectional/async HTTP.
>
> "The problem of firewalls and transparent proxies in the existing
> infrastructure" is an important topic to address and I think it
> deserves analysis _independent_ of API and low-level protocol.
>
> Here are a few things to consider:
>
>    Corporate firewalls sometimes block all TCP connections,
>    except for port 80 and 443.
>
>    Some firewalls sniff HTTP URLs, HTTP headers and even the content
>    (e.g. checking for viruses), and block based on those.
>
>    Transparent HTTP proxies and proxy-caches on port 80 are common
>    both in corporate settings and home internet connections.  Maybe
>    not the majority of connections (I have no figures), but enough to
>    matter.
>
>    Explicit HTTP proxies which pass HTTP CONNECT requests often
>    limit that to port 443.
>
>    I would be surprised if any widely deployed HTTP proxies, explicit
>    or transparent, currently work with HTTP UPGRADE.
>
>    The Bayeaux people seem to have decided that long-GET is more
>    reliable than slow-streaming-GET, because some HTTP proxies don't
>    forward any of the response until they have received the whole
>    response.  (Is that right?)
>
>    Any apparently-binary protocol which is opaque to intermediate
>    firewalls/proxies is likely to be blocked at many sites who are
>    currently able to use protocols which tunnel over HTTP.  I suspect
>    blanket blocking is more likely when there's no distinguishing
>    metadata in the protocol for a firewall to match on.
>
> > But I think there will only be 1 opportunity this decade to make
> > significant changes to the firewalls/gateways of the internet.
> > If that is true, then is websocket the best change to make?
>
> I think there is 0 opportunity to change the whole internet quickly.
>
> We need to at least _talk about_ whether we care about a new protocol
> interoperating with the existing firewalls/gateways, and if we want to
> provide an upgrade path to help those work with new protocols better.
>
> Issues for specific protocols to consider which might be of concern:
>
>    - Can you make it tunnel over HTTP when needed?
>    - How easily?
>    - What are the problems and other issues when doing that?
>
> It's always possible to tunnel over HTTP, because you can tunnel TCP
> over HTTP, even without CONNECT or UPGRADE), but for some protocols
> it's a poor fit (TCP!), and some fit tunnelling relatively well.  Some
> are even reasonable candidates for firewalls to filter selectively and
> proxies to intercept and cache.
>
> -- Jamie
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--00163646d987ca62ae0467870499
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

=A0=A0 - Can you make it tunnel over HTTP when needed?<br>
 =A0 =A0- How easily?<br>
 =A0 =A0- What are the problems and other issues when doing that?<br><br>Cu=
rrently, <a href=3D"http://js.io">js.io</a> works that way with WebSocket. =
If there is a native browser implementation available (none yet, obviousy),=
 it uses that to connect directly to the targeted WS server. Otherwise, it =
speaks WebSocket protocol on top of a TCPConnection object (from Orbited, A=
PE, or any other project imported into the global scope that implements the=
 API), which connects to its own TCP proxy that in turn connects to the Web=
Socket server. Easy Peasy.<br>
<br>It would be trivial to add a graceful fallback mechanism, so that it tr=
ies the native WebSocket (as soon as we get one...) and in case of timeout =
or immediate connection closure, falls back to HTTP tunneling.<br><br>Mario=
 Balibrera<br>
<br><div class=3D"gmail_quote">On Tue, Apr 14, 2009 at 8:32 AM, Jamie Lokie=
r <span dir=3D"ltr">&lt;<a href=3D"mailto:jamie@shareable.org">jamie@sharea=
ble.org</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D=
"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padd=
ing-left: 1ex;">
Greg Wilkins wrote:<br>
&gt; For example, XMPP has great market share - except when it comes to<br>
&gt; webchat - where there are now a plethora of alternative protocols<br>
&gt; for chat - all trying to work out there own way to tunnel through<br>
&gt; firewalls/gateways and get into the browser DOM.<br>
&gt;<br>
&gt; But fundamentally the issue we face is not *what* protocol to use,<br>
&gt; but how to get it through the internet and into the browser DOM.<br>
<br>
I foresee WebSocket in its current form will hit the same problem.<br>
How to get past firewalls and transparent HTTP proxies.<br>
<br>
So will any other proposal, including rHTTP and mine for<br>
bidirectional/async HTTP.<br>
<br>
&quot;The problem of firewalls and transparent proxies in the existing<br>
infrastructure&quot; is an important topic to address and I think it<br>
deserves analysis _independent_ of API and low-level protocol.<br>
<br>
Here are a few things to consider:<br>
<br>
 =A0 =A0Corporate firewalls sometimes block all TCP connections,<br>
 =A0 =A0except for port 80 and 443.<br>
<br>
 =A0 =A0Some firewalls sniff HTTP URLs, HTTP headers and even the content<b=
r>
 =A0 =A0(e.g. checking for viruses), and block based on those.<br>
<br>
 =A0 =A0Transparent HTTP proxies and proxy-caches on port 80 are common<br>
 =A0 =A0both in corporate settings and home internet connections. =A0Maybe<=
br>
 =A0 =A0not the majority of connections (I have no figures), but enough to<=
br>
 =A0 =A0matter.<br>
<br>
 =A0 =A0Explicit HTTP proxies which pass HTTP CONNECT requests often<br>
 =A0 =A0limit that to port 443.<br>
<br>
 =A0 =A0I would be surprised if any widely deployed HTTP proxies, explicit<=
br>
 =A0 =A0or transparent, currently work with HTTP UPGRADE.<br>
<br>
 =A0 =A0The Bayeaux people seem to have decided that long-GET is more<br>
 =A0 =A0reliable than slow-streaming-GET, because some HTTP proxies don&#39=
;t<br>
 =A0 =A0forward any of the response until they have received the whole<br>
 =A0 =A0response. =A0(Is that right?)<br>
<br>
 =A0 =A0Any apparently-binary protocol which is opaque to intermediate<br>
 =A0 =A0firewalls/proxies is likely to be blocked at many sites who are<br>
 =A0 =A0currently able to use protocols which tunnel over HTTP. =A0I suspec=
t<br>
 =A0 =A0blanket blocking is more likely when there&#39;s no distinguishing<=
br>
 =A0 =A0metadata in the protocol for a firewall to match on.<br>
<br>
&gt; But I think there will only be 1 opportunity this decade to make<br>
&gt; significant changes to the firewalls/gateways of the internet.<br>
&gt; If that is true, then is websocket the best change to make?<br>
<br>
I think there is 0 opportunity to change the whole internet quickly.<br>
<br>
We need to at least _talk about_ whether we care about a new protocol<br>
interoperating with the existing firewalls/gateways, and if we want to<br>
provide an upgrade path to help those work with new protocols better.<br>
<br>
Issues for specific protocols to consider which might be of concern:<br>
<br>
 =A0 =A0- Can you make it tunnel over HTTP when needed?<br>
 =A0 =A0- How easily?<br>
 =A0 =A0- What are the problems and other issues when doing that?<br>
<br>
It&#39;s always possible to tunnel over HTTP, because you can tunnel TCP<br=
>
over HTTP, even without CONNECT or UPGRADE), but for some protocols<br>
it&#39;s a poor fit (TCP!), and some fit tunnelling relatively well. =A0Som=
e<br>
are even reasonable candidates for firewalls to filter selectively and<br>
proxies to intercept and cache.<br>
<br>
-- Jamie<br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</blockquote></div><br>

--00163646d987ca62ae0467870499--

From jamie@shareable.org  Tue Apr 14 10:44:01 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 469A43A6E35 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:44:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.849
X-Spam-Level: 
X-Spam-Status: No, score=-3.849 tagged_above=-999 required=5 tests=[AWL=-1.250, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tL3iOjFp7d1a for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 10:44:00 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 42B4928C12E for <hybi@ietf.org>; Tue, 14 Apr 2009 10:44:00 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Ltmh3-0000b0-Of; Tue, 14 Apr 2009 18:45:09 +0100
Date: Tue, 14 Apr 2009 18:45:09 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mario Balibrera <mario.balibrera@gmail.com>
Message-ID: <20090414174509.GB32311@shareable.org>
References: <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <49E43D13.60308@webtide.com> <49513.193.253.216.132.1239694957.squirrel@mail1.webfaction.com> <20090414141744.GA26621@shareable.org> <49E4BE29.8030800@defuze.org> <79ea848f0904140957n670bf3f1v8e6f4051f6137a23@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <79ea848f0904140957n670bf3f1v8e6f4051f6137a23@mail.gmail.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 17:44:01 -0000

Mario Balibrera wrote:
> We don't have to create something incompatible with existing
> frameworks, but the very nature of CGI is in diametric opposition to
> the concept of a streaming protocol.

No.  It's different, but there are ways to bring them quite close
together for a lot of practical applications.  It's far from
diametrically opposed, you just need good glue in the middle to hide
the protocol differences.

> It's clear that people working in CGI-like frameworks will _have_ to
> change the way they program, because they will be writing
> applications that are fundamentally different.

I disagree.  Some applications must be structured differently to work
well.  But many don't need radical changes, especially when efficiency
at the server is not a great concern.

For example, an application with:

    - Data stored in a database and/or files.
    - A HTML templating system.
    - A template containing HTML, references to other templates,
      and snippets of code generating some HTML based on the stored data.
    - The output must appear as either a whole web page, or embedded
      in part of a larger one.

That comprises probably 99%+ of dynamically generated web pages at the
moment (which don't update themselves).  Nearly everything which isn't
an interactive form fits that pattern.

Now you, the (CGI-level :-) programmer, are asked to do the same, but
it must update changed parts of the page in the browser, immediately
when the stored data updates.  (I'd say a lot of collaborative
editing, chatting and data presentation falls into that category.
Many widgets could do.  Are we still at 99%?)

Despite the use of asynchronous server events on the wire, it's
possible to implement that in CGI with the right glue.

In fact it's possible in some cases without changing the CGI script.

I'm told Microsoft already has a version of ASP.NET which does it by
regenerating whole pages on the server side and comparing with
previous output to decide when to send an update, and what to send.  I
don't know if it's smart enough to figure out _when_ to recalculate
those pages.

There are cleverer ways to do it with less calculation and only when
the requisite data changes, while still structuring the code as if
it's classic CGI and templating + database/file queries.

I'm not saying all applications should be written that way, but when
it works and especially when it works efficiently - why not?  It's a
good way to write applications when it works.

> Note that different != harder. Anyone in the world can write a CGI script
> because anyone in the world can type "cgi script" into google and get a
> thousand code snippets to paste in. Writing actual network applications is a
> big secret only because no one's writing the tutorials. (We should start).

I expect the same snippets as soon as we have CGI scripts (+ Python,
Java frameworks etc.) which do clever two-way interactive things with
widgets.

The only difference will be "this script requires hybi-CGI support in
the web server for the dynamic parts to work" or whatever the
framework is called.

If people would like more discussion on techniques for client update
which don't require long-running server-side applications at the CGI
level (only the web server itself must manage those connections on
behalf of the application), I could expand on it but I expect there's
plenty of people here who have implemented that sort of thing already,
and the web isn't short of comet tutorials.

-- Jamie

From mike@dierken.com  Tue Apr 14 11:35:38 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 739A63A6E18 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 11:35:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id c+EWQcqVL18B for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 11:35:33 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.232]) by core3.amsl.com (Postfix) with ESMTP id 41AD828C1F8 for <hybi@ietf.org>; Tue, 14 Apr 2009 11:35:33 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2351413rvb.49 for <hybi@ietf.org>; Tue, 14 Apr 2009 11:36:44 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.140.139.4 with SMTP id m4mr3281817rvd.292.1239734204528; Tue,  14 Apr 2009 11:36:44 -0700 (PDT)
In-Reply-To: <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com>
Date: Tue, 14 Apr 2009 11:36:44 -0700
Message-ID: <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com>
From: Mike Dierken <mike@dierken.com>
To: sh@defuze.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 18:35:38 -0000

I was thinking the applet would expose the 'web socket' API (or
equivalent) and the app would be built with JavaScript. This would be
a way to prototype an approach that created a sibling of
XMLHttpRequest, but meant for receiving unsolicited requests in a page
- something like XMLHttpServer (but without the XML part).



On Tue, Apr 14, 2009 at 12:06 AM, Sylvain Hellegouarch <sh@defuze.org> wrote:
>
>>
>>> The question I have now is, what does precisely stop browsers
>>> from offering a raw TCP socket API at the Javascript level?
>>>
>> I thought that was available, via object/applet style plug-ins.
>>
>
> But if that was the solution we wouldn't have this discussion. It's
> because people want to/have to write their application in javascript that
> there is a need for a better solution than what exists now.
>
> Perhaps the reason BEEP has never taken off is because it's so versatile
> that it ultimately creates a bigger mismatch. I'm afraid WS might follow
> the same trend.
>
> - Sylvain
> --
> Sylvain Hellegouarch
> http://www.defuze.org
>

From jamie@shareable.org  Tue Apr 14 12:02:57 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8A35E28C260 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:02:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.775
X-Spam-Level: 
X-Spam-Status: No, score=-3.775 tagged_above=-999 required=5 tests=[AWL=-1.176, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id TS5CeGr2ucuY for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:02:56 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 8ECB628C265 for <hybi@ietf.org>; Tue, 14 Apr 2009 12:02:13 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Ltnum-0001Kk-5O; Tue, 14 Apr 2009 20:03:24 +0100
Date: Tue, 14 Apr 2009 20:03:24 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Mark Lentczner <markl@lindenlab.com>
Message-ID: <20090414190324.GH32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <20090413172412.GA31677@shareable.org> <C893F610-2839-4D5F-B13C-4CDC1D1293FF@lindenlab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <C893F610-2839-4D5F-B13C-4CDC1D1293FF@lindenlab.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 19:02:57 -0000

Mark Lentczner wrote:
> On Apr 13, 2009, at 10:24 AM, Jamie Lokier wrote:
> >... and are dispatched to the appropriate widget, by whatever ad hoc  
> >addressing scheme.
> >
> >For my own pages I'm combining multiple "group subscriptions" into a  
> >hierarchy of address/etag pairs.
> >... My scheme is efficient when you have hundreds/thousands of  
> >widgets,
> 
> How would such a think work over the model of WebSockets, where all  
> you get is a single string per event?

You can overlay any protocol over that, just as you can overlay any
protocol on top of TCP.  That's why WebSockets is acceptable for
people who are happy to build everything on top themselves.

Unfortunately it does require a lot of code on top of WebSocket to do
the kind of updates I'm doing.  (Ironic as the point is network
efficiency and minimal updates).

How it would work over WebSockets for me is:

   - Wrap subscription requests as generic events, and update
     notifications too.  (Effectively: Multiple responses per request).

   - Assign request ids, so that updates could come in any order.

   - Send small updates directly, with rate limiting, and big updates
     (e.g. images and large HTML) as merely a notification that
     there's a new one, unless requested to send the new update fully.

   - Keepalive events every 20 seconds to keep NATs open, and to
     recover from network changes.  Only needed when no other events
     transferred.

   - Interleave independent event sources fairly to prevent starvation
     by a subset.

   - Split large events into multiple small events, so that large
     events (e.g. images and large HTML) don't block the timely
     updates of small ones.

   - Serialise all those events into a byte stream, using whatever
     byte stream protocol seems like a good idea.  (Probably done at
     the same time as large event splitting.)

   - Split the byte stream either on event boundaries, if the large
     event splitting is good enough, or on artificial boundaries if I
     need better information on progress of large events than
     WebSockets provides with its per-message callback.

   - Encapsulate the split byte stream into WebSocket messages.

   - Stick it back together at the other end, into independent streams
     of events, and each event is an independent stream of bytes.  Bit
     like HTTP :-)

Hmm.  So I could _use_ Websockets, but I'd have to implement quite a
lot on top of it in Javascript on the browser side, and also on the
server side.  I'm basically looking at it as TCP with slight differences.

If you want my hopefully desirable widgeting methodology
then you'd probably want to use my Javascript as it looks a lot to do
from scratch.

Now I don't mind writing all that today, because I need it anyway for
tunnelling over HTTP in existing browsers.

But it would be nice if the (imho highly desirable ;-) features from
my protocol were available to others without them needing yet another
big Javascript library in future browsers, and then only perhaps with
one web server and application backend.

>From the above, my guess is WebSockets look compelling for the "3
widgets on a page need updates" case, and for the "I like writing my
own protocols and having full control" case.

But they don't look compelling for "I want to write an attractive web
application with lots of updating documents and tiny widgets without
using a pre-existing application framework or large bit of
Javascript".

-- Jamie

From jamie@shareable.org  Tue Apr 14 12:16:09 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3E25D3A6A2B for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:16:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.71
X-Spam-Level: 
X-Spam-Status: No, score=-3.71 tagged_above=-999 required=5 tests=[AWL=-1.111,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QSmB84WRmbQd for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:16:08 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 53AAD3A6A8B for <hybi@ietf.org>; Tue, 14 Apr 2009 12:16:08 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1Lto8D-0001Op-NT; Tue, 14 Apr 2009 20:17:17 +0100
Date: Tue, 14 Apr 2009 20:17:17 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090414191717.GI32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49E3C476.7090005@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 19:16:09 -0000

Greg Wilkins wrote:
> Anyway, my point is, that when considering multiplexing
> events onto the same transport, we could consider the
> multi frame/tab/window issues as well.

Interesting point.

This is another good reason to do multiplexing and connection sharing
in the browser itself, instead of in Javascript.

I personally have about 200 tabs open in 20 windows on 2 desktops,
most of the time.  That's perhaps unusual.  Mostly to different sites
- things I'm reading and want to go back to - but 10 tabs open to a
single site isn't uncommon.

If you have 10 tabs open to a single site, then you _must_ have the
flow of events (including large events like images) being independent
to each tab/window.

That's because each tab/window is expected to behave independently of
the others, even if the others get stuck, or are uploading/downloading
some big file.

You can do that by either:

   - The browser opens separate connections for each window/tab.

   - The browser shares connections for each window/tab and handles
     multiplexing automatically.  This requires the low-level
     protocols to support multiplexing.

   - The Javascript in every tab/window coordinates to somehow share
     connections.  This is extremely difficult to get right even when
     it's allowed by security rules.  What happens when the Javascript
     for the first tab/window is closed and it's acting for all the
     others?  Messy.

The same applies for IFRAMEs to different sites.  However, multiple
tab/window is surely even more popular.

(The same also applies to different parts of a page being authored by
different people using different Javascript frameworks, so not using
Javascript-level shared dispatch protocol.  But that's been covered
before.)

Right now, the browser handles sharing HTTP connections automatically
and mostly transparently.  This is good.

I think it should handle sharing of whatever else we're proposing too,
for all the advantages which come from that instead of forcing rather
complex Javascript implementations to do it badly - or fail to do it
reliably at all in many cases.  (Most Javascript authors are not experts).

The only realistically usable alternative is lots of separate
connections, and apart from anything else, that's a waste of
resources.  We're not talking 3 connections here, we're talking 100
idle connections for a single busy browser (10 tabs, 100 widgets from
10 different authors per tab).  Which is absurd, and won't scale over
the internet's NAT routers, let alone the actual computers and
connection setup time :-)

-- Jamie

From jamie@shareable.org  Tue Apr 14 12:21:25 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 60B6C3A6E31 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:21:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.652
X-Spam-Level: 
X-Spam-Status: No, score=-3.652 tagged_above=-999 required=5 tests=[AWL=-1.053, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Db9nP-njqaiu for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:21:24 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 838B53A6B4F for <hybi@ietf.org>; Tue, 14 Apr 2009 12:21:24 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LtoDJ-0001Qo-F1; Tue, 14 Apr 2009 20:22:33 +0100
Date: Tue, 14 Apr 2009 20:22:33 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Jonas Sicking <jonas@sicking.cc>
Message-ID: <20090414192233.GJ32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com> <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com> <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com> <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com> <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com> <63df84f0904131829j5bc1b85amf8d24fdd30644f1d@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <63df84f0904131829j5bc1b85amf8d24fdd30644f1d@mail.gmail.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 19:21:25 -0000

Jonas Sicking wrote:
> If the API could be something like
> 
> socket = new WebSocket('ws://example.com/chat', 'DCP');
> socket.onprotocol = chat_handler;
> 
> for the people that want checking. And
> 
> socket = new WebSocket('ws://example.com/chat');
> socket.onprotocol = iKnowWhoImTalkingToAndDontNeedNoStinkingChecking_handler;
> 
> for the people that are using their own home-brewed protocol. This
> could use some magic value as the type (a'la
> "application/octet-stream") or leave out the protocol indicator
> entirely.

I agree with all of this.

Think of it as really basic optional type-checking.

Handy when you're connecting to _someone else's_ service URL, which
might have been supplied by a random user.

It's also handy for protocol negotiation in future, if any protocols
are standardised it gives you a change to try for the favoured one
first then fall back, even if the standardised protocols forgot to
include any negotiation themselves.  (Or in other words: they don't have to).

> I don't have an opinion on if this type should be transferred to the
> server which would be responsible for the checking, or if the type is
> transferred to the client which would do the checking, or if its
> transferred both ways for added goodness.

I say send it to the server, which can check and refuse.  Same as the
URL.  That lets the server accept more than one protocol.

> But the point is that the type is transferred at a specified location.
> And IMHO that specified location should be separated from other fields
> with other purposes. My understanding is that the path is intended to
> allow for several websocket services from the same origin, thus I
> would prefer if the protocol is separate from the path.

Same reasoning here.  If the type is optional that's good for the
folks who don't care too.

Or instead of protocol type, a WebSockets method to add arbitrary HTTP
headers to the UPGRADE request, similar to XHR?

-- Jamie

From jamie@shareable.org  Tue Apr 14 12:29:41 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C751E28C235 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:29:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.599
X-Spam-Level: 
X-Spam-Status: No, score=-3.599 tagged_above=-999 required=5 tests=[AWL=-1.000, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id VPW410ad9Zap for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:29:41 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 0337728C231 for <hybi@ietf.org>; Tue, 14 Apr 2009 12:29:41 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LtoLK-0001Te-36; Tue, 14 Apr 2009 20:30:50 +0100
Date: Tue, 14 Apr 2009 20:30:50 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Sylvain Hellegouarch <sh@defuze.org>
Message-ID: <20090414193050.GK32311@shareable.org>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 19:29:41 -0000

Sylvain Hellegouarch wrote:
> Perhaps the reason BEEP has never taken off is because it's so versatile
> that it ultimately creates a bigger mismatch. I'm afraid WS might follow
> the same trend.

The BEEP web page suggests it's used more than you think.

I suspect not so much versatility (HTTP is versatile), as it's too
basic for a lot of applications.  You have to write your own stuff on
top of it, so you might as well go all the way and use TCP or SSL,
which isn't much harder and doesn't require you to learn another library.

Or at the other end of the scale, useing HTTP or something ugly like
SOAP containing a short string wrapped in 1k of pointlessness, because
that's provided by well known libraries and the handy
metadata/extensibility means you don't have to design your own way to
add those things ad hoc.

-- Jamie

From mike@dierken.com  Tue Apr 14 12:37:02 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1A7503A6B4F for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:37:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id q15CwNal9C5Q for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 12:37:01 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.236]) by core3.amsl.com (Postfix) with ESMTP id 276BA3A6A65 for <hybi@ietf.org>; Tue, 14 Apr 2009 12:37:01 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2374842rvb.49 for <hybi@ietf.org>; Tue, 14 Apr 2009 12:38:12 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.141.180.5 with SMTP id h5mr3306183rvp.223.1239737892841; Tue,  14 Apr 2009 12:38:12 -0700 (PDT)
In-Reply-To: <20090414191717.GI32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>
Date: Tue, 14 Apr 2009 12:38:12 -0700
Message-ID: <3f5bf96b0904141238u3b5b32f3rd53b9494bf8cc786@mail.gmail.com>
From: Mike Dierken <mike@dierken.com>
To: Jamie Lokier <jamie@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 19:37:02 -0000

> =A0 - The Javascript in every tab/window coordinates to somehow share
> =A0 =A0 connections. =A0This is extremely difficult to get right even whe=
n
> =A0 =A0 it's allowed by security rules. =A0What happens when the Javascri=
pt
> =A0 =A0 for the first tab/window is closed and it's acting for all the
> =A0 =A0 others? =A0Messy.
Yes - this is something I ran into many years ago and indeed gets
messy. Having the browser (or other user-agent) do this sort of heavy
lifting would be great.

From mjs@apple.com  Tue Apr 14 13:04:24 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AC3903A68DF for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 13:04:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id p2HNQUecEiuS for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 13:04:23 -0700 (PDT)
Received: from mail-out3.apple.com (mail-out3.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id 5ED963A699C for <hybi@ietf.org>; Tue, 14 Apr 2009 13:04:18 -0700 (PDT)
Received: from relay13.apple.com (relay13.apple.com [17.128.113.29]) by mail-out3.apple.com (Postfix) with ESMTP id 9FBFE5C6A646 for <hybi@ietf.org>; Tue, 14 Apr 2009 13:04:46 -0700 (PDT)
Received: from relay13.apple.com (unknown [127.0.0.1]) by relay13.apple.com (Symantec Brightmail Gateway) with ESMTP id 88E0428098 for <hybi@ietf.org>; Tue, 14 Apr 2009 13:04:46 -0700 (PDT)
X-AuditID: 1180711d-a76f2bb000000259-13-49e4ec5ec111
Received: from et.apple.com (et.apple.com [17.151.62.12]) by relay13.apple.com (Apple SCV relay) with ESMTP id 677D128085 for <hybi@ietf.org>; Tue, 14 Apr 2009 13:04:46 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [10.0.1.7] (c-69-181-43-20.hsd1.ca.comcast.net [69.181.43.20]) by et.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI300FL9X3W9W50@et.apple.com> for hybi@ietf.org; Tue, 14 Apr 2009 13:04:46 -0700 (PDT)
Message-id: <01EBA59B-6AB0-4E51-BFF0-20DFC002F916@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Anne van Kesteren <annevk@opera.com>
In-reply-to: <op.usde35il64w2qv@anne-van-kesterens-macbook.local>
Date: Tue, 14 Apr 2009 13:04:39 -0700
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <op.usc9z5al64w2qv@anne-van-kesterens-macbook.local> <694727E2-C898-4119-A265-6351E167A8A8@apple.com> <op.usde35il64w2qv@anne-van-kesterens-macbook.local>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 20:04:24 -0000

On Apr 14, 2009, at 2:53 AM, Anne van Kesteren wrote:

> On Tue, 14 Apr 2009 10:28:31 +0200, Maciej Stachowiak  
> <mjs@apple.com> wrote:
>> In my opinion, optional but well-defined subprotocol identification  
>> is another way, since it enabled building more structured protocols  
>> on top in a clear way.
>
> It seems to me that the path can be used for that, but that is not  
> formal enough for your taste currently as I understand it (it's not  
> well-defined). Any ideas as to what would work better? Or maybe as  
> to how to make the path fit the criteria?


Here's some possibilities that I think are better than leaving it  
completely undefined.

1) Optional HTTP response header for protocol type in the server  
handshake response.

2) Define that part of the path defines the type - this could be (a)  
the first path segment, (b) everything after the last dot (the  
"extension", though that is weird to use for protocol type since for  
server side scripts the extension often just reflects the  
implementation technology) or (c) a semicolon-delimited "parameter"  
which is otherwise not much used in the URI syntax.

3) A dedicated initial pair of messages over the channel with a  
specific format.

I think of these solutions, (1) is the cleanest. There's probably  
other possibilities I didn't think of.

Regards,
Maciej


From jonas@sicking.cc  Tue Apr 14 14:32:25 2009
Return-Path: <jonas@sicking.cc>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 022003A6A46 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 14:32:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3yvtXGf9Y7xy for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 14:32:24 -0700 (PDT)
Received: from mail-gx0-f208.google.com (mail-gx0-f208.google.com [209.85.217.208]) by core3.amsl.com (Postfix) with ESMTP id B5ADC3A68DF for <hybi@ietf.org>; Tue, 14 Apr 2009 14:32:23 -0700 (PDT)
Received: by gxk4 with SMTP id 4so6036648gxk.13 for <hybi@ietf.org>; Tue, 14 Apr 2009 14:33:35 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.150.153.7 with SMTP id a7mr15834225ybe.92.1239744815218; Tue,  14 Apr 2009 14:33:35 -0700 (PDT)
In-Reply-To: <20090414192233.GJ32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <A3699591-148C-4795-967A-6CDE23FE75F0@apple.com>  <Pine.LNX.4.62.0904122230550.10339@hixie.dreamhostps.com>  <71289452-5A19-48F4-9819-7FD9747EE9CA@apple.com> <Pine.LNX.4.62.0904122251070.10339@hixie.dreamhostps.com>  <1D801D05-C7F1-4DDC-8034-FAA458626F53@apple.com> <Pine.LNX.4.62.0904122350370.10339@hixie.dreamhostps.com>  <63df84f0904131829j5bc1b85amf8d24fdd30644f1d@mail.gmail.com>  <20090414192233.GJ32311@shareable.org>
From: Jonas Sicking <jonas@sicking.cc>
Date: Tue, 14 Apr 2009 14:33:20 -0700
Message-ID: <63df84f0904141433j740b886av600a9cd4c61e286e@mail.gmail.com>
To: Jamie Lokier <jamie@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 21:32:25 -0000

One thing that I forgot to include in my previous post was *why* I
thought that it was valuable with some basic semantic information :)

One aspect is the one Jamie is bringing up. We could allow for
protocol negotiation without having to have explicit support for that
in each protocol.

You could also imagine firewalls wanting to block some protocols,
either using white lists or black lists. While I'm personally not a
fan of companies filtering the actions of their employees, I do think
it should be allowed, and more importantly, trying to prevent it is
going to be unsuccessful at best, and harm the protocol at worst.

Security proxies could be made websocket aware and do filtering based
on the protocol that's used over websocket. Imagine that a chat
protocol becomes popular and that that chat protocol has the ability
to transfer files. A security product could scan such files for
viruses and malware if it knows how files are transferred. As I
understand it there are similar products for HTTP traffic. It'd be
much better for everyone involved if corporations did this rather than
block the chat protocol, or websockets, entirely.

I'm not sure that I can't think of a reason to do this, but in theory
a proxy could translate from one protocol to another if it knows what
protocols are being used.

The multiplexing channels issue that has been discussed in this and
other threads is also interesting. However I'd probably say that for
the initial version of the protocol we do not need support for having
UAs or proxies multiplex several connections into one transparently,
as long as we ensure that we can add support for that in a future
version if the need should arise.

/ Jonas

On Tue, Apr 14, 2009 at 12:22 PM, Jamie Lokier <jamie@shareable.org> wrote:
> Jonas Sicking wrote:
>> If the API could be something like
>>
>> socket =3D new WebSocket('ws://example.com/chat', 'DCP');
>> socket.onprotocol =3D chat_handler;
>>
>> for the people that want checking. And
>>
>> socket =3D new WebSocket('ws://example.com/chat');
>> socket.onprotocol =3D iKnowWhoImTalkingToAndDontNeedNoStinkingChecking_h=
andler;
>>
>> for the people that are using their own home-brewed protocol. This
>> could use some magic value as the type (a'la
>> "application/octet-stream") or leave out the protocol indicator
>> entirely.
>
> I agree with all of this.
>
> Think of it as really basic optional type-checking.
>
> Handy when you're connecting to _someone else's_ service URL, which
> might have been supplied by a random user.
>
> It's also handy for protocol negotiation in future, if any protocols
> are standardised it gives you a change to try for the favoured one
> first then fall back, even if the standardised protocols forgot to
> include any negotiation themselves. =A0(Or in other words: they don't hav=
e to).
>
>> I don't have an opinion on if this type should be transferred to the
>> server which would be responsible for the checking, or if the type is
>> transferred to the client which would do the checking, or if its
>> transferred both ways for added goodness.
>
> I say send it to the server, which can check and refuse. =A0Same as the
> URL. =A0That lets the server accept more than one protocol.
>
>> But the point is that the type is transferred at a specified location.
>> And IMHO that specified location should be separated from other fields
>> with other purposes. My understanding is that the path is intended to
>> allow for several websocket services from the same origin, thus I
>> would prefer if the protocol is separate from the path.
>
> Same reasoning here. =A0If the type is optional that's good for the
> folks who don't care too.
>
> Or instead of protocol type, a WebSockets method to add arbitrary HTTP
> headers to the UPGRADE request, similar to XHR?
>
> -- Jamie
>

From gregw@webtide.com  Tue Apr 14 15:36:16 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D9DA33A67D0 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:36:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.549
X-Spam-Level: 
X-Spam-Status: No, score=-2.549 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id a1VJJTi8pLN9 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:36:16 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.186]) by core3.amsl.com (Postfix) with ESMTP id DA1E13A6B34 for <hybi@ietf.org>; Tue, 14 Apr 2009 15:36:15 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so323657tim.25 for <hybi@ietf.org>; Tue, 14 Apr 2009 15:37:26 -0700 (PDT)
Received: by 10.110.42.17 with SMTP id p17mr9857993tip.3.1239748646777; Tue, 14 Apr 2009 15:37:26 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id d4sm1621856tib.28.2009.04.14.15.37.23 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 14 Apr 2009 15:37:25 -0700 (PDT)
Message-ID: <49E5101E.4040108@webtide.com>
Date: Wed, 15 Apr 2009 08:37:18 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com> <20090414153243.GD26621@shareable.org>
In-Reply-To: <20090414153243.GD26621@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 22:36:16 -0000

Jamie Lokier wrote:
> "The problem of firewalls and transparent proxies in the existing
> infrastructure" is an important topic to address and I think it
> deserves analysis _independent_ of API and low-level protocol.

I've added your list of points to the wiki:
  http://trac.tools.ietf.org/bof/trac/wiki/HyBi


>     The Bayeaux people seem to have decided that long-GET is more
>     reliable than slow-streaming-GET, because some HTTP proxies don't
>     forward any of the response until they have received the whole
>     response.  (Is that right?)

That is right.
long-POST for same domain. log-GET for cross domain.

There is actually surprisingly little problem with firewalls
for the streaming solutions.   The main problem is transparent
proxies (and there are enough of these to cause problems - at least
from my personal experience) and browsers that do not execute
javascript until they have all of it OR a buffer overflow.

> It's always possible to tunnel over HTTP, because you can tunnel TCP
> over HTTP, even without CONNECT or UPGRADE), but for some protocols
> it's a poor fit (TCP!), and some fit tunnelling relatively well.  Some
> are even reasonable candidates for firewalls to filter selectively and
> proxies to intercept and cache.

So I'm adding capability negotiation and fallback to HTTP and
candidate requirements on the wiki


cheers



From gregw@webtide.com  Tue Apr 14 15:48:04 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CD48E3A6A15 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:48:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.551
X-Spam-Level: 
X-Spam-Status: No, score=-2.551 tagged_above=-999 required=5 tests=[AWL=0.048,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0w5ALAnhMQTM for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:48:04 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.185]) by core3.amsl.com (Postfix) with ESMTP id D84423A6820 for <hybi@ietf.org>; Tue, 14 Apr 2009 15:48:03 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so324827tim.25 for <hybi@ietf.org>; Tue, 14 Apr 2009 15:49:15 -0700 (PDT)
Received: by 10.110.90.9 with SMTP id n9mr10981515tib.6.1239749354990; Tue, 14 Apr 2009 15:49:14 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id a14sm366338tia.37.2009.04.14.15.49.11 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 14 Apr 2009 15:49:13 -0700 (PDT)
Message-ID: <49E512E2.6000404@webtide.com>
Date: Wed, 15 Apr 2009 08:49:06 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>	 <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>	 <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>	 <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com>	 <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com>	 <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org>	 <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury>	 <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com> <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com>
In-Reply-To: <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 22:48:04 -0000

Mike Dierken wrote:
> I was thinking the applet would expose the 'web socket' API (or
> equivalent) and the app would be built with JavaScript. This would be
> a way to prototype an approach that created a sibling of
> XMLHttpRequest, but meant for receiving unsolicited requests in a page
> - something like XMLHttpServer (but without the XML part).


Having access to TCP does not really help much.

You still will not be able to tunnel through many firewalls
and proxies without pretending to be HTTP.

Also as a client/application API, raw TCP is pretty
difficult to code against.   What does an application do
with 2 bytes of a 6-byte UTF-8 character.

I very much like websockets abstraction of utf-8 datagrams
in the js/dom API.   This is very usable by applications
and/or frameworks (I just don't think the transport should
be skewed to support uncompressed UTF-8 datagrams without meta
data.)

cheers



From gregw@webtide.com  Tue Apr 14 15:57:17 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2E77A3A6A29 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:57:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.252
X-Spam-Level: 
X-Spam-Status: No, score=-2.252 tagged_above=-999 required=5 tests=[AWL=-0.253, BAYES_00=-2.599, J_CHICKENPOX_22=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id E+Fr2XC0Y9mG for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 15:57:11 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.184]) by core3.amsl.com (Postfix) with ESMTP id 1084B3A677D for <hybi@ietf.org>; Tue, 14 Apr 2009 15:57:10 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so325717tim.25 for <hybi@ietf.org>; Tue, 14 Apr 2009 15:58:22 -0700 (PDT)
Received: by 10.110.61.16 with SMTP id j16mr10968870tia.23.1239749901597; Tue, 14 Apr 2009 15:58:21 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id b7sm1469616tic.15.2009.04.14.15.58.19 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 14 Apr 2009 15:58:20 -0700 (PDT)
Message-ID: <49E51506.1020500@webtide.com>
Date: Wed, 15 Apr 2009 08:58:14 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com>	 <20090414153243.GD26621@shareable.org> <79ea848f0904141018j59659f21n82944f3723409991@mail.gmail.com>
In-Reply-To: <79ea848f0904141018j59659f21n82944f3723409991@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 22:57:17 -0000

Mario Balibrera wrote:
> Currently, js.io <http://js.io> works that way with WebSocket. If there 
> is a native browser implementation available (none yet, obviousy)


I find it really strange that there are no browsers that currently
implement Websocket.   It's been around for a while now and I've
frequently asked if there is a browser that supports it - even
a branch of one of the OS browsers.

So far I've drawn a blank... and because the browsers are not
implementing it, I've felt no great need to add it to my
Jetty server.

So there are some browser builders on this list - can any
of them volunteer why there has been no experimentation with
websocket?

Anybody know why the websocket designers have not tested their
proposal with an extension to firefox or similar?


cheers



From mjs@apple.com  Tue Apr 14 16:00:56 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9723A3A6AF6 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 16:00:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -105.999
X-Spam-Level: 
X-Spam-Status: No, score=-105.999 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, J_CHICKENPOX_22=0.6, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pvtceOUAhQr1 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 16:00:55 -0700 (PDT)
Received: from mail-out3.apple.com (mail-out3.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id A8D8B3A677D for <hybi@ietf.org>; Tue, 14 Apr 2009 16:00:34 -0700 (PDT)
Received: from relay10.apple.com (relay10.apple.com [17.128.113.47]) by mail-out3.apple.com (Postfix) with ESMTP id B262B5C74CBA for <hybi@ietf.org>; Tue, 14 Apr 2009 16:01:46 -0700 (PDT)
Received: from relay10.apple.com (unknown [127.0.0.1]) by relay10.apple.com (Symantec Brightmail Gateway) with ESMTP id 8C0362806A for <hybi@ietf.org>; Tue, 14 Apr 2009 16:01:46 -0700 (PDT)
X-AuditID: 1180712f-aa16dbb0000012d3-a9-49e515da577c
Received: from et.apple.com (et.apple.com [17.151.62.12]) by relay10.apple.com (Apple SCV relay) with ESMTP id 744C728057 for <hybi@ietf.org>; Tue, 14 Apr 2009 16:01:46 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from il0301a-dhcp77.apple.com (il0301a-dhcp77.apple.com [17.203.14.205]) by et.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI400FM45AY9W70@et.apple.com> for hybi@ietf.org; Tue, 14 Apr 2009 16:01:46 -0700 (PDT)
Message-id: <30F5867C-5081-422B-BF51-FE84EA6C686E@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Greg Wilkins <gregw@webtide.com>
In-reply-to: <49E51506.1020500@webtide.com>
Date: Tue, 14 Apr 2009 16:01:45 -0700
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com> <20090414153243.GD26621@shareable.org> <79ea848f0904141018j59659f21n82944f3723409991@mail.gmail.com> <49E51506.1020500@webtide.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 23:00:56 -0000

On Apr 14, 2009, at 3:58 PM, Greg Wilkins wrote:

> Mario Balibrera wrote:
>> Currently, js.io <http://js.io> works that way with WebSocket. If  
>> there is a native browser implementation available (none yet,  
>> obviousy)
>
>
> I find it really strange that there are no browsers that currently
> implement Websocket.   It's been around for a while now and I've
> frequently asked if there is a browser that supports it - even
> a branch of one of the OS browsers.
>
> So far I've drawn a blank... and because the browsers are not
> implementing it, I've felt no great need to add it to my
> Jetty server.
>
> So there are some browser builders on this list - can any
> of them volunteer why there has been no experimentation with
> websocket?
>
> Anybody know why the websocket designers have not tested their
> proposal with an extension to firefox or similar?

We're pacing ourselves on adding new features. We can only add so much  
per release cycle, and WebSocket has been significantly redesigned  
fairly recently (from a state we would have considered unacceptable).  
However, I think most browser vendors / developers consider support  
for bidirectional communication a fairly high priority - certainly we  
are very interested in it for Safari and WebKit, and WebSocket seems  
like a good solution in this space.

Regards,
Maciej


From frank.salim@kaazing.com  Tue Apr 14 16:48:42 2009
Return-Path: <frank.salim@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 469993A6DBB for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 16:48:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.376
X-Spam-Level: 
X-Spam-Status: No, score=-1.376 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, J_CHICKENPOX_22=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 46DeEP1xlFlJ for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 16:48:41 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.178]) by core3.amsl.com (Postfix) with ESMTP id 52D523A6B2B for <hybi@ietf.org>; Tue, 14 Apr 2009 16:48:41 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so1387934waf.5 for <hybi@ietf.org>; Tue, 14 Apr 2009 16:49:53 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.114.181.6 with SMTP id d6mr3969771waf.94.1239752993296; Tue,  14 Apr 2009 16:49:53 -0700 (PDT)
In-Reply-To: <49E51506.1020500@webtide.com>
References: <49E3D363.8030803@mozilla.com> <49E3D937.5020009@webtide.com>  <20090414153243.GD26621@shareable.org> <79ea848f0904141018j59659f21n82944f3723409991@mail.gmail.com>  <49E51506.1020500@webtide.com>
From: Frank Salim <frank.salim@kaazing.com>
Date: Tue, 14 Apr 2009 16:49:38 -0700
Message-ID: <b7197f6c0904141649n6e5e9441pd54c1383ba698b9@mail.gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: multipart/alternative; boundary=00163646d987c6540604678c7cbb
Cc: hybi@ietf.org
Subject: Re: [hybi] Working over existing firewalls and proxies
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2009 23:48:42 -0000

--00163646d987c6540604678c7cbb
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

There is a patch for Firefox:
https://bugzilla.mozilla.org/show_bug.cgi?id=472529
-Frank

On Tue, Apr 14, 2009 at 3:58 PM, Greg Wilkins <gregw@webtide.com> wrote:

> Mario Balibrera wrote:
>
>> Currently, js.io <http://js.io> works that way with WebSocket. If there
>> is a native browser implementation available (none yet, obviousy)
>>
>
>
> I find it really strange that there are no browsers that currently
> implement Websocket.   It's been around for a while now and I've
> frequently asked if there is a browser that supports it - even
> a branch of one of the OS browsers.
>
> So far I've drawn a blank... and because the browsers are not
> implementing it, I've felt no great need to add it to my
> Jetty server.
>
> So there are some browser builders on this list - can any
> of them volunteer why there has been no experimentation with
> websocket?
>
> Anybody know why the websocket designers have not tested their
> proposal with an extension to firefox or similar?
>
>
> cheers
>
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--00163646d987c6540604678c7cbb
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

There is a patch for Firefox:<br><a href=3D"https://bugzilla.mozilla.org/sh=
ow_bug.cgi?id=3D472529">https://bugzilla.mozilla.org/show_bug.cgi?id=3D4725=
29</a><br>-Frank<br><br><div class=3D"gmail_quote">On Tue, Apr 14, 2009 at =
3:58 PM, Greg Wilkins <span dir=3D"ltr">&lt;<a href=3D"mailto:gregw@webtide=
.com">gregw@webtide.com</a>&gt;</span> wrote:<br>

<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Mario Balibrera w=
rote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Currently, <a href=3D"http://js.io" target=3D"_blank">js.io</a> &lt;<a href=
=3D"http://js.io" target=3D"_blank">http://js.io</a>&gt; works that way wit=
h WebSocket. If there is a native browser implementation available (none ye=
t, obviousy)<br>


</blockquote>
<br>
<br>
I find it really strange that there are no browsers that currently<br>
implement Websocket. =A0 It&#39;s been around for a while now and I&#39;ve<=
br>
frequently asked if there is a browser that supports it - even<br>
a branch of one of the OS browsers.<br>
<br>
So far I&#39;ve drawn a blank... and because the browsers are not<br>
implementing it, I&#39;ve felt no great need to add it to my<br>
Jetty server.<br>
<br>
So there are some browser builders on this list - can any<br>
of them volunteer why there has been no experimentation with<br>
websocket?<br>
<br>
Anybody know why the websocket designers have not tested their<br>
proposal with an extension to firefox or similar?<br>
<br>
<br>
cheers<div><div></div><div class=3D"h5"><br>
<br>
<br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org" target=3D"_blank">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--00163646d987c6540604678c7cbb--

From mike@dierken.com  Tue Apr 14 17:03:09 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3E0393A6A24 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 17:03:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id I7nGbnTOakgu for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 17:03:08 -0700 (PDT)
Received: from qw-out-2122.google.com (qw-out-2122.google.com [74.125.92.27]) by core3.amsl.com (Postfix) with ESMTP id 4517A3A6ACA for <hybi@ietf.org>; Tue, 14 Apr 2009 17:03:08 -0700 (PDT)
Received: by qw-out-2122.google.com with SMTP id 9so2417754qwb.31 for <hybi@ietf.org>; Tue, 14 Apr 2009 17:04:19 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.220.91.205 with SMTP id o13mr8219432vcm.92.1239753859724; Tue,  14 Apr 2009 17:04:19 -0700 (PDT)
In-Reply-To: <49E512E2.6000404@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com> <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com> <49E512E2.6000404@webtide.com>
Date: Tue, 14 Apr 2009 17:04:19 -0700
Message-ID: <3f5bf96b0904141704h35864b2en18187276a96d388@mail.gmail.com>
From: Mike Dierken <mike@dierken.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 00:03:09 -0000

I was just pointing out that if someone wanted TCP access, they could
do that easily with an very small applet.

I personally wouldn't want just TCP access - I would prefer that the
browser make it easy for a web app author to connect to a 'helper'
service (due to firewalls not allowing inbound connections), register
that the web app hooks into a particular URI namespace (so other
participants in the application have a URI to send messages to), and
for the browser to deliver "Request" objects (and maybe a Response
object as well) to a registered handler function implemented in
JavaScript. The per-app or per-URI dispatching to multiple handlers is
simple enough that this wouldn't need to be standardized up front and
the app developer could do it however they want - as long as the
Request object has the key elements of URI, method, headers and
content.

Is that too much to ask for? ;)




On Tue, Apr 14, 2009 at 3:49 PM, Greg Wilkins <gregw@webtide.com> wrote:
> Mike Dierken wrote:
>>
>> I was thinking the applet would expose the 'web socket' API (or
>> equivalent) and the app would be built with JavaScript. This would be
>> a way to prototype an approach that created a sibling of
>> XMLHttpRequest, but meant for receiving unsolicited requests in a page
>> - something like XMLHttpServer (but without the XML part).
>
>
> Having access to TCP does not really help much.
>
> You still will not be able to tunnel through many firewalls
> and proxies without pretending to be HTTP.
>
> Also as a client/application API, raw TCP is pretty
> difficult to code against. =A0 What does an application do
> with 2 bytes of a 6-byte UTF-8 character.
>
> I very much like websockets abstraction of utf-8 datagrams
> in the js/dom API. =A0 This is very usable by applications
> and/or frameworks (I just don't think the transport should
> be skewed to support uncompressed UTF-8 datagrams without meta
> data.)
>
> cheers
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

From mike@dierken.com  Tue Apr 14 17:26:44 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D465F3A6DF0 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 17:26:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zBv9JefKo6Re for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 17:26:44 -0700 (PDT)
Received: from mail-qy0-f134.google.com (mail-qy0-f134.google.com [209.85.221.134]) by core3.amsl.com (Postfix) with ESMTP id E42C33A692F for <hybi@ietf.org>; Tue, 14 Apr 2009 17:26:43 -0700 (PDT)
Received: by qyk40 with SMTP id 40so1630661qyk.29 for <hybi@ietf.org>; Tue, 14 Apr 2009 17:27:55 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.220.87.1 with SMTP id u1mr8256682vcl.2.1239755273819; Tue, 14  Apr 2009 17:27:53 -0700 (PDT)
In-Reply-To: <F813335479CB4F70A58EC314CF9D362B@mercury>
References: <Acm7GxK/UkDXJP+cQf6D5zE05/aM/A==> <F813335479CB4F70A58EC314CF9D362B@mercury>
Date: Tue, 14 Apr 2009 17:27:53 -0700
Message-ID: <3f5bf96b0904141727s1998edf2n15ec986f175d8f4a@mail.gmail.com>
From: Mike Dierken <mike@dierken.com>
To: hybi@ietf.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Question about charter/mission/summary/etc.
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 00:26:44 -0000

I found the Wiki, so that should provide the background and context I
was looking for.


On Sat, Apr 11, 2009 at 8:02 PM, S. Mike Dierken <mike@dierken.com> wrote:
> I may have joined this mailing list a bit late, and I'm a bit lost trying to
> find a place with a summary of the goal, the players, overview, history and
> whatnot.
>
> So far, what I'm reading sounds just like the what-wg 'event source' stuff
> from a few years back and I was wondering what happened with that, and how
> HyBi relates to that. Can someone share this background and/or point me to a
> page with more info?
>
> Back then, I had written down some thoughts and these may be relevant to the
> current hybi discussion, especially using HTTP message framing for the
> server-to-client messages (though I see it as still client-to-server but the
> server is on the desktop)
>
> http://www.searchalert.net/dierken/eventsource/
> http://korrespondence.blogspot.com/2005/06/whatwg-and-event-source-design.html

From frank.salim@kaazing.com  Tue Apr 14 19:32:04 2009
Return-Path: <frank.salim@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7F35A3A68A8 for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 19:32:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.677
X-Spam-Level: 
X-Spam-Status: No, score=-1.677 tagged_above=-999 required=5 tests=[AWL=0.301,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ha-2cqzLKOxS for <hybi@core3.amsl.com>; Tue, 14 Apr 2009 19:32:03 -0700 (PDT)
Received: from mail-gx0-f208.google.com (mail-gx0-f208.google.com [209.85.217.208]) by core3.amsl.com (Postfix) with ESMTP id ACA493A6859 for <hybi@ietf.org>; Tue, 14 Apr 2009 19:32:03 -0700 (PDT)
Received: by gxk4 with SMTP id 4so6305351gxk.13 for <hybi@ietf.org>; Tue, 14 Apr 2009 19:33:15 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.68.12 with SMTP id q12mr7171693aga.50.1239762795293; Tue,  14 Apr 2009 19:33:15 -0700 (PDT)
In-Reply-To: <20090414191717.GI32311@shareable.org>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>  <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>  <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>
From: Frank Salim <frank.salim@kaazing.com>
Date: Tue, 14 Apr 2009 19:33:00 -0700
Message-ID: <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>
To: Jamie Lokier <jamie@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 02:32:04 -0000

I think that connection sharing and coordinating amongst tabs is well
served by workers, as Ian pointed out. I do not think it should affect
the protocol design.

> What happens when the Javascript
> for the first tab/window is closed and it's acting for all the
> others?  Messy.

I believe this is addressed by workers.

> Unfortunately it does require a lot of code on top of WebSocket to do
> the kind of updates I'm doing.  (Ironic as the point is network
> efficiency and minimal updates).

Code is not bytes on the wire! Addressing your use case in the
protocol makes everyone pay for the features you need, and the cost is
bandwidth instead of code. It is better to address them at a higher
layer with another protocol (AMQP? BEEP?). Likewise, the multiple
widget case is best served by an application protocol with
publish/subscribe semantics.

> From the above, my guess is WebSockets look compelling for the "3
> widgets on a page need updates" case, and for the "I like writing my
> own protocols and having full control" case.
>
> But they don't look compelling for "I want to write an attractive web
> application with lots of updating documents and tiny widgets without
> using a pre-existing application framework or large bit of
> Javascript".

I think that this is correct. However, I think that there is nothing
wrong with using JavaScript to implement an application protocol. This
will be reusable code that is cached in CDNs. It will be much like
using a library that implements a protocol in C.

-Frank

From sh@defuze.org  Wed Apr 15 00:16:05 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 627D93A6AD3 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 00:16:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.232
X-Spam-Level: 
X-Spam-Status: No, score=-2.232 tagged_above=-999 required=5 tests=[AWL=0.033,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HC9cFVOQbQwR for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 00:16:04 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 8C9E93A6828 for <hybi@ietf.org>; Wed, 15 Apr 2009 00:15:56 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3F7H8Bh026752; Wed, 15 Apr 2009 02:17:08 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n3F7H7pP026750; Wed, 15 Apr 2009 09:17:07 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Wed, 15 Apr 2009 09:17:07 +0200 (CEST)
Message-ID: <53386.193.253.216.132.1239779827.squirrel@mail1.webfaction.com>
In-Reply-To: <20090414193050.GK32311@shareable.org>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com> <20090414193050.GK32311@shareable.org>
Date: Wed, 15 Apr 2009 09:17:07 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: "Jamie Lokier" <jamie@shareable.org>
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 07:16:05 -0000

> Sylvain Hellegouarch wrote:
>> Perhaps the reason BEEP has never taken off is because it's so versatile
>> that it ultimately creates a bigger mismatch. I'm afraid WS might follow
>> the same trend.
>
> The BEEP web page suggests it's used more than you think.

Well to be fair, those applications don't look so much as being used on
the web itself but rather internally. It's quite easy to adapt to a
protocol when you don't have the constraints of a global network you don't
control.

>
> I suspect not so much versatility (HTTP is versatile), as it's too
> basic for a lot of applications.  You have to write your own stuff on
> top of it, so you might as well go all the way and use TCP or SSL,
> which isn't much harder and doesn't require you to learn another library.

Much agreed and that's the point I was tyring to make :)

>
> Or at the other end of the scale, useing HTTP or something ugly like
> SOAP containing a short string wrapped in 1k of pointlessness, because
> that's provided by well known libraries and the handy
> metadata/extensibility means you don't have to design your own way to
> add those things ad hoc.

You don't say :)

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From sh@defuze.org  Wed Apr 15 00:33:51 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E67363A6B37 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 00:33:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.235
X-Spam-Level: 
X-Spam-Status: No, score=-2.235 tagged_above=-999 required=5 tests=[AWL=0.030,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id OgDjgF8yUepF for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 00:33:51 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 0504C3A67EE for <hybi@ietf.org>; Wed, 15 Apr 2009 00:33:50 -0700 (PDT)
Received: from mail1.webfaction.com (localhost.localdomain [127.0.0.1]) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3F7Z2od031164; Wed, 15 Apr 2009 02:35:02 -0500
Received: (from apache@localhost) by mail1.webfaction.com (8.12.11.20060308/8.12.11/Submit) id n3F7Z2Ts031162; Wed, 15 Apr 2009 09:35:02 +0200
X-Authentication-Warning: mail1.webfaction.com: apache set sender to sh@defuze.org using -f
Received: from 193.253.216.132 (SquirrelMail authenticated user platom_sylvain) by mail1.webfaction.com with HTTP; Wed, 15 Apr 2009 09:35:02 +0200 (CEST)
Message-ID: <54156.193.253.216.132.1239780902.squirrel@mail1.webfaction.com>
In-Reply-To: <49E512E2.6000404@webtide.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com> <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com> <49E512E2.6000404@webtide.com>
Date: Wed, 15 Apr 2009 09:35:02 +0200 (CEST)
From: "Sylvain Hellegouarch" <sh@defuze.org>
To: "Greg Wilkins" <gregw@webtide.com>
User-Agent: SquirrelMail/1.4.8-6.el3
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
Importance: Normal
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: sh@defuze.org
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 07:33:52 -0000

> Mike Dierken wrote:
>> I was thinking the applet would expose the 'web socket' API (or
>> equivalent) and the app would be built with JavaScript. This would be
>> a way to prototype an approach that created a sibling of
>> XMLHttpRequest, but meant for receiving unsolicited requests in a page
>> - something like XMLHttpServer (but without the XML part).
>
>
> Having access to TCP does not really help much.
>
> You still will not be able to tunnel through many firewalls
> and proxies without pretending to be HTTP.

But that's exactly the point, isn't it? Instead of reinventing the wheel
to go through firewalls and proxies, you reuse what's been defined,
refined and tested over decades of networking.

Corporations will always want at some point to control their traffic, they
have already administrators that know how to handle the basic TCP routing.
Here we are trying to defeat that but ultimately software will be upgraded
to filter WS anyway if required. You just add a layer of complexity to the
game. I fail to understand the value in that. Why not educating players in
the game to play fair? Port 80 is spoiled more and more and that means
it's more and more complex to perform proper security without basically
shutting down almost anything going through.

>
> Also as a client/application API, raw TCP is pretty
> difficult to code against.   What does an application do
> with 2 bytes of a 6-byte UTF-8 character.

It doesn't seem to have blocked the software industry so far. Libraries
handle those issues and offer higher level of abstraction for any given
protocol. JavaScript is short in those so far precisely due to the
constraints of not having access to raw TCP sockets I believe.

Why is the browser so different from anything else?

>
> I very much like websockets abstraction of utf-8 datagrams
> in the js/dom API.   This is very usable by applications
> and/or frameworks (I just don't think the transport should
> be skewed to support uncompressed UTF-8 datagrams without meta
> data.)

>From my perspective it's just one more layer of complexity.

- Sylvain
-- 
Sylvain Hellegouarch
http://www.defuze.org

From gregw@webtide.com  Wed Apr 15 02:35:02 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C9CF13A6A67 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 02:35:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.544
X-Spam-Level: 
X-Spam-Status: No, score=-2.544 tagged_above=-999 required=5 tests=[AWL=0.055,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uvD8niLeUEWe for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 02:35:02 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.227]) by core3.amsl.com (Postfix) with ESMTP id 19B333A693C for <hybi@ietf.org>; Wed, 15 Apr 2009 02:35:01 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2639078rvb.49 for <hybi@ietf.org>; Wed, 15 Apr 2009 02:36:14 -0700 (PDT)
Received: by 10.140.134.15 with SMTP id h15mr268632rvd.190.1239788174063; Wed, 15 Apr 2009 02:36:14 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id c20sm17254576rvf.0.2009.04.15.02.36.10 (version=TLSv1/SSLv3 cipher=RC4-MD5); Wed, 15 Apr 2009 02:36:12 -0700 (PDT)
Message-ID: <49E5AA86.9040100@webtide.com>
Date: Wed, 15 Apr 2009 19:36:06 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Frank Salim <frank.salim@kaazing.com>, hybi@ietf.org
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>
In-Reply-To: <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 09:35:02 -0000

Frank Salim wrote:
> I think that connection sharing and coordinating amongst tabs is well
> served by workers, as Ian pointed out. I do not think it should affect
> the protocol design.

Frank,

while workers might make such sharing technically possible, I think there
is still an issue since the websocket API is both an application API
and a protocol API.    Developers who just want UTF-8 datagrams and
don't care about the transport will use the websocket API directly and
will thus inadvertently make a decision about the creating a connection
to the server!.

Even XHR abstracts the requests made away from the connections used.

I think in this regard websocket API is a step in the wrong direction
as it makes the developer directly request a connection.   Instead
they should be directly requesting a channel and it should be up to
the browser to decide if that is a new connection or a shared multiplexed
connection.

At the very least, it should provide an SPI so that frameworks can
get underneath the users trying to consume datagrams and manage
the connections for them.

regards





From mario.balibrera@gmail.com  Wed Apr 15 09:41:25 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D6E613A6936 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 09:41:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.198
X-Spam-Level: 
X-Spam-Status: No, score=-2.198 tagged_above=-999 required=5 tests=[AWL=-0.200, BAYES_00=-2.599, HTML_MESSAGE=0.001, J_CHICKENPOX_64=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5gS32NpXqdSZ for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 09:41:24 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.230]) by core3.amsl.com (Postfix) with ESMTP id D818A3A6BBB for <hybi@ietf.org>; Wed, 15 Apr 2009 09:41:24 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so2788812rvb.49 for <hybi@ietf.org>; Wed, 15 Apr 2009 09:42:37 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=ys+DyWG0PcqHg59fyfTovBRpki3bPYQUDonArSsiDEo=; b=pfjECJtsTcnzzWpuVfwMo44GlPxwbxuLI5SLTj/AelRTPg8wQYHJn6t5hh+C6YYoMx eXYbypTAnKnkAZlp1Rm0GQJceKHfkPL8aRi4t0QYNd573IQEdk+7YFu4GjGRHf2bMsNt mGEsR5lEquQrUgDGTv/ctcC9UqwIEVSZ4GgXg=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=bFo7QMP3zWS3MvqQkniLpLGt6IrCRgoXcJ7dY0EQ2eBYbPQY850DxDecXvxPjuNWew aB1zfK6OUiIUu00ZOJYoZwoSyXV/KoWkCl3ulY46HFQ3AuNAyk3VBMCUEKLGG46jLO3l OmAKPDRUWcaGo3pzTITQHCnercSsWYK3dEoBo=
MIME-Version: 1.0
Received: by 10.114.200.2 with SMTP id x2mr153484waf.83.1239813757338; Wed, 15  Apr 2009 09:42:37 -0700 (PDT)
In-Reply-To: <49E5AA86.9040100@webtide.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com> <49E5AA86.9040100@webtide.com>
Date: Wed, 15 Apr 2009 09:42:37 -0700
Message-ID: <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: multipart/alternative; boundary=00163645800098036f04679aa202
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 16:41:25 -0000

--00163645800098036f04679aa202
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Greg Wilkins wrote:
I think there
is still an issue since the websocket API is both an application API
and a protocol API.

Frank's point is that WebSocket is _not_ an application API. Despite the
migration of this thread (from "Apples and Orangutans" to "Addressing
Minimal Events"), we're still discussing every proposal that comes up as if
it fit in the same layer. Do socket.send and socket.recv constitute an
application API? I mean, come on guys, why can't my socket speak XMPP and
support transactions?

GW:
Developers who just want UTF-8 datagrams and
don't care about the transport will use the websocket API directly and
will thus inadvertently make a decision about the creating a connection
to the server!.

In 99% of cases, this is not a problem. Any server supporting a streaming
connection -oriented application will not be affected if a user has 2 or 10
or 20 connections open instead of 1. And you know what? If someone really
needs his 1000 connections multiplexed, he can implement that on top of
WebSocket, because it is a _transport_ protocol.

Mario Balibrera

On Wed, Apr 15, 2009 at 2:36 AM, Greg Wilkins <gregw@webtide.com> wrote:

> Frank Salim wrote:
>
>> I think that connection sharing and coordinating amongst tabs is well
>> served by workers, as Ian pointed out. I do not think it should affect
>> the protocol design.
>>
>
> Frank,
>
> while workers might make such sharing technically possible, I think there
> is still an issue since the websocket API is both an application API
> and a protocol API.    Developers who just want UTF-8 datagrams and
> don't care about the transport will use the websocket API directly and
> will thus inadvertently make a decision about the creating a connection
> to the server!.
>
> Even XHR abstracts the requests made away from the connections used.
>
> I think in this regard websocket API is a step in the wrong direction
> as it makes the developer directly request a connection.   Instead
> they should be directly requesting a channel and it should be up to
> the browser to decide if that is a new connection or a shared multiplexed
> connection.
>
> At the very least, it should provide an SPI so that frameworks can
> get underneath the users trying to consume datagrams and manage
> the connections for them.
>
> regards
>
>
>
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--00163645800098036f04679aa202
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div class=3D"im">Greg Wilkins wrote:<br>I think there<br>
is still an issue since the websocket API is both an application API<br>
and a protocol API.<br><br></div>Frank&#39;s
point is that WebSocket is _not_ an application API. Despite the
migration of this thread (from &quot;Apples and Orangutans&quot; to &quot;A=
ddressing
Minimal Events&quot;), we&#39;re still discussing every proposal that comes=
 up
as if it fit in the same layer. Do socket.send and socket.recv
constitute an application API? I mean, come on guys, why can&#39;t my
socket speak XMPP and support transactions?<br>
<br>GW:<div class=3D"im"><br>Developers who just want UTF-8 datagrams and<b=
r>
don&#39;t care about the transport will use the websocket API directly and<=
br>
will thus inadvertently make a decision about the creating a connection<br>
to the server!.<br>
<br></div>In 99% of cases, this is not a problem. Any server supporting
a streaming connection -oriented application will not be affected if a
user has 2 or 10 or 20 connections open instead of 1. And you know
what? If someone really needs his 1000 connections multiplexed, he can
implement that on top of WebSocket, because it is a _transport_
protocol.<br><font color=3D"#888888">
<br>Mario Balibrera</font><br><br><div class=3D"gmail_quote">On Wed, Apr 15=
, 2009 at 2:36 AM, Greg Wilkins <span dir=3D"ltr">&lt;<a href=3D"mailto:gre=
gw@webtide.com">gregw@webtide.com</a>&gt;</span> wrote:<br><blockquote clas=
s=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margi=
n: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">Frank Salim wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I think that connection sharing and coordinating amongst tabs is well<br>
served by workers, as Ian pointed out. I do not think it should affect<br>
the protocol design.<br>
</blockquote>
<br></div>
Frank,<br>
<br>
while workers might make such sharing technically possible, I think there<b=
r>
is still an issue since the websocket API is both an application API<br>
and a protocol API. =A0 =A0Developers who just want UTF-8 datagrams and<br>
don&#39;t care about the transport will use the websocket API directly and<=
br>
will thus inadvertently make a decision about the creating a connection<br>
to the server!.<br>
<br>
Even XHR abstracts the requests made away from the connections used.<br>
<br>
I think in this regard websocket API is a step in the wrong direction<br>
as it makes the developer directly request a connection. =A0 Instead<br>
they should be directly requesting a channel and it should be up to<br>
the browser to decide if that is a new connection or a shared multiplexed<b=
r>
connection.<br>
<br>
At the very least, it should provide an SPI so that frameworks can<br>
get underneath the users trying to consume datagrams and manage<br>
the connections for them.<br>
<br>
regards<div><div></div><div class=3D"h5"><br>
<br>
<br>
<br>
<br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org" target=3D"_blank">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--00163645800098036f04679aa202--

From gregw@webtide.com  Wed Apr 15 16:28:29 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A20A73A6846 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 16:28:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.546
X-Spam-Level: 
X-Spam-Status: No, score=-2.546 tagged_above=-999 required=5 tests=[AWL=0.053,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eSuIxWsucS3W for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 16:28:28 -0700 (PDT)
Received: from ti-out-0910.google.com (ti-out-0910.google.com [209.85.142.189]) by core3.amsl.com (Postfix) with ESMTP id AD9C33A6AC6 for <hybi@ietf.org>; Wed, 15 Apr 2009 16:28:28 -0700 (PDT)
Received: by ti-out-0910.google.com with SMTP id 11so40555tim.25 for <hybi@ietf.org>; Wed, 15 Apr 2009 16:29:40 -0700 (PDT)
Received: by 10.110.69.5 with SMTP id r5mr730774tia.19.1239838180417; Wed, 15 Apr 2009 16:29:40 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id a14sm1146245tia.21.2009.04.15.16.29.37 (version=TLSv1/SSLv3 cipher=RC4-MD5); Wed, 15 Apr 2009 16:29:39 -0700 (PDT)
Message-ID: <49E66DDD.8030201@webtide.com>
Date: Thu, 16 Apr 2009 09:29:33 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Mario Balibrera <mario.balibrera@gmail.com>
References: <49DEF171.4080506@mozilla.com>	 <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>	 <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>	 <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>	 <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>	 <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>	 <49E5AA86.9040100@webtide.com> <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com>
In-Reply-To: <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2009 23:28:29 -0000

Mario Balibrera wrote:
> In 99% of cases, this is not a problem. Any server supporting a 
> streaming connection -oriented application will not be affected if a 
> user has 2 or 10 or 20 connections open instead of 1. And you know what? 
> If someone really needs his 1000 connections multiplexed, he can 
> implement that on top of WebSocket, because it is a _transport_ protocol.


I have to disagree on both counts.

It is my experience in deploying many Comet Ajax Push applications that
scalability is limited either by message rate (for apps like stock tickers)
or by number of connections (apps like auctions, IM and/or chat).
Connections are a big problem and will only become more so if creating them
become easier.

Websocket has a transport protocol, but it also has an API designed
to be used by javascript developers that directly relates to that
transport.   Developers will directly use that API from applications
and frameworks.    If sockets are to be shared, then it will have to
be as a result of those applications and frameworks voluntarily and
explicitly taking part in some multi-plexing scheme - just like in
the good old days when multi-tasking was done by applications
voluntarily giving back the CPU!

Connections are a real problem and I believe that the Websocket
API will encourage the usage of more sockets.

cheers







From mnot@mnot.net  Wed Apr 15 17:18:09 2009
Return-Path: <mnot@mnot.net>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 688103A6BE2 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:18:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.242
X-Spam-Level: 
X-Spam-Status: No, score=-5.242 tagged_above=-999 required=5 tests=[AWL=-2.643, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7vHJHP+xThl1 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:18:08 -0700 (PDT)
Received: from mxout-08.mxes.net (mxout-08.mxes.net [216.86.168.183]) by core3.amsl.com (Postfix) with ESMTP id 64B803A67AD for <hybi@ietf.org>; Wed, 15 Apr 2009 17:18:08 -0700 (PDT)
Received: from [192.168.1.1] (unknown [209.131.62.115]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by smtp.mxes.net (Postfix) with ESMTPSA id 61FE8D05A8; Wed, 15 Apr 2009 20:19:17 -0400 (EDT)
Message-Id: <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net>
From: Mark Nottingham <mnot@mnot.net>
To: Jamie Lokier <jamie@shareable.org>
In-Reply-To: <20090414191717.GI32311@shareable.org>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Thu, 16 Apr 2009 10:19:15 +1000
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>
X-Mailer: Apple Mail (2.930.3)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 00:18:09 -0000

What I find interesting is that these are all API / implementation /  
policy issues, not wire protocol issues.

I.e., based on what I've read so far, "minimal, event" can be achieved  
by HTTP on the wire using Comet (perhaps with best practices /  
extensions to augment it); it's mostly the browser-side model for  
using HTTP that needs work here.

I know that Ian (and others?) also want to allow someone more advanced  
to create a bespoke protocol that goes beyond "minimal events", but I  
think that discussion should be separated; i.e., don't assume that  
addressing minimal events requires a new wire protocol.



On 15/04/2009, at 5:17 AM, Jamie Lokier wrote:

> Greg Wilkins wrote:
>> Anyway, my point is, that when considering multiplexing
>> events onto the same transport, we could consider the
>> multi frame/tab/window issues as well.
>
> Interesting point.
>
> This is another good reason to do multiplexing and connection sharing
> in the browser itself, instead of in Javascript.
>
> I personally have about 200 tabs open in 20 windows on 2 desktops,
> most of the time.  That's perhaps unusual.  Mostly to different sites
> - things I'm reading and want to go back to - but 10 tabs open to a
> single site isn't uncommon.
>
> If you have 10 tabs open to a single site, then you _must_ have the
> flow of events (including large events like images) being independent
> to each tab/window.
>
> That's because each tab/window is expected to behave independently of
> the others, even if the others get stuck, or are uploading/downloading
> some big file.
>
> You can do that by either:
>
>   - The browser opens separate connections for each window/tab.
>
>   - The browser shares connections for each window/tab and handles
>     multiplexing automatically.  This requires the low-level
>     protocols to support multiplexing.
>
>   - The Javascript in every tab/window coordinates to somehow share
>     connections.  This is extremely difficult to get right even when
>     it's allowed by security rules.  What happens when the Javascript
>     for the first tab/window is closed and it's acting for all the
>     others?  Messy.
>
> The same applies for IFRAMEs to different sites.  However, multiple
> tab/window is surely even more popular.
>
> (The same also applies to different parts of a page being authored by
> different people using different Javascript frameworks, so not using
> Javascript-level shared dispatch protocol.  But that's been covered
> before.)
>
> Right now, the browser handles sharing HTTP connections automatically
> and mostly transparently.  This is good.
>
> I think it should handle sharing of whatever else we're proposing too,
> for all the advantages which come from that instead of forcing rather
> complex Javascript implementations to do it badly - or fail to do it
> reliably at all in many cases.  (Most Javascript authors are not  
> experts).
>
> The only realistically usable alternative is lots of separate
> connections, and apart from anything else, that's a waste of
> resources.  We're not talking 3 connections here, we're talking 100
> idle connections for a single busy browser (10 tabs, 100 widgets from
> 10 different authors per tab).  Which is absurd, and won't scale over
> the internet's NAT routers, let alone the actual computers and
> connection setup time :-)
>
> -- Jamie
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi


--
Mark Nottingham     http://www.mnot.net/


From annevk@opera.com  Wed Apr 15 17:26:51 2009
Return-Path: <annevk@opera.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D93B93A67AD for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:26:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -8.599
X-Spam-Level: 
X-Spam-Status: No, score=-8.599 tagged_above=-999 required=5 tests=[AWL=-2.000, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QbGpMP4LfTYs for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:26:51 -0700 (PDT)
Received: from smtp.opera.com (sam.opera.com [213.236.208.81]) by core3.amsl.com (Postfix) with ESMTP id B25DD3A67B5 for <hybi@ietf.org>; Wed, 15 Apr 2009 17:26:50 -0700 (PDT)
Received: from annevk-t60.oslo.opera.com (53568A94.cable.casema.nl [83.86.138.148]) (authenticated bits=0) by smtp.opera.com (8.13.4/8.13.4/Debian-3sarge3) with ESMTP id n3G0Rw2M030813 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 16 Apr 2009 00:28:00 GMT
Date: Thu, 16 Apr 2009 02:27:56 +0200
To: "Mark Nottingham" <mnot@mnot.net>, "Jamie Lokier" <jamie@shareable.org>
From: "Anne van Kesteren" <annevk@opera.com>
Organization: Opera Software ASA
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net>
Content-Transfer-Encoding: 7bit
Message-ID: <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com>
In-Reply-To: <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net>
User-Agent: Opera Mail/9.62 (Linux)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 00:26:51 -0000

On Thu, 16 Apr 2009 02:19:15 +0200, Mark Nottingham <mnot@mnot.net> wrote:
> I.e., based on what I've read so far, "minimal, event" can be achieved  
> by HTTP on the wire using Comet (perhaps with best practices /  
> extensions to augment it); it's mostly the browser-side model for using  
> HTTP that needs work here.

That's just polling data over HTTP. There's an API specification for that  
already:

   http://dev.w3.org/html5/eventsource/


> I know that Ian (and others?) also want to allow someone more advanced  
> to create a bespoke protocol that goes beyond "minimal events", but I  
> think that discussion should be separated; i.e., don't assume that  
> addressing minimal events requires a new wire protocol.

I might be missing something but I thought this was about bidirectional  
minimal events.


-- 
Anne van Kesteren
http://annevankesteren.nl/

From mnot@mnot.net  Wed Apr 15 17:31:49 2009
Return-Path: <mnot@mnot.net>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 375333A6A2C for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:31:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.2
X-Spam-Level: 
X-Spam-Status: No, score=-5.2 tagged_above=-999 required=5 tests=[AWL=-2.601,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uz8vdODRbnke for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:31:48 -0700 (PDT)
Received: from mxout-08.mxes.net (mxout-08.mxes.net [216.86.168.183]) by core3.amsl.com (Postfix) with ESMTP id 6D0353A691E for <hybi@ietf.org>; Wed, 15 Apr 2009 17:31:48 -0700 (PDT)
Received: from [192.168.1.1] (unknown [209.131.62.115]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by smtp.mxes.net (Postfix) with ESMTPSA id AA0E1D0562; Wed, 15 Apr 2009 20:32:58 -0400 (EDT)
Message-Id: <FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net>
From: Mark Nottingham <mnot@mnot.net>
To: Anne van Kesteren <annevk@opera.com>
In-Reply-To: <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Thu, 16 Apr 2009 10:32:55 +1000
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net> <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com>
X-Mailer: Apple Mail (2.930.3)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 00:31:49 -0000

On 16/04/2009, at 10:27 AM, Anne van Kesteren wrote:

> On Thu, 16 Apr 2009 02:19:15 +0200, Mark Nottingham <mnot@mnot.net>  
> wrote:
>> I.e., based on what I've read so far, "minimal, event" can be  
>> achieved by HTTP on the wire using Comet (perhaps with best  
>> practices / extensions to augment it); it's mostly the browser-side  
>> model for using HTTP that needs work here.
>
> That's just polling data over HTTP. There's an API specification for  
> that already:
>
>  http://dev.w3.org/html5/eventsource/

Thanks, wasn't aware of that one.


>> I know that Ian (and others?) also want to allow someone more  
>> advanced to create a bespoke protocol that goes beyond "minimal  
>> events", but I think that discussion should be separated; i.e.,  
>> don't assume that addressing minimal events requires a new wire  
>> protocol.
>
> I might be missing something but I thought this was about  
> bidirectional minimal events.

Does anything need to be said -- at a protocol level -- about sending  
events in the other direction over HTTP?


--
Mark Nottingham     http://www.mnot.net/


From frank.salim@kaazing.com  Wed Apr 15 17:52:17 2009
Return-Path: <frank.salim@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B8E8B3A6BDA for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:52:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.827
X-Spam-Level: 
X-Spam-Status: No, score=-1.827 tagged_above=-999 required=5 tests=[AWL=0.150,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yxiuGCSLPnms for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 17:52:17 -0700 (PDT)
Received: from mail-gx0-f160.google.com (mail-gx0-f160.google.com [209.85.217.160]) by core3.amsl.com (Postfix) with ESMTP id 1163E3A69E7 for <hybi@ietf.org>; Wed, 15 Apr 2009 17:52:16 -0700 (PDT)
Received: by gxk4 with SMTP id 4so462116gxk.13 for <hybi@ietf.org>; Wed, 15 Apr 2009 17:53:29 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.75.13 with SMTP id x13mr946845aga.71.1239843209225; Wed, 15  Apr 2009 17:53:29 -0700 (PDT)
In-Reply-To: <49E66DDD.8030201@webtide.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>  <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>  <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>  <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>  <49E5AA86.9040100@webtide.com> <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com>  <49E66DDD.8030201@webtide.com>
From: Frank Salim <frank.salim@kaazing.com>
Date: Wed, 15 Apr 2009 17:53:14 -0700
Message-ID: <b7197f6c0904151753ne354d8cl62d0e66501e3b80f@mail.gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 00:52:17 -0000

> Connections are a real problem and I believe that the Websocket
> API will encourage the usage of more sockets.

WebSockets are not subject to the browsers' restrictions on the number
of HTTP connections. I don't think this is a problem.

-Frank

From markl@lindenlab.com  Wed Apr 15 19:01:16 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2D3EA3A6961 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:01:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.999
X-Spam-Level: 
X-Spam-Status: No, score=-2.999 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, J_CHICKENPOX_13=0.6, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ntbzuJA77ZR8 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:01:15 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 548ED3A67FC for <hybi@ietf.org>; Wed, 15 Apr 2009 19:01:15 -0700 (PDT)
Received: from [192.168.1.192] (m208-18.dsl.rawbw.com [198.144.208.18]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 842D21414002 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:02:27 -0700 (PDT)
Message-Id: <146BB157-60BF-4942-869B-18F7BA7474F7@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <54156.193.253.216.132.1239780902.squirrel@mail1.webfaction.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
X-Priority: 3 (Normal)
Date: Wed, 15 Apr 2009 18:35:10 -0700
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <49E3D66C.5060002@webtide.com> <49E3D731.30305@mozilla.com> <79ea848f0904131727w5d4bc0d8kc9914d26080a01fc@mail.gmail.com> <49E3DB47.5060801@webtide.com> <49E428DD.3070803@defuze.org> <086D9C6DD2C9407D8D8CE0601FF0FD4A@mercury> <50044.193.253.216.132.1239692803.squirrel@mail1.webfaction.com> <3f5bf96b0904141136j6bc6e179w34af8ebb449b8d70@mail.gmail.com> <49E512E2.6000404@webtide.com> <54156.193.253.216.132.1239780902.squirrel@mail1.webfaction.com>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 02:01:16 -0000

Wow - the only traffic jam seems to be in my in-box! My this mailing  
list has gotten popular!

I tend to agree with the arguments against a minimal protocol such as  
WebSockets:
+1: there's nothing wrong using existing libraries for things like  
HTTP, XML, UTF-8, etc...
+1: inexperienced programmers are going to code to an API like CGI,  
not write the whole server
+1: you can make a good, easy to use API even for a rich protocol: to  
wit: CGI
+1: if the protocol is too minimal, the commonly used stuff will be  
implemented up the stack, which means:
	a) it will be non-interoperable
and	b)the infrastructure will not be able to interact with it

Other things I noticed:
- HTTP over SCTP *does* seem interesting. It might be nice to separate  
out the channeling and asynchrony from the bi-directionality. Now if  
we could do that, and have a fall back for today (not much SCTP in  
stock OSes...)... that is an interesting direction...
- Tunneling through existing firewall and proxy constraints seems to  
keep rearing its head. This feels to me like a separate concern: We  
should design what we think is useful, expect that in a few years  
firewalls and proxies will adapt, *and* design a pure HTTP fallback  
that will allow us to "get on with it" today.

	- Mark

Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From markl@lindenlab.com  Wed Apr 15 19:01:16 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8E2AB3A67FC for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:01:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.299
X-Spam-Level: 
X-Spam-Status: No, score=-3.299 tagged_above=-999 required=5 tests=[AWL=0.299,  BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id C9k4c8Be29Zz for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:01:15 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id C2BC73A6934 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:01:15 -0700 (PDT)
Received: from [192.168.1.192] (m208-18.dsl.rawbw.com [198.144.208.18]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 890A31414002 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:02:28 -0700 (PDT)
Message-Id: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
Content-Type: multipart/alternative; boundary=Apple-Mail-20--453798811
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Wed, 15 Apr 2009 18:46:46 -0700
X-Mailer: Apple Mail (2.930.3)
Subject: [hybi] pointers, please?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 02:01:16 -0000

--Apple-Mail-20--453798811
Content-Type: text/plain;
	charset=US-ASCII;
	format=flowed;
	delsp=yes
Content-Transfer-Encoding: 7bit

> The same-origin security policy.
Can someone point me at a text on this?

Also, while I understand the difficulties that the two connection  
limit imposes on applications doing Comet/AJAX style operations, I  
couldn't seem to find a discussion of why that limit was added to HTTP?

I don't want to incite re-opening of these issues here - I'm just  
looking for pointers....   at least for now :-)

	- Mark


Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






--Apple-Mail-20--453798811
Content-Type: text/html;
	charset=US-ASCII
Content-Transfer-Encoding: quoted-printable

<html><body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; =
-webkit-line-break: after-white-space; "><blockquote type=3D"cite">The =
same-origin security policy.<br></blockquote><div><font =
class=3D"Apple-style-span" color=3D"#144FAE">Can someone point me at a =
text on this?</font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE"><br></font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE">Also, while I understand the difficulties that the two =
connection limit imposes on applications doing Comet/AJAX style =
operations, I couldn't seem to find a discussion of why that limit was =
added to HTTP?</font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE"><br></font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE">I don't want to incite re-opening of these issues here =
- I'm just looking for pointers.... &nbsp; at least for now =
:-)</font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE"><br></font></div><div><font class=3D"Apple-style-span" =
color=3D"#144FAE"><span class=3D"Apple-tab-span" =
style=3D"white-space:pre">	</span>- Mark<br></font></div><div><font =
class=3D"Apple-style-span" color=3D"#144FAE"><br></font></div><div><font =
class=3D"Apple-style-span" color=3D"#144FAE"><br></font></div><div =
apple-content-edited=3D"true"> <span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; color: rgb(0, 0, 0); font-family: =
Helvetica; font-size: 12px; font-style: normal; font-variant: normal; =
font-weight: normal; letter-spacing: normal; line-height: normal; =
orphans: 2; text-align: auto; text-indent: 0px; text-transform: none; =
white-space: normal; widows: 2; word-spacing: 0px; =
-webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: =
0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: =
auto; -webkit-text-stroke-width: 0; "><div style=3D"word-wrap: =
break-word; -webkit-nbsp-mode: space; -webkit-line-break: =
after-white-space; "><span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; color: rgb(0, 0, 0); font-family: =
Helvetica; font-size: 12px; font-style: normal; font-variant: normal; =
font-weight: normal; letter-spacing: normal; line-height: normal; =
orphans: 2; text-indent: 0px; text-transform: none; white-space: normal; =
widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; =
-webkit-border-vertical-spacing: 0px; =
-webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: =
auto; -webkit-text-stroke-width: 0px; "><div style=3D"word-wrap: =
break-word; -webkit-nbsp-mode: space; -webkit-line-break: =
after-white-space; "><span class=3D"Apple-style-span" =
style=3D"border-collapse: separate; -webkit-border-horizontal-spacing: =
0px; -webkit-border-vertical-spacing: 0px; color: rgb(0, 0, 0); =
font-family: Helvetica; font-size: 12px; font-style: normal; =
font-variant: normal; font-weight: normal; letter-spacing: normal; =
line-height: normal; -webkit-text-decorations-in-effect: none; =
text-indent: 0px; -webkit-text-size-adjust: auto; text-transform: none; =
orphans: 2; white-space: normal; widows: 2; word-spacing: 0px; "><div =
style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; =
-webkit-line-break: after-white-space; "><div>Mark =
Lentczner</div><div>Sr. Systems Architect</div><div>Technology =
Integration</div><div>Linden Lab</div><div><br =
class=3D"khtml-block-placeholder"></div><div><a =
href=3D"mailto:markl@lindenlab.com">markl@lindenlab.com</a></div><div><br =
class=3D"khtml-block-placeholder"></div><div>Zero Linden</div><div><a =
href=3D"mailto:zero.linden@secondlife.com">zero.linden@secondlife.com</a><=
/div><div><br class=3D"khtml-block-placeholder"></div></div><br =
class=3D"Apple-interchange-newline"></span></div></span><br =
class=3D"Apple-interchange-newline"></div></span><br =
class=3D"Apple-interchange-newline"> </div><br></body></html>=

--Apple-Mail-20--453798811--

From Martin.Thomson@andrew.com  Wed Apr 15 19:03:17 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 362DD3A6BAF for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:03:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.535
X-Spam-Level: 
X-Spam-Status: No, score=-2.535 tagged_above=-999 required=5 tests=[AWL=0.063,  BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2VlXp4waMnYA for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:03:16 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 217603A6B9C for <hybi@ietf.org>; Wed, 15 Apr 2009 19:03:16 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_15_21_25_00
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from aopexbh2.andrew.com [10.86.20.25] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Wed, 15 Apr 2009 21:24:59 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by aopexbh2.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Wed, 15 Apr 2009 21:04:28 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C9BE37.ABF7C8A1"
Date: Wed, 15 Apr 2009 21:04:26 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF105A0CCF3@AHQEX1.andrew.com>
In-Reply-To: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] pointers, please?
Thread-Index: Acm+N24qfQY/euamSmapreoP1IGGeAAADDdw
References: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Mark Lentczner" <markl@lindenlab.com>, <hybi@ietf.org>
X-OriginalArrivalTime: 16 Apr 2009 02:04:28.0169 (UTC) FILETIME=[AC672790:01C9BE37]
Subject: Re: [hybi] pointers, please?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 02:03:17 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01C9BE37.ABF7C8A1
Content-Type: text/plain;
	charset="utf-8"
Content-Transfer-Encoding: base64
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------_=_NextPart_001_01C9BE37.ABF7C8A1
Content-Type: text/html;
	charset="utf-8"
Content-Transfer-Encoding: base64
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==

------_=_NextPart_001_01C9BE37.ABF7C8A1--


From markl@lindenlab.com  Wed Apr 15 19:19:39 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AE71D3A6905 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:19:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.449
X-Spam-Level: 
X-Spam-Status: No, score=-3.449 tagged_above=-999 required=5 tests=[AWL=0.150,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id y+5lhZooWRci for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:19:39 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 15D3E3A682B for <hybi@ietf.org>; Wed, 15 Apr 2009 19:19:39 -0700 (PDT)
Received: from [192.168.1.192] (m208-18.dsl.rawbw.com [198.144.208.18]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id 074D31414003 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:02:28 -0700 (PDT)
Message-Id: <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <20090414191717.GI32311@shareable.org>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Wed, 15 Apr 2009 18:46:26 -0700
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 02:19:39 -0000

On Apr 14, 2009, at 12:17 PM, Jamie Lokier wrote:
>   - The browser shares connections for each window/tab and handles
>     multiplexing automatically.  This requires the low-level
>     protocols to support multiplexing.


On Apr 15, 2009, at 2:36 AM, Greg Wilkins wrote:
> I think in this regard websocket API is a step in the wrong direction
> as it makes the developer directly request a connection.   Instead
> they should be directly requesting a channel and it should be up to
> the browser to decide if that is a new connection or a shared  
> multiplexed
> connection.

These are dead-on and an example of where, if the very commonly used  
features are part of the protocol, then the infrastructure can do a  
better job of handling the traffic.

Developing a protocol with an understanding of addressing of the bi- 
directional flow so that the infrastructure, being client environments  
(browsers being but one), proxies, caches and server frameworks can  
make different tradeoffs to achieve efficiency would be a great outcome.

This discussion seems convincing that clear abstraction of the  
interaction, and visibility of addressing would be of benefit to the  
infrastructure. Seems to me likely that meta-data (especially  
caching), content-type, and transfer encoding (especially chunked)  
would also be of benefit to the infrastructure.

	- Mark



Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From jonas@sicking.cc  Wed Apr 15 19:33:50 2009
Return-Path: <jonas@sicking.cc>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DD4413A6811 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:33:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id VP0Jk3CCT9Jd for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 19:33:50 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.29]) by core3.amsl.com (Postfix) with ESMTP id 0EC9B3A6961 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:33:49 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 5so302966ywh.49 for <hybi@ietf.org>; Wed, 15 Apr 2009 19:35:02 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.150.52.2 with SMTP id z2mr1379457ybz.28.1239849302205; Wed, 15  Apr 2009 19:35:02 -0700 (PDT)
In-Reply-To: <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>  <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>  <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>  <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com>
From: Jonas Sicking <jonas@sicking.cc>
Date: Wed, 15 Apr 2009 19:34:47 -0700
Message-ID: <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>
To: Mark Lentczner <markl@lindenlab.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 02:33:51 -0000

On Wed, Apr 15, 2009 at 6:46 PM, Mark Lentczner <markl@lindenlab.com> wrote=
:
> On Apr 14, 2009, at 12:17 PM, Jamie Lokier wrote:
>>
>> =A0- The browser shares connections for each window/tab and handles
>> =A0 =A0multiplexing automatically. =A0This requires the low-level
>> =A0 =A0protocols to support multiplexing.
>
>
> On Apr 15, 2009, at 2:36 AM, Greg Wilkins wrote:
>>
>> I think in this regard websocket API is a step in the wrong direction
>> as it makes the developer directly request a connection. =A0 Instead
>> they should be directly requesting a channel and it should be up to
>> the browser to decide if that is a new connection or a shared multiplexe=
d
>> connection.
>
> These are dead-on and an example of where, if the very commonly used
> features are part of the protocol, then the infrastructure can do a bette=
r
> job of handling the traffic.
>
> Developing a protocol with an understanding of addressing of the
> bi-directional flow so that the infrastructure, being client environments
> (browsers being but one), proxies, caches and server frameworks can make
> different tradeoffs to achieve efficiency would be a great outcome.
>
> This discussion seems convincing that clear abstraction of the interactio=
n,
> and visibility of addressing would be of benefit to the infrastructure.
> Seems to me likely that meta-data (especially caching), content-type, and
> transfer encoding (especially chunked) would also be of benefit to the
> infrastructure.

Do you, or anyone else, have suggestions for concrete changes to the
current websockets draft?

/ Jonas

From ian@hixie.ch  Wed Apr 15 22:00:19 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 044FE3A6C7E for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:00:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.281
X-Spam-Level: 
X-Spam-Status: No, score=-3.281 tagged_above=-999 required=5 tests=[AWL=-0.682, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id f2dGYNAwAqrg for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:00:18 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 125E03A6A48 for <hybi@ietf.org>; Wed, 15 Apr 2009 22:00:17 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 3419C125B60; Wed, 15 Apr 2009 22:01:30 -0700 (PDT)
Date: Thu, 16 Apr 2009 05:01:30 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Mark Nottingham <mnot@mnot.net>
In-Reply-To: <FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net>
Message-ID: <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net> <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com> <FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 05:00:19 -0000

On Thu, 16 Apr 2009, Mark Nottingham wrote:
> 
> Does anything need to be said -- at a protocol level -- about sending 
> events in the other direction over HTTP?

I don't understand the relevance of HTTP to this discussion. HTTP is about 
requests/response for REST-like sets of resources. The model that I'm 
interested in here is bidirectional full-duplex arbitrary data transfer, 
typically without any resources (in the REST sense) being involved. For 
example, for a multiplayer game, sending user commands and receiving 
continuous game state updates.

Trying to adapt HTTP to this purpose seems to me to make about as much 
sense as trying to adapt a library catalogue system for use as a telephone 
interchange.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From jamie@shareable.org  Wed Apr 15 22:01:16 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E374D3A6BF2 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:01:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.551
X-Spam-Level: 
X-Spam-Status: No, score=-3.551 tagged_above=-999 required=5 tests=[AWL=-0.952, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id F3JF0Sui2NjV for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:01:16 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 008AD3A6A48 for <hybi@ietf.org>; Wed, 15 Apr 2009 22:01:16 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LuJk1-0004xM-Ei; Thu, 16 Apr 2009 06:02:25 +0100
Date: Thu, 16 Apr 2009 06:02:25 +0100
From: Jamie Lokier <jamie@shareable.org>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
Message-ID: <20090416050225.GA18883@shareable.org>
References: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0CCF3@AHQEX1.andrew.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105A0CCF3@AHQEX1.andrew.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] pointers, please?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 05:01:17 -0000

Thomson, Martin wrote:
> Congestion control was the reason cited at the last meeting.Â
> It's a good reason.

(Another reason: additional connections add considerable overhead just
in packets transmitted: 6 for setup and teardown per connection).

Some people have said there is no need to limit the number of
connections made for WebSocket-style things: say 100 sockets from a
single client to a single server are not a problem.

(That's from 10 per window/tab (different app authors on the same page),
times 10 windows/tabs open to the same server).

Does congestion control mean it is important to reduce the number of
connections below that sort of level?

I think so, but I'd like to read other opinions.

-- Jam

From mike@dierken.com  Wed Apr 15 22:56:08 2009
Return-Path: <mike@dierken.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D1D613A684A for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:56:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.134
X-Spam-Level: 
X-Spam-Status: No, score=-2.134 tagged_above=-999 required=5 tests=[AWL=0.465,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id n1V67pqLmnnz for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 22:56:08 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.237]) by core3.amsl.com (Postfix) with ESMTP id F2E3D3A682A for <hybi@ietf.org>; Wed, 15 Apr 2009 22:56:07 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id k40so258674rvb.49 for <hybi@ietf.org>; Wed, 15 Apr 2009 22:57:20 -0700 (PDT)
Received: by 10.141.142.1 with SMTP id u1mr275645rvn.18.1239861440743; Wed, 15 Apr 2009 22:57:20 -0700 (PDT)
Received: from mercury (c-76-121-195-84.hsd1.wa.comcast.net [76.121.195.84]) by mx.google.com with ESMTPS id g22sm1874452rvb.13.2009.04.15.22.57.19 (version=SSLv3 cipher=RC4-MD5); Wed, 15 Apr 2009 22:57:19 -0700 (PDT)
From: "S. Mike Dierken" <mike@dierken.com>
To: "'Ian Hickson'" <ian@hixie.ch>, "'Mark Nottingham'" <mnot@mnot.net>
References: <49DEF171.4080506@mozilla.com><c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com><op.usaa24vt64w2qv@anne-van-kesterens-macbook.local><170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com><49E3C476.7090005@webtide.com><20090414191717.GI32311@shareable.org><AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net><op.usgd8uif64w2qv@annevk-t60.oslo.opera.com><FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net> <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com>
Date: Wed, 15 Apr 2009 22:57:23 -0700
Message-ID: <C91453ADE37A47B48960387C81BF5F0C@mercury>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
In-Reply-To: <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com>
Thread-Index: Acm+UGsVDcYYeAyTQ3yCIwdTXgsuYAABxBiQ
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 05:56:08 -0000

If I were writing a multi-player game, I'd definitely want to identify
portions of game state and media assets with something like a URI, have
various standardized content types, consistent methods (uniform interfaces),
etc. So HTTP concepts sound helpful.
The part that would really suck would be to do all this with a single
connection - having both machines talking to each other at the same time on
a single connection doesn't work with HTTP (receiving media assets like
textures would block outbound messages or higher priority inbound messages).
I'm not sure having two connections would help - it might, but I don't know.



> -----Original Message-----
> From: hybi-bounces@ietf.org [mailto:hybi-bounces@ietf.org] On 
> Behalf Of Ian Hickson
> Sent: Wednesday, April 15, 2009 10:02 PM
> To: Mark Nottingham
> Cc: hybi@ietf.org
> Subject: Re: [hybi] Addressing Minimal Events
> 
> On Thu, 16 Apr 2009, Mark Nottingham wrote:
> > 
> > Does anything need to be said -- at a protocol level -- 
> about sending 
> > events in the other direction over HTTP?
> 
> I don't understand the relevance of HTTP to this discussion. 
> HTTP is about requests/response for REST-like sets of 
> resources. The model that I'm interested in here is 
> bidirectional full-duplex arbitrary data transfer, typically 
> without any resources (in the REST sense) being involved. For 
> example, for a multiplayer game, sending user commands and 
> receiving continuous game state updates.
> 
> Trying to adapt HTTP to this purpose seems to me to make 
> about as much sense as trying to adapt a library catalogue 
> system for use as a telephone interchange.
> 
> -- 
> Ian Hickson               U+1047E                
> )\._.,--....,'``.    fL
> http://ln.hixie.ch/       U+263A                /,   _.. \   
> _\  ;`._ ,.
> Things that are impossible just take longer.   
> `._.-(,_..'--(,_..'`-.;.'
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
> 


From mjs@apple.com  Wed Apr 15 23:56:41 2009
Return-Path: <mjs@apple.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EA4553A69D1 for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 23:56:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.292
X-Spam-Level: 
X-Spam-Status: No, score=-106.292 tagged_above=-999 required=5 tests=[AWL=0.307, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id EpTpw1Vi70DC for <hybi@core3.amsl.com>; Wed, 15 Apr 2009 23:56:41 -0700 (PDT)
Received: from mail-out3.apple.com (mail-out3.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id 3593B3A67B4 for <hybi@ietf.org>; Wed, 15 Apr 2009 23:56:41 -0700 (PDT)
Received: from relay10.apple.com (relay10.apple.com [17.128.113.47]) by mail-out3.apple.com (Postfix) with ESMTP id 1CE7B5CCF8EE for <hybi@ietf.org>; Wed, 15 Apr 2009 23:57:54 -0700 (PDT)
Received: from relay10.apple.com (unknown [127.0.0.1]) by relay10.apple.com (Symantec Brightmail Gateway) with ESMTP id 03B2B28059 for <hybi@ietf.org>; Wed, 15 Apr 2009 23:57:54 -0700 (PDT)
X-AuditID: 1180712f-a6966bb0000012d3-1b-49e6d6f1de3f
Received: from gertie.apple.com (gertie.apple.com [17.151.62.15]) by relay10.apple.com (Apple SCV relay) with ESMTP id D1ACD28054 for <hybi@ietf.org>; Wed, 15 Apr 2009 23:57:53 -0700 (PDT)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Received: from [17.151.123.186] by gertie.apple.com (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008; 32bit)) with ESMTPSA id <0KI600L0TM0HFE00@gertie.apple.com> for hybi@ietf.org; Wed, 15 Apr 2009 23:57:53 -0700 (PDT)
Message-id: <08302DB5-806D-4A2A-89B1-089EB2D41537@apple.com>
From: Maciej Stachowiak <mjs@apple.com>
To: Mark Lentczner <markl@lindenlab.com>
In-reply-to: <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com>
Date: Wed, 15 Apr 2009 23:57:52 -0700
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com>
X-Mailer: Apple Mail (2.930.3)
X-Brightmail-Tracker: AAAAAA==
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 06:56:42 -0000

On Apr 15, 2009, at 6:46 PM, Mark Lentczner wrote:

> On Apr 14, 2009, at 12:17 PM, Jamie Lokier wrote:
>>  - The browser shares connections for each window/tab and handles
>>    multiplexing automatically.  This requires the low-level
>>    protocols to support multiplexing.
>
>
> On Apr 15, 2009, at 2:36 AM, Greg Wilkins wrote:
>> I think in this regard websocket API is a step in the wrong direction
>> as it makes the developer directly request a connection.   Instead
>> they should be directly requesting a channel and it should be up to
>> the browser to decide if that is a new connection or a shared  
>> multiplexed
>> connection.
>
> These are dead-on and an example of where, if the very commonly used  
> features are part of the protocol, then the infrastructure can do a  
> better job of handling the traffic.
>
> Developing a protocol with an understanding of addressing of the bi- 
> directional flow so that the infrastructure, being client  
> environments (browsers being but one), proxies, caches and server  
> frameworks can make different tradeoffs to achieve efficiency would  
> be a great outcome.
>
> This discussion seems convincing that clear abstraction of the  
> interaction, and visibility of addressing would be of benefit to the  
> infrastructure. Seems to me likely that meta-data (especially  
> caching), content-type, and transfer encoding (especially chunked)  
> would also be of benefit to the infrastructure.

It seems to me that if WebSocket is extended to support subprotocol  
negotiation, then in the future we could develop a Multiplexed- 
WebSocket-Over-WebSocket protocol and user agents could use it  
transparently when available to handle multiple channels from the same  
security origin to the same host/port. Nothing about the API directly  
exposes the wire protocol format or the fact that there is a dedicated  
TCP socket. This is part of why I think subprotocol negotiation is  
essential even for the first version of the spec, but multiplexing is  
not.

Regards,
Maciej


From gregw@webtide.com  Thu Apr 16 00:06:13 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6A3013A6B2E for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:06:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.548
X-Spam-Level: 
X-Spam-Status: No, score=-2.548 tagged_above=-999 required=5 tests=[AWL=0.051,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uTGuQNBxAgOr for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:06:12 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.232]) by core3.amsl.com (Postfix) with ESMTP id 6EE513A682E for <hybi@ietf.org>; Thu, 16 Apr 2009 00:06:12 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id b25so2917967rvf.5 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:07:25 -0700 (PDT)
Received: by 10.115.32.18 with SMTP id k18mr573554waj.98.1239865645182; Thu, 16 Apr 2009 00:07:25 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id n20sm3908025pof.10.2009.04.16.00.07.22 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 16 Apr 2009 00:07:24 -0700 (PDT)
Message-ID: <49E6D923.5030406@webtide.com>
Date: Thu, 16 Apr 2009 17:07:15 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com>
In-Reply-To: <BF52D439-13B9-4E77-BE0A-2CB17D61E284@lindenlab.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] pointers, please?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 07:06:13 -0000

Mark Lentczner wrote:
>> The same-origin security policy.
> Can someone point me at a text on this?
> 
> Also, while I understand the difficulties that the two connection limit 
> imposes on applications doing Comet/AJAX style operations, I couldn't 
> seem to find a discussion of why that limit was added to HTTP?
> 
> I don't want to incite re-opening of these issues here - I'm just 
> looking for pointers....   at least for now :-)


I don't know if limiting connections on the server was a motivation
for the 2 connection limit, but as a developer of a HTTP server, I
appreciate the limit as every connection to the server consumes some
kernel resources and buffers.

Typically I aim of a server node to handle several 10s of thousands
of simultaneous users.   If each of those users starts using 8 connections
instead of 2, then my FD limit needs to be increased from 40k-ish
to 120,000k-ish and my buffer memory requirements are worst case x 4!
That is possible to do, but I'm not exactly happy that all that extra
memory and server capabilities I now have are being consumed essentially
because client side can't coordinate in order to multiplex.

I'd much rather have a few extra bytes per datagram indicating a
channel id. rather than an extra buffer server side for every connection
created in lieu of channels.


I've also seen people saying that increasing the number of connections
from their browser improves the throughput they receive from the network.
eg http://www.speedguide.net/read_articles.php?id=2448
This may be for a congested network trying to fair share capacity between
many connections - but it will not work if everybody on the network
increases their limit to 8 connections.... then somebody will increase
to 16 and let the arms race commence!

I really love how you see pages like this:
   http://support.microsoft.com/default.aspx?scid=KB;en-us;q183110
which demonstrates that the browser 2 connection limit is well
known, so that modern browsers that violate it are doing so
very knowingly against the RFC recommendations.



cheers

From gregw@webtide.com  Thu Apr 16 00:16:41 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3B7D03A6B40 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:16:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.549
X-Spam-Level: 
X-Spam-Status: No, score=-2.549 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id sepCnWDeSlCn for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:16:40 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.181]) by core3.amsl.com (Postfix) with ESMTP id 12D653A6B2E for <hybi@ietf.org>; Thu, 16 Apr 2009 00:16:40 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so137980waf.5 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:17:52 -0700 (PDT)
Received: by 10.115.108.1 with SMTP id k1mr576933wam.122.1239866272906; Thu, 16 Apr 2009 00:17:52 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id m26sm4257433pof.14.2009.04.16.00.17.48 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 16 Apr 2009 00:17:51 -0700 (PDT)
Message-ID: <49E6DB96.3090203@webtide.com>
Date: Thu, 16 Apr 2009 17:17:42 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <49DEF171.4080506@mozilla.com>	<c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>	<op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>	<170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>	<49E3C476.7090005@webtide.com>	<20090414191717.GI32311@shareable.org>	<AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net>	<op.usgd8uif64w2qv@annevk-t60.oslo.opera.com>	<FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net> <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 07:16:41 -0000

Ian Hickson wrote:
> I don't understand the relevance of HTTP to this discussion. 

HTTP is our starting point.

It is how our browsers initiate connections
It is the protocol that is most likely to get through firewalls
It is a brilliant example of how an extensible flexible protocol can
adapt to changing usage.


So if you want to have a bidirection full-duplex protocol, then the
chances are the connection will be created by a HTTP client talking
to a HTTP server and then doing an upgrade.

For that upgrade to be successful, we are going to need to engage
all the firewall/gateway/proxy/cache venders and developers so that
they will support upgrade either to a specific protocol or to
a class of protocols.


Websocket has been put forward as a candidate for being such
a specific protocol.     However, while there is a lot of goodness
in websocket, I don't think it has the extensibility and flexibility
needed for it to be successful as the only available upgrade target.

If websocket is just one of many possible upgrade targets, then
I don't think that is a problem and websocket can continue to focus
on it's prime use-cases.    However, if websocket is to be the
only widely supported upgrade protocol, then I think it needs to
learn some lessons from HTTP as to how to keep most people mostly
happy most of the time.

cheers









From gregw@webtide.com  Thu Apr 16 00:18:44 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 15A103A6B40 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:18:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.55
X-Spam-Level: 
X-Spam-Status: No, score=-2.55 tagged_above=-999 required=5 tests=[AWL=0.049,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rNhrkTVKGmxF for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:18:43 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.239]) by core3.amsl.com (Postfix) with ESMTP id 6BCC43A6B14 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:18:43 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id b25so2921918rvf.5 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:19:50 -0700 (PDT)
Received: by 10.114.26.11 with SMTP id 11mr583799waz.38.1239866390872; Thu, 16 Apr 2009 00:19:50 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id m29sm3912960poh.26.2009.04.16.00.19.47 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 16 Apr 2009 00:19:50 -0700 (PDT)
Message-ID: <49E6DC0E.2070809@webtide.com>
Date: Thu, 16 Apr 2009 17:19:42 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com> <49E5AA86.9040100@webtide.com> <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com> <49E66DDD.8030201@webtide.com> <b7197f6c0904151753ne354d8cl62d0e66501e3b80f@mail.gmail.com>
In-Reply-To: <b7197f6c0904151753ne354d8cl62d0e66501e3b80f@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 07:18:44 -0000

Frank Salim wrote:
>> Connections are a real problem and I believe that the Websocket
>> API will encourage the usage of more sockets.
> 
> WebSockets are not subject to the browsers' restrictions on the number
> of HTTP connections. I don't think this is a problem.

Frank,

Avoiding the limit doesn't avoid the good reasons the limit was put there
in the first place.

regards



From gregw@webtide.com  Thu Apr 16 00:36:37 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CF4C73A6D05 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:36:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.552
X-Spam-Level: 
X-Spam-Status: No, score=-2.552 tagged_above=-999 required=5 tests=[AWL=0.047,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id x-k0MfpJEpWF for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 00:36:37 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.178]) by core3.amsl.com (Postfix) with ESMTP id 23F0F3A69D2 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:36:37 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so141022waf.5 for <hybi@ietf.org>; Thu, 16 Apr 2009 00:37:49 -0700 (PDT)
Received: by 10.115.32.8 with SMTP id k8mr591015waj.54.1239867469756; Thu, 16 Apr 2009 00:37:49 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id a8sm3958275poa.7.2009.04.16.00.37.45 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 16 Apr 2009 00:37:48 -0700 (PDT)
Message-ID: <49E6E042.8020006@webtide.com>
Date: Thu, 16 Apr 2009 17:37:38 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Jonas Sicking <jonas@sicking.cc>, hybi@ietf.org
References: <49DEF171.4080506@mozilla.com>	<c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local>	<170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com> <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>
In-Reply-To: <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 07:36:37 -0000

Jonas Sicking wrote:

> 
> Do you, or anyone else, have suggestions for concrete changes to the
> current websockets draft?

Jonas,

certainly:

Support for a frame type that supports meta data plus
binary content in a standard format (HTTP-style would be fine).

Support for multiplexing datagrams over a single connection.
This could be on a subset of the datagram types (eg all those
with high bit of frame type vale set would have a channel number
sent and then the normal frame).

Decouple of the javascript API from connection state.  Javascript
developers should be able to develop against a datagrams centric
API and not be concerned if they are given a dedicated connection
or a multiplexed connection.

readyState, onopen, onclose and disconnect() should be on a separate
interface to onmessage and postMessage().  Frameworks and/or browsers
should have the opportunity to manage connections and use a
multiplexing datagram type.

A readable structured specification in RFC style rather than
narrative style (eg 1 send 0x00, 2, encode data, 3. send 0xff)

I'm happy to write these up in more details. Where is the best place
to send such feedback?   Also where is the latest draft of the
spec - I can find numerous versions, but all a little different.

regards

From markl@lindenlab.com  Thu Apr 16 09:31:34 2009
Return-Path: <markl@lindenlab.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 93F413A67EC for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 09:31:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.558
X-Spam-Level: 
X-Spam-Status: No, score=-3.558 tagged_above=-999 required=5 tests=[AWL=0.041,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HoQszn9cYSNQ for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 09:31:33 -0700 (PDT)
Received: from tammy.lindenlab.com (tammy.lindenlab.com [64.154.223.128]) by core3.amsl.com (Postfix) with ESMTP id 054DB28C0E8 for <hybi@ietf.org>; Thu, 16 Apr 2009 09:30:42 -0700 (PDT)
Received: from nil.lindenlab.com (nil.lindenlab.com [10.1.16.4]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by tammy.lindenlab.com (Postfix) with ESMTP id ED9D73DBC449 for <hybi@ietf.org>; Thu, 16 Apr 2009 09:31:55 -0700 (PDT)
Message-Id: <F7BE232F-EAA1-44E1-89EB-D78B46AC7178@lindenlab.com>
From: Mark Lentczner <markl@lindenlab.com>
To: hybi@ietf.org
In-Reply-To: <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Thu, 16 Apr 2009 09:31:55 -0700
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com> <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>
X-Mailer: Apple Mail (2.930.3)
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 16:31:34 -0000

On Apr 15, 2009, at 7:34 PM, Jonas Sicking wrote:
> Do you, or anyone else, have suggestions for concrete changes to the
> current websockets draft?

I'm putting together a draft, incorporating many of the ideas and  
thoughts I've read here. However, it is more in style to my Reverse  
HTTP draft[1], than WebSockets. It will incorporate HTTP message  
semantics, bidirectionally, synchronicity, request/response and event  
support. I think it will do all that without being significantly more  
complex than HTTP, and do so by making use of the bulk of the HTTP spec.

Give me about a week or so....

	- Mark

[1] http://tools.ietf.org/html/draft-lentczner-rhttp-00

Mark Lentczner
Sr. Systems Architect
Technology Integration
Linden Lab

markl@lindenlab.com

Zero Linden
zero.linden@secondlife.com






From jonas@sicking.cc  Thu Apr 16 10:52:32 2009
Return-Path: <jonas@sicking.cc>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BE0083A6B3D for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 10:52:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id juLKrUWziz9C for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 10:52:31 -0700 (PDT)
Received: from mail-qy0-f123.google.com (mail-qy0-f123.google.com [209.85.221.123]) by core3.amsl.com (Postfix) with ESMTP id B81DA3A69E6 for <hybi@ietf.org>; Thu, 16 Apr 2009 10:52:31 -0700 (PDT)
Received: by qyk29 with SMTP id 29so125056qyk.29 for <hybi@ietf.org>; Thu, 16 Apr 2009 10:53:44 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.224.2.138 with SMTP id 10mr2144766qaj.299.1239904424313; Thu,  16 Apr 2009 10:53:44 -0700 (PDT)
In-Reply-To: <49E6E042.8020006@webtide.com>
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com>  <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>  <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>  <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com> <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com>  <49E6E042.8020006@webtide.com>
From: Jonas Sicking <jonas@sicking.cc>
Date: Thu, 16 Apr 2009 10:53:29 -0700
Message-ID: <63df84f0904161053l42ec666l91422cde418033ad@mail.gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 17:52:32 -0000

On Thu, Apr 16, 2009 at 12:37 AM, Greg Wilkins <gregw@webtide.com> wrote:
> Jonas Sicking wrote:
>> Do you, or anyone else, have suggestions for concrete changes to the
>> current websockets draft?
>
> Support for a frame type that supports meta data plus
> binary content in a standard format (HTTP-style would be fine).
>
> Support for multiplexing datagrams over a single connection.
> This could be on a subset of the datagram types (eg all those
> with high bit of frame type vale set would have a channel number
> sent and then the normal frame).
>
> Decouple of the javascript API from connection state. =A0Javascript
> developers should be able to develop against a datagrams centric
> API and not be concerned if they are given a dedicated connection
> or a multiplexed connection.
>
> readyState, onopen, onclose and disconnect() should be on a separate
> interface to onmessage and postMessage(). =A0Frameworks and/or browsers
> should have the opportunity to manage connections and use a
> multiplexing datagram type.
>
> A readable structured specification in RFC style rather than
> narrative style (eg 1 send 0x00, 2, encode data, 3. send 0xff)
>
> I'm happy to write these up in more details. Where is the best place
> to send such feedback? =A0 Also where is the latest draft of the
> spec - I can find numerous versions, but all a little different.

I believe the latest version can be found at
http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol

As I understand it feedback is supposed to go to the list we're
currently talking on.

/ Jonas

From prescod@gmail.com  Thu Apr 16 11:25:15 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DFB263A6C57 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:25:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.715
X-Spam-Level: 
X-Spam-Status: No, score=-1.715 tagged_above=-999 required=5 tests=[AWL=0.262,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 76Cti9t8dX+J for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:25:11 -0700 (PDT)
Received: from mail-gx0-f160.google.com (mail-gx0-f160.google.com [209.85.217.160]) by core3.amsl.com (Postfix) with ESMTP id 5B0F83A6CEB for <hybi@ietf.org>; Thu, 16 Apr 2009 11:25:09 -0700 (PDT)
Received: by gxk4 with SMTP id 4so1282344gxk.13 for <hybi@ietf.org>; Thu, 16 Apr 2009 11:26:22 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=WYD9aGoGxR43SF5Z0LSDSKpgleEzHWE+L1l5RCUDMmo=; b=o0eGu6bgO7oqI7AwLP0b9I1zXhDY0JQ7QZY/HeMbh1ImkwtKuUPGmvdlOoFlRa2RRs YF12lO10Gs3/1Yh06f7gSrw5pZtkTW5RNErORcDd9RVcCmiAZ/3zdnpyeqNpeUCBkiYs RVmriJSR2zdrf4Ilj8rJMD/KVIkpaAyAiHNNY=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=TA2SV/Bf79TCA5TZN0vy7InK4qaxM+gF9hVUrtZWLXBwU99oVfQqFXNN55ebrPXngo JqN5sNJMV98h4+4AttJg2UdbLWNroOWoXeQ1HhnC9jszVMsgZYtlMyPK8BEqVqIK2dLv 6pPnvRS8vcNM8Qc/CNaEqIwOZzABbzT4Ju5dA=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.101.68.19 with SMTP id v19mr2388367ank.70.1239906381895; Thu,  16 Apr 2009 11:26:21 -0700 (PDT)
In-Reply-To: <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>
Date: Thu, 16 Apr 2009 11:26:21 -0700
X-Google-Sender-Auth: 011f3d4f8b5f2d8b
Message-ID: <1cb725390904161126ke0a158y50a011a6632a2fb0@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 18:25:15 -0000

It is obvious that the websockets design is heavily influenced by the
use case of a relatively inexperienced web programmer programming
directly against the API. Therefore, I would greatly appreciate an
answer to the question of how *precisely* that would work. If the
requirement to be directly implementable by inexperienced programmers
is key to the design, then please describe how this happens in your
proposed design?

Code or pseudocode or narrative are all fine. I'd just like to know
how you go about writing a WebSockets server in PHP. Roughly how many
lines of code is it, what libraries do you need to use, how do you
handle multiple overlapping incoming connections, and how do you
manage state?

On Mon, Apr 13, 2009 at 6:14 PM, Paul Prescod <paul@prescod.net> wrote:
> ...
>
> Please be very specific about HOW a 12-year old who knows PHP would
> write a multiplayer game using PHP and WebSockets.
>
> =C2=A0Paul Prescod
>

From frank.salim@kaazing.com  Thu Apr 16 11:43:38 2009
Return-Path: <frank.salim@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 97F3D3A6C92 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:43:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.977
X-Spam-Level: 
X-Spam-Status: No, score=-1.977 tagged_above=-999 required=5 tests=[AWL=0.001,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GpZYbA3XiQSB for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:43:37 -0700 (PDT)
Received: from yx-out-2324.google.com (yx-out-2324.google.com [74.125.44.30]) by core3.amsl.com (Postfix) with ESMTP id B71F73A6B69 for <hybi@ietf.org>; Thu, 16 Apr 2009 11:43:37 -0700 (PDT)
Received: by yx-out-2324.google.com with SMTP id 8so770172yxm.49 for <hybi@ietf.org>; Thu, 16 Apr 2009 11:44:50 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.49.8 with SMTP id w8mr2001076agw.117.1239907490423; Thu, 16  Apr 2009 11:44:50 -0700 (PDT)
In-Reply-To: <49E6DC0E.2070809@webtide.com>
References: <49DEF171.4080506@mozilla.com> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com>  <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org>  <b7197f6c0904141933o14c2cc53qf687458d3f68745e@mail.gmail.com>  <49E5AA86.9040100@webtide.com> <79ea848f0904150942h6038bf3bs9fc867b918c87a36@mail.gmail.com>  <49E66DDD.8030201@webtide.com> <b7197f6c0904151753ne354d8cl62d0e66501e3b80f@mail.gmail.com>  <49E6DC0E.2070809@webtide.com>
From: Frank Salim <frank.salim@kaazing.com>
Date: Thu, 16 Apr 2009 11:44:35 -0700
Message-ID: <b7197f6c0904161144m447049d6t1599943a7c552c54@mail.gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 18:43:38 -0000

There are different reasons for imposing a limit on the number of HTTP
connections. A page with 100 images might innocently cause 100 HTTP
connections to open simultaneously. A malicious page might
intentionally do the same with embedded content.

In the case of malicious cross-origin connections, a WebSocket server
should have all the information it needs to close them quickly.

As far as client-side congestion goes, we leave it up to authors to
police themselves regarding their CPU usage and memory consumption.
Badly behaved sites get complaints and drive away their users. I see
this the same way.

-Frank

On Thu, Apr 16, 2009 at 12:19 AM, Greg Wilkins <gregw@webtide.com> wrote:
> Frank Salim wrote:
>>>
>>> Connections are a real problem and I believe that the Websocket
>>> API will encourage the usage of more sockets.
>>
>> WebSockets are not subject to the browsers' restrictions on the number
>> of HTTP connections. I don't think this is a problem.
>
> Frank,
>
> Avoiding the limit doesn't avoid the good reasons the limit was put there
> in the first place.
>
> regards
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

From ian@hixie.ch  Thu Apr 16 11:56:21 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E58413A6C57 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:56:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.269
X-Spam-Level: 
X-Spam-Status: No, score=-3.269 tagged_above=-999 required=5 tests=[AWL=-0.669, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ilu6wKjo1psa for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 11:56:21 -0700 (PDT)
Received: from looneymail-a1.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 8A5393A6CBE for <hybi@ietf.org>; Thu, 16 Apr 2009 11:56:11 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a1.g.dreamhost.com (Postfix) with ESMTP id 9742415D56C; Thu, 16 Apr 2009 11:57:00 -0700 (PDT)
Date: Thu, 16 Apr 2009 18:57:00 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904161126ke0a158y50a011a6632a2fb0@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904161856060.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>  <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>  <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>  <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <1cb725390904161126ke0a158y50a011a6632a2fb0@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 18:56:22 -0000

On Thu, 16 Apr 2009, Paul Prescod wrote:
>
> It is obvious that the websockets design is heavily influenced by the 
> use case of a relatively inexperienced web programmer programming 
> directly against the API. Therefore, I would greatly appreciate an 
> answer to the question of how *precisely* that would work. If the 
> requirement to be directly implementable by inexperienced programmers is 
> key to the design, then please describe how this happens in your 
> proposed design?

I haven't ignored this e-mail, I just can't send out a reply to it in 
three minutes, and have a couple of tasks I need to get done ASAP. I 
intend to reply to this e-mail and a number of others later this week or 
early next week.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Thu Apr 16 12:00:21 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5F2613A6997 for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 12:00:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.731
X-Spam-Level: 
X-Spam-Status: No, score=-1.731 tagged_above=-999 required=5 tests=[AWL=0.246,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kyRO5xqwKtZP for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 12:00:20 -0700 (PDT)
Received: from mail-gx0-f160.google.com (mail-gx0-f160.google.com [209.85.217.160]) by core3.amsl.com (Postfix) with ESMTP id 8F99528C2BB for <hybi@ietf.org>; Thu, 16 Apr 2009 12:00:12 -0700 (PDT)
Received: by gxk4 with SMTP id 4so1316775gxk.13 for <hybi@ietf.org>; Thu, 16 Apr 2009 12:01:25 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type:content-transfer-encoding; bh=zuhFZ0wA4zNcsq4z5p1pK3r5AC9pRwEttyidMgx/rQY=; b=WOPHk7MiZlRl4YluQbkV+SYSvdmETUV1nJdB6vVtTVtM7ZP23xQ4WVizNhqIjm7YxV DincMuMu7SejNjGZBAPIZt9OLPbA+VWkF9lXt+4vT55+1ZtbALzkErtykkEMKK/FpptU o4fMXKpR+oefnXP4xHI03kYSpaARu72coVLXg=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=fOFquwQvGpvZqRtYKEvlajjFCHKo5S7sq0wuIEyb433Kkup7m5bTx44uuDUXq1ujWf J9s7C/OioOsI9OZJcA0CAXXHHsbS2WQqx11aR5983TSMKoniv3FsB+vmZHqXyzYtazKV 2w5/AbcrnQE+f8ISPkVKP7j6sG7Vv8m23CL4Q=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.100.251.8 with SMTP id y8mr2448914anh.19.1239908485375; Thu,  16 Apr 2009 12:01:25 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904161856060.10339@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <1cb725390904161126ke0a158y50a011a6632a2fb0@mail.gmail.com> <Pine.LNX.4.62.0904161856060.10339@hixie.dreamhostps.com>
Date: Thu, 16 Apr 2009 12:01:25 -0700
X-Google-Sender-Auth: 0b5e5426d99bfad7
Message-ID: <1cb725390904161201m37229da5j6ac2d6280a42d195@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 19:00:21 -0000

Thank you.  I only brought it up again because the pace of email is so
high that posts from a few days ago seem "buried" and lost.  I
understand that one cannot do everything at once.

 Paul Prescod

From jamie@shareable.org  Thu Apr 16 14:56:22 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 070323A6B4A for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 14:56:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.508
X-Spam-Level: 
X-Spam-Status: No, score=-3.508 tagged_above=-999 required=5 tests=[AWL=-0.909, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7CkR0QNUTBYz for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 14:56:21 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 2F2823A69F7 for <hybi@ietf.org>; Thu, 16 Apr 2009 14:56:21 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LuZaJ-0003qD-Ps; Thu, 16 Apr 2009 22:57:27 +0100
Date: Thu, 16 Apr 2009 22:57:27 +0100
From: Jamie Lokier <jamie@shareable.org>
To: "S. Mike Dierken" <mike@dierken.com>
Message-ID: <20090416215727.GB10578@shareable.org>
References: <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com> <C91453ADE37A47B48960387C81BF5F0C@mercury>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <C91453ADE37A47B48960387C81BF5F0C@mercury>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 21:56:22 -0000

S. Mike Dierken wrote:
> If I were writing a multi-player game, I'd definitely want to identify
> portions of game state and media assets with something like a URI, have
> various standardized content types, consistent methods (uniform interfaces),
> etc. So HTTP concepts sound helpful.

I agree.  But I know some people don't want those concepts; to them,
they are overhead.  I'm leaning towards a protocol which makes the
concepts easily available (and I think most people will use them), but
makes them optional.

> The part that would really suck would be to do all this with a single
> connection - having both machines talking to each other at the same time on
> a single connection doesn't work with HTTP (receiving media assets like
> textures would block outbound messages or higher priority inbound messages).
> I'm not sure having two connections would help - it might, but I don't know.

That's why we talk about multiplexing and flow control.  A large media
asset does not block small messages over a well designed multiplexing
protocol, if it's also well implemented.  It does delay them a bit,
but the same happens with separate connections (with different
charactistics in detail).

If you've ever made a tunnel using "ssh -R" or "ssh -L" you'll have
seen this working ok.

You might be thinking of HTTP pipelining, where large assets do block
everything; that is one of many problems with HTTP pipelining, which
is why I don't like it :-)

Actually I lied.  For a game you would might benefit from two TCP
connections with IPTOS = INTERACTIVE and IPTOS = BULK, to help
separate the streams and improve prioritising at routers.  That isn't
in the WebSockets API at the moment.

(Though a _really_ real-time game would use UDP anyway...)

Having two connections isn't a problem.  Having many connections (say
20) is worse for delays, if they are all active.  It tends to increase
packet delivery jitter for all of them, making TCP's heuristics less
effective.

Also you need to send 2*number_of_connections keepalive packets every
30 seconds or so to keep NATs open.  With many connections that adds up.

-- Jamie

From jamie@shareable.org  Thu Apr 16 15:11:44 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 709F828C0EB for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 15:11:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.469
X-Spam-Level: 
X-Spam-Status: No, score=-3.469 tagged_above=-999 required=5 tests=[AWL=-0.870, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id q+McAI9jgLSa for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 15:11:43 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id B259828C0E9 for <hybi@ietf.org>; Thu, 16 Apr 2009 15:11:43 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LuZpG-0003tX-TB; Thu, 16 Apr 2009 23:12:54 +0100
Date: Thu, 16 Apr 2009 23:12:54 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090416221254.GC10578@shareable.org>
References: <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net> <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com> <FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net> <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com> <49E6DB96.3090203@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49E6DB96.3090203@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 22:11:44 -0000

Greg Wilkins wrote:
> Ian Hickson wrote:
> So if you want to have a bidirection full-duplex protocol, then the
> chances are the connection will be created by a HTTP client talking
> to a HTTP server and then doing an upgrade.
> 
> For that upgrade to be successful, we are going to need to engage
> all the firewall/gateway/proxy/cache venders and developers so that
> they will support upgrade either to a specific protocol or to
> a class of protocols.

That's one way, but there are other more current-proxy-compatible ways
to do it.  They are hacks, but work.

-- Jamie

From mario.balibrera@gmail.com  Thu Apr 16 16:28:16 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id F31653A6DDF for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 16:28:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.448
X-Spam-Level: 
X-Spam-Status: No, score=-2.448 tagged_above=-999 required=5 tests=[AWL=0.150,  BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Vf2XrxN2nqkd for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 16:28:13 -0700 (PDT)
Received: from mail-qy0-f136.google.com (mail-qy0-f136.google.com [209.85.221.136]) by core3.amsl.com (Postfix) with ESMTP id B1A913A67C0 for <hybi@ietf.org>; Thu, 16 Apr 2009 16:28:13 -0700 (PDT)
Received: by qyk42 with SMTP id 42so167765qyk.29 for <hybi@ietf.org>; Thu, 16 Apr 2009 16:29:26 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=ZEqdZK/i8Sd/S3CM0jTRxz+zPsY9fFCmpOE7MYXig18=; b=ZHs6KzCa1On3XUZUA1KPZgMV98PS/R/GmQHi1saaBg46iL1/QvkWN02SlLDbDwDCBz gitqsX9De7aOEFE6UBct1yS/SFjzC8uS+b3qY2iDA/1qGJZ28NRbQULa+Xdzab8rqEKW W4XXNM1ae6Bchozau15kvKmCPdY8bg8mdupBQ=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=rDWb2jp+Tmat6b2MTYHNlfEozD2DsxJe+9rAf3I8Ip61GN23QNAkKpLVmeyKc43uYF qyYp/jbSwW0/2lu9exTitET1OwNemV36gXDOpvQyK3LNbY+Z2EJySFl+xqi4ab94ZUJJ u5qrucH81rPM/2wpJ5CkX4e1fRZuSl02i9xAU=
MIME-Version: 1.0
Received: by 10.220.74.3 with SMTP id s3mr1318734vcj.54.1239924566570; Thu, 16  Apr 2009 16:29:26 -0700 (PDT)
In-Reply-To: <20090416221254.GC10578@shareable.org>
References: <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <AE64FFAD-05CC-473C-AB13-145A4FE4FF4E@mnot.net> <op.usgd8uif64w2qv@annevk-t60.oslo.opera.com> <FBE1482E-1E56-4A3A-BFBA-8187EAC14D95@mnot.net> <Pine.LNX.4.62.0904160448080.10339@hixie.dreamhostps.com> <49E6DB96.3090203@webtide.com> <20090416221254.GC10578@shareable.org>
Date: Thu, 16 Apr 2009 16:29:26 -0700
Message-ID: <79ea848f0904161629q2abd8bc1l9b3030bb89d66cd9@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Jamie Lokier <jamie@shareable.org>
Content-Type: multipart/alternative; boundary=0016362850fe56b5c20467b46f29
Cc: hybi@ietf.org
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Apr 2009 23:28:16 -0000

--0016362850fe56b5c20467b46f29
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

> The part that would really suck would be to do all this with a single
> connection - having both machines talking to each other at the same time
on
> a single connection doesn't work with HTTP (receiving media assets like
> textures would block outbound messages or higher priority inbound
messages).
> I'm not sure having two connections would help - it might, but I don't
know.

That's why we talk about multiplexing and flow control.  A large media
asset does not block small messages over a well designed multiplexing
protocol, if it's also well implemented.  It does delay them a bit,
but the same happens with separate connections (with different
charactistics in detail).

I think an important fact is sort of being overlooked here. When you add
something new to your toolset, the rest of your tools don't disappear. If
your application needs to transmit a file, you should consider using a
protocol designed for that exact thing: HTTP. Of course, you can send
anything you want over your streaming connection, but if for your particular
application some other tool is a better fit, don't be shy! Consider how XMPP
decided to support buddy icons (http://xmpp.org/extensions/xep-0084.html).
They thought it was best to provide two options: you can either send the
file over the wire, or simply transmit a url.

As Jamie points out, another solution involves separate connections, which
rather makes sense, as game logic and images will very likely be on separate
servers.

No one here is proposing eliminating HTTP. Ideally, these discussions will
produce something that perfectly complements it - that fills in where HTTP
is lacking without too much overlap. That said, I certainly don't mean to
suggest that HTTP has no place in this conversation (WebSocket, after all,
uses HTTP conventions to establish a connection). I simply think that every
feature of HTTP doesn't necessarily belong in the protocol we choose to sit
beside it, and that the various features of HTTP will continue to work, even
if we settle on Streaming Pig Latin Sockets.

As Frank pointed out, giving developers a flexible tool is a GOOD IDEA, and
if we don't, they'll just end up hacking out the features they would have
implemented reasonably given something more general (see: Comet).

Mario Balibrera

On Thu, Apr 16, 2009 at 3:12 PM, Jamie Lokier <jamie@shareable.org> wrote:

> Greg Wilkins wrote:
> > Ian Hickson wrote:
> > So if you want to have a bidirection full-duplex protocol, then the
> > chances are the connection will be created by a HTTP client talking
> > to a HTTP server and then doing an upgrade.
> >
> > For that upgrade to be successful, we are going to need to engage
> > all the firewall/gateway/proxy/cache venders and developers so that
> > they will support upgrade either to a specific protocol or to
> > a class of protocols.
>
> That's one way, but there are other more current-proxy-compatible ways
> to do it.  They are hacks, but work.
>
> -- Jamie
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--0016362850fe56b5c20467b46f29
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

&gt; The part that would really suck would be to do all this with a single<=
br>
&gt; connection - having both machines talking to each other at the same ti=
me on<br>
&gt; a single connection doesn&#39;t work with HTTP (receiving media assets=
 like<br>
&gt; textures would block outbound messages or higher priority inbound mess=
ages).<br>
&gt; I&#39;m not sure having two connections would help - it might, but I d=
on&#39;t know.<br>
<br>
That&#39;s why we talk about multiplexing and flow control. =A0A large medi=
a<br>
asset does not block small messages over a well designed multiplexing<br>
protocol, if it&#39;s also well implemented. =A0It does delay them a bit,<b=
r>
but the same happens with separate connections (with different<br>
charactistics in detail).<br><br>I think an important fact is sort of being=
 overlooked here. When you add something new to your toolset, the rest of y=
our tools don&#39;t disappear. If your application needs to transmit a file=
, you should consider using a protocol designed for that exact thing: HTTP.=
 Of course, you can send anything you want over your streaming connection, =
but if for your particular application some other tool is a better fit, don=
&#39;t be shy! Consider how XMPP decided to support buddy icons (<a href=3D=
"http://xmpp.org/extensions/xep-0084.html">http://xmpp.org/extensions/xep-0=
084.html</a>). They thought it was best to provide two options: you can eit=
her send the file over the wire, or simply transmit a url.<br>
<br>As Jamie points out, another solution involves separate connections, wh=
ich rather makes sense, as game logic and images will very likely be on sep=
arate servers.<br><br>No one here is proposing eliminating HTTP. Ideally, t=
hese discussions will produce something that perfectly complements it - tha=
t fills in where HTTP is lacking without too much overlap. That said, I cer=
tainly don&#39;t mean to suggest that HTTP has no place in this conversatio=
n (WebSocket, after all, uses HTTP conventions to establish a connection). =
I simply think that every feature of HTTP doesn&#39;t necessarily belong in=
 the protocol we choose to sit beside it, and that the various features of =
HTTP will continue to work, even if we settle on Streaming Pig Latin Socket=
s.<br>
<br>As Frank pointed out, giving developers a flexible tool is a GOOD IDEA,=
 and if we don&#39;t, they&#39;ll just end up hacking out the features they=
 would have implemented reasonably given something more general (see: Comet=
).<br>
<br>Mario Balibrera<br><br><div class=3D"gmail_quote">On Thu, Apr 16, 2009 =
at 3:12 PM, Jamie Lokier <span dir=3D"ltr">&lt;<a href=3D"mailto:jamie@shar=
eable.org">jamie@shareable.org</a>&gt;</span> wrote:<br><blockquote class=
=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin=
: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Greg Wilkins wrote:<br>
<div class=3D"im">&gt; Ian Hickson wrote:<br>
&gt; So if you want to have a bidirection full-duplex protocol, then the<br=
>
&gt; chances are the connection will be created by a HTTP client talking<br=
>
&gt; to a HTTP server and then doing an upgrade.<br>
&gt;<br>
&gt; For that upgrade to be successful, we are going to need to engage<br>
&gt; all the firewall/gateway/proxy/cache venders and developers so that<br=
>
&gt; they will support upgrade either to a specific protocol or to<br>
&gt; a class of protocols.<br>
<br>
</div>That&#39;s one way, but there are other more current-proxy-compatible=
 ways<br>
to do it. =A0They are hacks, but work.<br>
<font color=3D"#888888"><br>
-- Jamie<br>
</font><div><div></div><div class=3D"h5">__________________________________=
_____________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--0016362850fe56b5c20467b46f29--

From gregw@webtide.com  Thu Apr 16 22:50:30 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CB16A28B23E for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 22:50:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.553
X-Spam-Level: 
X-Spam-Status: No, score=-2.553 tagged_above=-999 required=5 tests=[AWL=0.046,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id SHQb2PI7mLeE for <hybi@core3.amsl.com>; Thu, 16 Apr 2009 22:50:27 -0700 (PDT)
Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.179]) by core3.amsl.com (Postfix) with ESMTP id E5B8B3A69A9 for <hybi@ietf.org>; Thu, 16 Apr 2009 22:50:26 -0700 (PDT)
Received: by wa-out-1112.google.com with SMTP id l35so369865waf.5 for <hybi@ietf.org>; Thu, 16 Apr 2009 22:51:40 -0700 (PDT)
Received: by 10.114.192.3 with SMTP id p3mr1200326waf.166.1239947499763; Thu, 16 Apr 2009 22:51:39 -0700 (PDT)
Received: from ?10.10.1.12? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id q20sm9159231pog.17.2009.04.16.22.51.36 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 16 Apr 2009 22:51:38 -0700 (PDT)
Message-ID: <49E818E2.8060701@webtide.com>
Date: Fri, 17 Apr 2009 15:51:30 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com> <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com> <49E6E042.8020006@webtide.com> <63df84f0904161053l42ec666l91422cde418033ad@mail.gmail.com>
In-Reply-To: <63df84f0904161053l42ec666l91422cde418033ad@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 17 Apr 2009 05:50:30 -0000

Jonas Sicking wrote:

> I believe the latest version can be found at
> http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol
> 
> As I understand it feedback is supposed to go to the list we're
> currently talking on.


Great!


Then I reiterate my last feedback - the spec is
written is a way that is very difficult to understand.

I'm sorry to be critical of the document - as I know it
is exceedingly difficult to write good specifications
that are both understandable and precise, while not
be too prescriptive of things like ordering and algorithms.


The spec really needs to have the key concepts described
in introductory sections rather than by pseudo code
repeated several times in the body of the document.


For example, if I want to find out what the frame type byte
is, I have to go to section 3.2   where I learn that
it is the first byte read when reading bytes from the
server.

But to know what that byte means, I then need to step
through the pseudo code of what to do if high bit is
set and then mentally run the iterations in order to
work out what it means.

I think have to read the same pseudo code in section 4.2
to see if the server implements the same state
machine(s) and thus has the same framing.  I'm still not
sure if the framing of 3.2 and 4.2 are exactly the
same?


It would have been so much simpler just to say:


3 Protocol Parameters

   ...

   3.x Frame Type

   The first byte of any websocket message is the frame type,
   which is used to indicate both the content type and framing
   of the messasge. Frame types 0x00 to 0x7f (high-order bit clear)
   represent sentinel framed data. Frame type 0x00 to 0x7f (high-order
   bit set), represent length encoded framing.

   3.x.1 Known frame types

   0x00  sentinel framed UTF-8 data
    ...   ????
   0x80  length encoded framed UTF-8 data


4 Web Socket Message

   4.1 Sentinel Framing
   Blah blah blah 0x00 blah blah blah 0xFF

   4.2 Length Encoded Framing
   First bytes represent the message length
   encoded as a variable length integer.  Each byte contributes
   7 bits to the length and the first byte with the high-order bit
   set is the last of the length bytes.







Another problem is that the handshake is defined by
explicitly saying what bytes to send: eg:

    5.   Send the following bytes to the remote side (the server):
            47 45 54 20
         Send the /resource name/ value, encoded as US-ASCII.

         Send the following bytes:

            20 48 54 54 50 2f 31 2e  31 0d 0a 55 70 67 72 61
            64 65 3a 20 57 65 62 53  6f 63 6b 65 74 0d 0a 43
            6f 6e 6e 65 63 74 69 6f  6e 3a 20 55 70 67 72 61
            64 65 0d 0a

         NOTE: The string "GET ", the path, " HTTP/1.1", CRLF, the string
         "Upgrade: WebSocket", CRLF, and the string "Connection:
          Upgrade", CRLF.

     6.  Send the following bytes:
            48 6f 73 74 3a 20
         Send the /host/ value, encoded as US-ASCII.

         Send the following bytes:
            0d 0a
         NOTE: The string "Host: ", the host, and CRLF.

    7.   Send the following bytes:
            4f 72 69 67 69 6e 3a 20
         Send the /origin/ value, encoded as US-ASCII.

         NOTE: The /origin/ value is a string that was passed to this algorithm.

         Send the following bytes:
            0d 0a
         NOTE: The string "Origin: ", the origin, and CRLF.


That's a really long winded way of saying send a HTTP/1.1
request with some headers set.   More over, that specification
implies that the order has to be:

  GET /path HTTP/1.1
  Upgrade: WebSocket
  Connection: Upgrade
  Host: acme.com
  Origin: origin

and thus it would be illegal to send:

  GET /path HTTP/1.1
  Origin:                origin
  Host:
    acme.com:80
  Upgrade:                 WebSocket
  Connection:      Upgrade


This is a perfectly legal HTTP request and should
be able to be accepted to initialize a websocket
upgrade.




regards

































From David.Linner9ab33xy531fokus.fraunhofer.de@bounce.antispameurope.de  Mon Apr 20 02:32:55 2009
Return-Path: <David.Linner9ab33xy531fokus.fraunhofer.de@bounce.antispameurope.de>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C6D3728C0EA for <hybi@core3.amsl.com>; Mon, 20 Apr 2009 02:32:55 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.764
X-Spam-Level: 
X-Spam-Status: No, score=-2.764 tagged_above=-999 required=5 tests=[AWL=-0.654, BAYES_05=-1.11, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gUnNCvMB5RPk for <hybi@core3.amsl.com>; Mon, 20 Apr 2009 02:32:54 -0700 (PDT)
Received: from relay01-haj2.antispameurope.com (relay01-haj2.antispameurope.com [83.246.65.51]) by core3.amsl.com (Postfix) with ESMTP id 8BE8E28C0E5 for <hybi@ietf.org>; Mon, 20 Apr 2009 02:32:53 -0700 (PDT)
Received: by relay01-haj2.antispameurope.com (ASE-Secure-MTA, from userid 1000) id C60D9940AB; Mon, 20 Apr 2009 11:34:07 +0200 (CEST)
Received: from pluto.fokus.fraunhofer.de (pluto.fokus.fraunhofer.de [195.37.77.164]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by relay01-haj2.antispameurope.com (ASE-Secure-MTA) with ESMTP id 56FA7940AB; Mon, 20 Apr 2009 11:34:07 +0200 (CEST)
Received: from EXCHSRV.fokus.fraunhofer.de (bohr [10.147.9.231]) by pluto.fokus.fraunhofer.de (8.13.7/8.13.7) with SMTP id n3K9Y7BF018224; Mon, 20 Apr 2009 11:34:07 +0200 (MEST)
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Date: Mon, 20 Apr 2009 11:34:01 +0200
Message-ID: <79C0EA6E7AD7CE4A85EDAF482B5456B202173E5E@EXCHSRV.fokus.fraunhofer.de>
In-Reply-To: <20090411033345.GD9817@voodoowarez.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Re: [hybi] Apples and Orangutans
Thread-Index: Acm6bp1ZOvZXKcN/R5OSdPlrThjWygHKUL0A
References: <1E5B078C-3C9F-4934-A96E-56E57879A31D@lindenlab.com><1cb725390904081302k36614a1cq5715bc6afb96c7a2@mail.gmail.com><49DD6B77.50104@it.aoyama.ac.jp> <20090411033345.GD9817@voodoowarez.com>
From: "Linner, David" <David.Linner@fokus.fraunhofer.de>
To: "rektide" <rektide@voodoowarez.com>, "Jamie Lokier" <jamie@shareable.org>,  "S. Mike Dierken" <mike@dierken.com>
Cc: hybi@ietf.org
Subject: Re: [hybi] Apples and Orangutans
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 20 Apr 2009 09:32:55 -0000

On Sat, Apr 11, 2009 rektide wrote:
> This seems like something that would be ideal to layer on top of a
> HyBi/"Server Initiated
> HTTP" transport.  It doesnt sound like something that everyone using
> duplex HTTP would want.
> Whether or not the conversation about this possible protocol belongs
> here, I truly do not
> know, but I want to call it out right now as something I see as
> decidedly "not transport",
> which was the question I interpretted as lying at the heart of Apples
> and Orangutans.
I don't agree. Some application requirements are better answered on
transport level. You can realize something like Martin's "synchronize
this part of this page (i.e. DOM) with this part of that page over at
that client" with HTTP Channels on top of a bidirectional message
protocol over TCP. But how many clients can be served? What latency do
users experience?=20

If you have a use case such as real-time state sharing (e.g. multiplayer
games) you need to care about fast handling of concurrent writes to the
same resource by several users, but you perhaps don't have to care about
lost packets with updates too much.
For example we implemented a round based client-server protocol for Web
state sharing on UDP. The server manages the actual states of all
resources and moderates the synchronization rounds. A round has a length
of 100-300ms and comprises three phases, i) call clients to instantly
send updates, ii) wait a predefined time (>2*network delay) and collect
update requests from all clients, iii) apply updates, resolve conflicts,
send difference to last state to all clients (and kick-off next round).
The communication is client initiated; the short round time ensures NATs
are kept open. Packet loss means that contained updates are communicated
one round later if still required. Clients always only get the
difference of their state to the current state at the server (no already
overwritten states). Tests with a Firefox extension and some real-time
Web applications went very well.

Protocols like the above are not suitable for bi-directional
communication of messages with arbitrary length, but they meet pretty
well the requirements of a real-time game or an application for
real-time collaborative work. If the WebSocket aims at such use cases, I
think it is at least worth to discuss if some protocol concepts or
parameters can be included with the specs discussed on this mailing list
to explicitly support state sharing on transport level. Our experiments
clearly showed that any protocol semantic coded in JavaScript and
interpreted at runtime affects the communication performance.

On Thu, Apr 16, 2009 Jamie Lokier wrote
> S. Mike Dierken wrote:
> > If I were writing a multi-player game, I'd definitely want to
> identify
> > portions of game state and media assets with something like a URI,
> have
> > various standardized content types, consistent methods (uniform
> interfaces),
> > etc. So HTTP concepts sound helpful.
>=20
> I agree.  But I know some people don't want those concepts; to them,
> they are overhead.  I'm leaning towards a protocol which makes the
> concepts easily available (and I think most people will use them), but
> makes them optional.
That would be a good starting point.=20

Regards,
David

p.s. Sorry for the late reply, I didn't find time last week.=20

From gregw@webtide.com  Thu Apr 23 17:42:26 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4A4B928C6C0 for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 17:42:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.554
X-Spam-Level: 
X-Spam-Status: No, score=-2.554 tagged_above=-999 required=5 tests=[AWL=0.045,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2MelmubygRVi for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 17:42:25 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.28]) by core3.amsl.com (Postfix) with ESMTP id C4BBF28C545 for <hybi@ietf.org>; Thu, 23 Apr 2009 17:42:24 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 3so658653ywj.49 for <hybi@ietf.org>; Thu, 23 Apr 2009 17:43:42 -0700 (PDT)
Received: by 10.90.97.16 with SMTP id u16mr1882591agb.105.1240533822121; Thu, 23 Apr 2009 17:43:42 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 29sm1214806agd.41.2009.04.23.17.43.39 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 23 Apr 2009 17:43:41 -0700 (PDT)
Message-ID: <49F10B37.3020904@webtide.com>
Date: Fri, 24 Apr 2009 10:43:35 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <49DEF171.4080506@mozilla.com> <c5b3a7130904100019ged7ec45o749213482f12c878@mail.gmail.com> <op.usaa24vt64w2qv@anne-van-kesterens-macbook.local> <170C0E82-633A-4CA9-B26E-38015BFC7C0D@lindenlab.com> <49E3C476.7090005@webtide.com> <20090414191717.GI32311@shareable.org> <BB63DF55-2385-4995-A856-E1EAD67F2D4A@lindenlab.com> <63df84f0904151934n19b65b1bm4188667eac91287e@mail.gmail.com> <49E6E042.8020006@webtide.com> <63df84f0904161053l42ec666l91422cde418033ad@mail.gmail.com> <49E818E2.8060701@webtide.com>
In-Reply-To: <49E818E2.8060701@webtide.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [hybi] Websocket feedback was: Addressing Minimal Events
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 00:42:26 -0000

Jonas, Ian,

just to continue on expanding my feedback on websocket


It would be great if the websocket proposal could include
standard definitions for mime encoded datagrams.

Current frame types are:

   0x00  - sentinel framed UTF-8 message
   0x80  - length framed binary data.

I'd like to see two additional frame types supported
by default:

   0x01  - sentinel framed UTF-8 encoded MIME message
   0x81  - length framed MIME message.


Both these data types would contain a data that commenced
with a standard mime header (RFC 2045).   The header is optional
and terminated by CR LF CR LF.  Thus these types have a minimal
overhead of 4 bytes.

For both these types, any Content-Length header will be
ignored and the length indicated by the websocket framing
minus the header length will be used.

For 0x01 types the content type is assumed to be "text/plain; charset=utf-8"
If a content type header is specified, it must be "text/????; charset=utf-8"

For 0x81 the content type is assumed to be application/octet-stream unless
otherwise indicated.


The websocket API would need to be slightly extended to support some
common types of message.

I would suggest that onmessage always be called for all text
mime types, but with some additional parameters: eg.

   onmessage(text,mimetype,headers)

The browser would be responsible for converting the transported
charset to the charset of javascript. If the conversion could not
be done, then the message would be discarded.

Additional events could be supported if you want the browser/server
to do the parsing for your.   For text/xml & text/html:

   ondocument(dom,headers)

and for text/json

   onobject(object,headers)


To send such messages, the API would also need to support

void postMessage(data,headers);



I think this is a minimal change to websocket and would go a long
way to address many of the concerns raised here.    With the ability
to send standardized meta data, then the job of coming up with
standardized multiplexing is much much simpler.



cheers



Greg Wilkins wrote:
> Jonas Sicking wrote:
> 
>> I believe the latest version can be found at
>> http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol
>>
>> As I understand it feedback is supposed to go to the list we're
>> currently talking on.
> 
> 
> Great!
> 
> 
> Then I reiterate my last feedback - the spec is
> written is a way that is very difficult to understand.
> 
> I'm sorry to be critical of the document - as I know it
> is exceedingly difficult to write good specifications
> that are both understandable and precise, while not
> be too prescriptive of things like ordering and algorithms.
> 
> 
> The spec really needs to have the key concepts described
> in introductory sections rather than by pseudo code
> repeated several times in the body of the document.
> 
> 
> For example, if I want to find out what the frame type byte
> is, I have to go to section 3.2   where I learn that
> it is the first byte read when reading bytes from the
> server.
> 
> But to know what that byte means, I then need to step
> through the pseudo code of what to do if high bit is
> set and then mentally run the iterations in order to
> work out what it means.
> 
> I think have to read the same pseudo code in section 4.2
> to see if the server implements the same state
> machine(s) and thus has the same framing.  I'm still not
> sure if the framing of 3.2 and 4.2 are exactly the
> same?
> 
> 
> It would have been so much simpler just to say:
> 
> 
> 3 Protocol Parameters
> 
>   ...
> 
>   3.x Frame Type
> 
>   The first byte of any websocket message is the frame type,
>   which is used to indicate both the content type and framing
>   of the messasge. Frame types 0x00 to 0x7f (high-order bit clear)
>   represent sentinel framed data. Frame type 0x00 to 0x7f (high-order
>   bit set), represent length encoded framing.
> 
>   3.x.1 Known frame types
> 
>   0x00  sentinel framed UTF-8 data
>    ...   ????
>   0x80  length encoded framed UTF-8 data
> 
> 
> 4 Web Socket Message
> 
>   4.1 Sentinel Framing
>   Blah blah blah 0x00 blah blah blah 0xFF
> 
>   4.2 Length Encoded Framing
>   First bytes represent the message length
>   encoded as a variable length integer.  Each byte contributes
>   7 bits to the length and the first byte with the high-order bit
>   set is the last of the length bytes.
> 
> 
> 
> 
> 
> 
> 
> Another problem is that the handshake is defined by
> explicitly saying what bytes to send: eg:
> 
>    5.   Send the following bytes to the remote side (the server):
>            47 45 54 20
>         Send the /resource name/ value, encoded as US-ASCII.
> 
>         Send the following bytes:
> 
>            20 48 54 54 50 2f 31 2e  31 0d 0a 55 70 67 72 61
>            64 65 3a 20 57 65 62 53  6f 63 6b 65 74 0d 0a 43
>            6f 6e 6e 65 63 74 69 6f  6e 3a 20 55 70 67 72 61
>            64 65 0d 0a
> 
>         NOTE: The string "GET ", the path, " HTTP/1.1", CRLF, the string
>         "Upgrade: WebSocket", CRLF, and the string "Connection:
>          Upgrade", CRLF.
> 
>     6.  Send the following bytes:
>            48 6f 73 74 3a 20
>         Send the /host/ value, encoded as US-ASCII.
> 
>         Send the following bytes:
>            0d 0a
>         NOTE: The string "Host: ", the host, and CRLF.
> 
>    7.   Send the following bytes:
>            4f 72 69 67 69 6e 3a 20
>         Send the /origin/ value, encoded as US-ASCII.
> 
>         NOTE: The /origin/ value is a string that was passed to this 
> algorithm.
> 
>         Send the following bytes:
>            0d 0a
>         NOTE: The string "Origin: ", the origin, and CRLF.
> 
> 
> That's a really long winded way of saying send a HTTP/1.1
> request with some headers set.   More over, that specification
> implies that the order has to be:
> 
>  GET /path HTTP/1.1
>  Upgrade: WebSocket
>  Connection: Upgrade
>  Host: acme.com
>  Origin: origin
> 
> and thus it would be illegal to send:
> 
>  GET /path HTTP/1.1
>  Origin:                origin
>  Host:
>    acme.com:80
>  Upgrade:                 WebSocket
>  Connection:      Upgrade
> 
> 
> This is a perfectly legal HTTP request and should
> be able to be accepted to initialize a websocket
> upgrade.
> 
> 
> 
> 
> regards
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 


From ian@hixie.ch  Thu Apr 23 23:10:08 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BD6383A6E35 for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 23:10:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.257
X-Spam-Level: 
X-Spam-Status: No, score=-3.257 tagged_above=-999 required=5 tests=[AWL=-0.658, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xwlPrrhuamUA for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 23:10:08 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id F3B763A6D76 for <hybi@ietf.org>; Thu, 23 Apr 2009 23:10:07 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 3461C12241A; Thu, 23 Apr 2009 23:11:26 -0700 (PDT)
Date: Fri, 24 Apr 2009 06:11:25 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>  <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>  <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 06:10:08 -0000

On Mon, 13 Apr 2009, Paul Prescod wrote:
> 
> Please be very specific about HOW a 12-year old who knows PHP would
> write a multiplayer game using PHP and WebSockets.

I don't know about a 12-year old who knows PHP, but a 12-year old who 
knows, say, Perl, could do it something like this for the server side:

   http://www.damowmow.com/playground/demos/websocket/tic-tac-toe.pl

...and something like this for the client side:

   http://www.damowmow.com/playground/demos/websocket/tic-tac-toe.html

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From Martin.Thomson@andrew.com  Thu Apr 23 23:44:58 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3DF9C3A6FE8 for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 23:44:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.234
X-Spam-Level: 
X-Spam-Status: No, score=-2.234 tagged_above=-999 required=5 tests=[AWL=-0.235, BAYES_00=-2.599, J_CHICKENPOX_39=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Pq3lg0fh8EPJ for <hybi@core3.amsl.com>; Thu, 23 Apr 2009 23:44:57 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 1D3083A69FA for <hybi@ietf.org>; Thu, 23 Apr 2009 23:44:57 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_24_02_06_55
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from acdcexbh1.andrew.com [10.86.20.91] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Fri, 24 Apr 2009 02:06:55 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by acdcexbh1.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Fri, 24 Apr 2009 01:46:14 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Fri, 24 Apr 2009 01:38:07 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com>
In-Reply-To: <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] Is there a traffic jam?
Thread-Index: AcnEo4HArUWzFy6rT7W2HhvQFmXuQgAAsBIA
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com><1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Ian Hickson" <ian@hixie.ch>, "Paul Prescod" <paul@prescod.net>
X-OriginalArrivalTime: 24 Apr 2009 06:46:14.0594 (UTC) FILETIME=[5CB8E220:01C9C4A8]
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 06:44:58 -0000
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From ian@hixie.ch  Fri Apr 24 00:01:09 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 239FA28C6EA for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:01:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.245
X-Spam-Level: 
X-Spam-Status: No, score=-3.245 tagged_above=-999 required=5 tests=[AWL=-0.646, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 8kLUvEkWpU82 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:01:08 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 3DBF328C71F for <hybi@ietf.org>; Fri, 24 Apr 2009 00:01:08 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 5E79B27B7D; Fri, 24 Apr 2009 00:02:25 -0700 (PDT)
Date: Fri, 24 Apr 2009 07:02:25 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com>
Message-ID: <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C46E@AHQEX1.andrew.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com><1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 07:01:09 -0000

On Fri, 24 Apr 2009, Thomson, Martin wrote:
>
> Because all 12 year olds use strict and -T.

I was using their equivalents in the language(s) I used when I was 12. 
It's good hygiene. But it doesn't really matter either way for this code.


> I must admit that my perl is getting a wee bit rusty, but that's a 
> non-trivial piece of code.

Come now. It's barely 100 lines of code, just doing basic network I/O. My 
friends and I were writing far more complex code at 12, and I'm hardly a 
good programmer. Most of the constructs are copied straight from online 
tutorials on Perl networking.


> I hope that you are about to argue that most of that code is going to be 
> provided by Net::WebSocket or something similar.

It seems reasonable for such a library to exist, but I don't think it is 
reasonable to expect that libraries will exist, no. That's the point. You 
don't need a library. It's no more than a couple of hours' worth of coding 
for a novice like me (I'm no professional programmer).


> Otherwise, there's more than enough rope left there to hang a number of 
> software engineers that I know (admittedly, not the better ones).

The same could be said of "Hello World" in most languages, especially 
Perl. Just think how much more complex this example would be if the 
protocol was more than just sentinel-delimited frames. :-)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Fri Apr 24 00:12:29 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 62A983A69FF for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:12:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.976
X-Spam-Level: 
X-Spam-Status: No, score=-1.976 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id C5iG-PMeBFnk for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:12:28 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.29]) by core3.amsl.com (Postfix) with ESMTP id EAD333A6AF0 for <hybi@ietf.org>; Fri, 24 Apr 2009 00:11:02 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 3so821126ywj.49 for <hybi@ietf.org>; Fri, 24 Apr 2009 00:12:21 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=n6h0QbNj2AIob97g6E2I++ix1PTmFPV8GKB852vDfJ8=; b=iSku8cgqHSGAbFxa9mvmqg90PmWyuwoKvcEzyabM9r8mDAt1W26HnwsXJnbYHuN6Lv ji6IJkwJNsZM2efilNYOmqB0wnW7VsUvUXpcEpwdIOxR4Vz3152/gstgOJfSEVkNA+Xh efv95G6CLhqdYjNQtSZMnEg5bY/nwd5rm9faE=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=d8ugygDfIT/Cav0Q0m5Re5mY8R333qFS/CB8GZg30SKCtIlRpLajrOU4O0fI/17BYm KzgOzej5F+U2yQUJIho1TacEm2EPsfPhUXLMG0sL9c1g2UBQCC8hJq/ZETnX1ljPE+5B vuAwwEAGR7+OAGKCVL1pMVeOrQCOj0B+JR1e0=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.100.151.8 with SMTP id y8mr2743479and.94.1240557140271; Fri,  24 Apr 2009 00:12:20 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>
Date: Fri, 24 Apr 2009 00:12:20 -0700
X-Google-Sender-Auth: 0b8d30f2d736bb63
Message-ID: <1cb725390904240012q781391a1v924cbc1349952499@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=001485f94562ab64be046847b742
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 07:12:29 -0000

--001485f94562ab64be046847b742
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

A few thoughts:

 1. It boggles my mind that you think that that is the kind of code that
"novices" will write. This is my main point. Everything that follows is of
lesser emphasis.

 2. I do not understand why you wish to inflict this kind of low-level Unix
hacking on novices. Even presuming that they can learn it from "WebSockets
for dummies", they will likely just skip the explanation and type it in as
cargo cult software.

 3. What happens when your tic-tac-toe server gets popular? Does this code
scale?

 4. Proper software engineering would suggest that you should split the
reusable bits from the application logic. Having done this, you will have a
"WebSockets library". At that point, you will never need to worry about 1
through 3 again, and neither will anyone else. But the problem is, once
you've done this, you've undone the very reason that you made the protocol
so simple-seeming in the first place. So you, Ian Hickson probably won't do
that. But someone will do it and no Perl programmer will ever care about the
details of how to parse WebSockets again. On the other hand, when they get
out their trusty TCPDUMP tool, they'll see this crud on the wire:

 0x48,0x54,0x54,0x50,0x2f,0x31,0x2e,0x31,
0x20,0x31,0x30,0x31,0x20,0x57,0x65,0x62,

0x20,0x53,0x6f,0x63,0x6b,0x65,0x74,0x20,
0x50,0x72,0x6f,0x74,0x6f,0x63,0x6f,0x6c,

0x20,0x48,0x61,0x6e,0x64,0x73,0x68,0x61,
0x6b,0x65,0x0d,0x0a,0x55,0x70,0x67,0x72,

0x61,0x64,0x65,0x3a,0x20,0x57,0x65,0x62,
0x53,0x6f,0x63,0x6b,0x65,0x74,0x0d,0x0a,

0x43,0x6f,0x6e,0x6e,0x65,0x63,0x74,0x69,
0x6f,0x6e,0x3a,0x20,0x55,0x70,0x67,0x72,
                                             0x61,0x64,0x65,0x0d,0x0a


In other words, I strongly believe that you are optimizing for the wrong
thing. Slightly easier parsing is not as important as readability of the
protocol, in my opinion. I cannot tell you how many times in a month I do
this:

# telnet localhost 11211
Trying 127.0.0.1...
Connected to localhost.localdomain (127.0.0.1).
stats
STAT pid 2202
STAT uptime 7567475
...

(that's memcached -- but I've never written a memcached client)

# telnet localhost 80
Trying 127.0.0.1...
Connected to localhost.localdomain (127.0.0.1).
Escape character is '^]'.
GET / HTTP/1.1
Host: www.slate.com

# telnet localhost 22
Trying 127.0.0.1...
Connected to localhost.localdomain (127.0.0.1).
Escape character is '^]'.
SSH-2.0-OpenSSH_4.3

Protocol mismatch.
Connection closed by foreign host.

Even SSH at least gives me a readable introductory string to tell me what
protocol is running and that the server is alive and kicking.


On Thu, Apr 23, 2009 at 11:11 PM, Ian Hickson <ian@hixie.ch> wrote:

> On Mon, 13 Apr 2009, Paul Prescod wrote:
> >
> > Please be very specific about HOW a 12-year old who knows PHP would
> > write a multiplayer game using PHP and WebSockets.
>
> I don't know about a 12-year old who knows PHP, but a 12-year old who
> knows, say, Perl, could do it something like this for the server side:
>
>   http://www.damowmow.com/playground/demos/websocket/tic-tac-toe.pl
>
> ...and something like this for the client side:
>
>   http://www.damowmow.com/playground/demos/websocket/tic-tac-toe.html
>
> --
> Ian Hickson               U+1047E                )\._.,--....,'``.    fL
> http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
> Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
>

--001485f94562ab64be046847b742
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

A few thoughts:<br><br>=C2=A01. It boggles my mind that you think that that=
 is the kind of code that &quot;novices&quot; will write. This is my main p=
oint. Everything that follows is of lesser emphasis.<br><br>=C2=A02. I do n=
ot understand why you wish to inflict this kind of low-level Unix hacking o=
n novices. Even presuming that they can learn it from &quot;WebSockets for =
dummies&quot;, they will likely just skip the explanation and type it in as=
 cargo cult software. <br>

<br>=C2=A03. What happens when your tic-tac-toe server gets popular? Does t=
his code scale?<br><br>=C2=A04. Proper software engineering would suggest t=
hat you should split the reusable bits from the application logic. Having d=
one this, you will have a &quot;WebSockets library&quot;. At that point, yo=
u will never need to worry about 1 through 3 again, and neither will anyone=
 else. But the problem is, once you&#39;ve done this, you&#39;ve undone the=
 very reason that you made the protocol so simple-seeming in the first plac=
e. So you, Ian Hickson probably won&#39;t do that. But someone will do it a=
nd no Perl programmer will ever care about the details of how to parse WebS=
ockets again. On the other hand, when they get out their trusty TCPDUMP too=
l, they&#39;ll see this crud on the wire:<br>
<br><pre> 0x48,0x54,0x54,0x50,0x2f,0x31,0x2e,0x31, 0x20,0x31,0x30,0x31,0x20=
,0x57,0x65,0x62,<br>                                             0x20,0x53,=
0x6f,0x63,0x6b,0x65,0x74,0x20, 0x50,0x72,0x6f,0x74,0x6f,0x63,0x6f,0x6c,<br>
                                             0x20,0x48,0x61,0x6e,0x64,0x73,=
0x68,0x61, 0x6b,0x65,0x0d,0x0a,0x55,0x70,0x67,0x72,<br>                    =
                         0x61,0x64,0x65,0x3a,0x20,0x57,0x65,0x62, 0x53,0x6f=
,0x63,0x6b,0x65,0x74,0x0d,0x0a,<br>
                                             0x43,0x6f,0x6e,0x6e,0x65,0x63,=
0x74,0x69, 0x6f,0x6e,0x3a,0x20,0x55,0x70,0x67,0x72,<br>                    =
                         0x61,0x64,0x65,0x0d,0x0a</pre><br>
In other words, I strongly believe that you are optimizing for the wrong th=
ing. Slightly easier parsing is not as important as readability of the prot=
ocol, in my opinion. I cannot tell you how many times in a month I do this:=
<br>
<br># telnet localhost 11211<br>Trying 127.0.0.1...<br>Connected to localho=
st.localdomain (127.0.0.1).<br>stats<br>STAT pid 2202<br>STAT uptime 756747=
5<br>...<br><br>(that&#39;s memcached -- but I&#39;ve never written a memca=
ched client)<br>
<br># telnet localhost 80<br>Trying 127.0.0.1...<br>Connected to localhost.=
localdomain (127.0.0.1).<br>Escape character is &#39;^]&#39;.<br>GET / HTTP=
/1.1<br>Host: <a href=3D"http://www.slate.com">www.slate.com</a><br><br># t=
elnet localhost 22<br>
Trying 127.0.0.1...<br>Connected to localhost.localdomain (127.0.0.1).<br>E=
scape character is &#39;^]&#39;.<br>SSH-2.0-OpenSSH_4.3<br><br>Protocol mis=
match.<br>Connection closed by foreign host.<br><br>Even SSH at least gives=
 me a readable introductory string to tell me what protocol is running and =
that the server is alive and kicking.<br>
<br><br><div class=3D"gmail_quote">On Thu, Apr 23, 2009 at 11:11 PM, Ian Hi=
ckson <span dir=3D"ltr">&lt;ian@hixie.ch&gt;</span> wrote:<br><blockquote c=
lass=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); ma=
rgin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">On Mon, 13 Apr 2009, Paul Prescod wrote:<br>
&gt;<br>
</div><div class=3D"im">&gt; Please be very specific about HOW a 12-year ol=
d who knows PHP would<br>
&gt; write a multiplayer game using PHP and WebSockets.<br>
<br>
</div>I don&#39;t know about a 12-year old who knows PHP, but a 12-year old=
 who<br>
knows, say, Perl, could do it something like this for the server side:<br>
<br>
 =C2=A0 <a href=3D"http://www.damowmow.com/playground/demos/websocket/tic-t=
ac-toe.pl" target=3D"_blank">http://www.damowmow.com/playground/demos/webso=
cket/tic-tac-toe.pl</a><br>
<br>
...and something like this for the client side:<br>
<br>
 =C2=A0 <a href=3D"http://www.damowmow.com/playground/demos/websocket/tic-t=
ac-toe.html" target=3D"_blank">http://www.damowmow.com/playground/demos/web=
socket/tic-tac-toe.html</a><br>
<div><div></div><div class=3D"h5"><br>
--<br>
Ian Hickson =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 U+1047E =C2=A0=
 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0)\._.,--....,&#39;``. =C2=
=A0 =C2=A0fL<br>
<a href=3D"http://ln.hixie.ch/" target=3D"_blank">http://ln.hixie.ch/</a> =
=C2=A0 =C2=A0 =C2=A0 U+263A =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0/, =C2=A0 _.. \ =C2=A0 _\ =C2=A0;`._ ,.<br>
Things that are impossible just take longer. =C2=A0 `._.-(,_..&#39;--(,_..&=
#39;`-.;.&#39;<br>
</div></div></blockquote></div><br>

--001485f94562ab64be046847b742--

From prescod@gmail.com  Fri Apr 24 00:17:13 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E9CF328C689 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:17:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.976
X-Spam-Level: 
X-Spam-Status: No, score=-1.976 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RzgssGdUnHeU for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 00:17:12 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.31]) by core3.amsl.com (Postfix) with ESMTP id 743DE28C718 for <hybi@ietf.org>; Fri, 24 Apr 2009 00:16:42 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 3so823143ywj.49 for <hybi@ietf.org>; Fri, 24 Apr 2009 00:18:00 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=88QjfOflOn9bCja0HsQdxrE1CaHilJYyWlVfgHQNxXM=; b=a/Z/HOElDnRvb6252MGd1Xtkz7ICh4iyF2j4aIQRYtSyZnQcUdNMf+n2z5/hyTxauN c4P6grw+UMSjaiAZOjXYrZ4e6hFcPyvMLmT4g+Ckxtz0UOVzJtKhg9TIRTTaz7EP4OeI WnbSduwTNwLb+BmqQQwAlqzKHdX9zJt4r0xKk=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=b5pF8xN93I+TR13udmEcOeSrgx7QwhYDfefAYcfaCo+2FEjJ7EodwFW1qSy9qWcwUF SZ9aI6+8ODyrUXEjOUgIA5x3kmiDT9rOVoaFW966EAea8ufhYCcE7poNuo/VjhhJiFgh tnxKg6rh1ZgcjMgTbEpybz0/2CIa3ALMTirXE=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.100.8.17 with SMTP id 17mr2707205anh.130.1240557480510; Fri,  24 Apr 2009 00:18:00 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com>
Date: Fri, 24 Apr 2009 00:18:00 -0700
X-Google-Sender-Auth: d1358613dc840e42
Message-ID: <1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=001485f86f4ef3097f046847cb72
Cc: hybi@ietf.org, "Thomson, Martin" <Martin.Thomson@andrew.com>
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 07:17:14 -0000

--001485f86f4ef3097f046847cb72
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On Fri, Apr 24, 2009 at 12:02 AM, Ian Hickson <ian@hixie.ch> wrote:


>
> > I hope that you are about to argue that most of that code is going to be
> > provided by Net::WebSocket or something similar.
>
> It seems reasonable for such a library to exist, but I don't think it is
> reasonable to expect that libraries will exist, no.


WHY NOT????

Libraries for UTF-8 exist. Libraries for XML exist. Libraries for EDI exist.
Libraries for SGML exist. Libraries for TCP sockets exist. Libraries for
every important standard under the sun EXIST. Why will WebSockets be a
unique and special case?

Why wouldn't libraries for Net::WebSocket exist? What kind of bizarro
universe do you anticipate where people prefer to go hunt down bits of code
from tutorials AND like to read protocol specs AND like to implement them
*themselves*, rather than just install libraries from CPAN? Are these
novices masochists?

 Paul Prescod

--001485f86f4ef3097f046847cb72
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Fri, Apr 24, 2009 at 12:02 AM, Ian Hickson <span dir=3D"ltr">&lt;ian@hix=
ie.ch&gt;</span> wrote:<br><div class=3D"gmail_quote"><br><blockquote class=
=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin=
: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im"><br>
<br>
&gt; I hope that you are about to argue that most of that code is going to =
be<br>
&gt; provided by Net::WebSocket or something similar.<br>
<br>
</div>It seems reasonable for such a library to exist, but I don&#39;t thin=
k it is<br>
reasonable to expect that libraries will exist, no. </blockquote><div><br>W=
HY NOT????<br><br>Libraries for UTF-8 exist. Libraries for XML exist. Libra=
ries for EDI exist. Libraries for SGML exist. Libraries for TCP sockets exi=
st. Libraries for every important standard under the sun EXIST. Why will We=
bSockets be a unique and special case?<br>
<br>Why wouldn&#39;t libraries for Net::WebSocket exist? What kind of bizar=
ro universe do you anticipate where people prefer to go hunt down bits of c=
ode from tutorials AND like to read protocol specs AND like to implement th=
em *themselves*, rather than just install libraries from CPAN? Are these no=
vices masochists?<br>
<br>=C2=A0Paul Prescod<br><br></div></div>

--001485f86f4ef3097f046847cb72--

From ian@hixie.ch  Fri Apr 24 01:02:51 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 28AC83A69C2 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 01:02:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.233
X-Spam-Level: 
X-Spam-Status: No, score=-3.233 tagged_above=-999 required=5 tests=[AWL=-0.635, BAYES_00=-2.599, WEIRD_PORT=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fZo5Cra5qKp5 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 01:02:50 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id D55B23A67D0 for <hybi@ietf.org>; Fri, 24 Apr 2009 01:02:48 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 2B450125C66; Fri, 24 Apr 2009 01:04:07 -0700 (PDT)
Date: Fri, 24 Apr 2009 08:04:05 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904240012q781391a1v924cbc1349952499@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904240735290.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <Pine.LNX.4.62.0904132352430.10339@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105A0C476@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>  <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>  <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>  <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>  <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <1cb725390904240012q781391a1v924cbc1349952499@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 08:02:51 -0000

On Fri, 24 Apr 2009, Paul Prescod wrote:
> 
> 1. It boggles my mind that you think that that is the kind of code that 
> "novices" will write. This is my main point. Everything that follows is 
> of lesser emphasis.

I guess this is a matter of opinion. As I said in an earlier e-mail, most 
of the sample code was written by copying-and-pasting from tutorials.


> 2. I do not understand why you wish to inflict this kind of low-level 
> Unix hacking on novices. Even presuming that they can learn it from 
> "WebSockets for dummies", they will likely just skip the explanation and 
> type it in as cargo cult software.

Well, as you say, for people who want them, libraries will probably exist.


> 3. What happens when your tic-tac-toe server gets popular? Does this 
> code scale?

No, of course not. It's written in perl. Writing code for 10,000 QPS loads 
is a whole different ball game than writing a tic tac toe game for one's 
friends at school.


> 4. Proper software engineering would suggest that you should split the 
> reusable bits from the application logic. Having done this, you will 
> have a "WebSockets library". At that point, you will never need to worry 
> about 1 through 3 again, and neither will anyone else.

The same could be said for CGI, yet people still write their own CGI 
parsers. The same could be said for URLs, or XML, or JSON, or any number 
of other Web technologies, yet people write their own URL parsers, their 
own XML parsers (heaven forfend), their own JSON parsers, etc.


> # telnet localhost 80
> Trying 127.0.0.1...
> Connected to localhost.localdomain (127.0.0.1).
> Escape character is '^]'.
> GET / HTTP/1.1
> Host: www.slate.com
> 
> # telnet localhost 22
> Trying 127.0.0.1...
> Connected to localhost.localdomain (127.0.0.1).
> Escape character is '^]'.
> SSH-2.0-OpenSSH_4.3
> 
> Protocol mismatch.
> Connection closed by foreign host.
> 
> Even SSH at least gives me a readable introductory string to tell me 
> what protocol is running and that the server is alive and kicking.

$ telnet damowmow.com 11111
Trying 208.113.210.27...   
Connected to damowmow.com.
Escape character is '^]'. 
HTTP/1.1 101 Web Socket Protocol Handshake
Upgrade: WebSocket
Connection: Upgrade
WebSocket-Origin: http://damowmow.com
WebSocket-Location: ws://damowmow.com:11111/tic-tac-toe


I don't really understand the problem.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From ian@hixie.ch  Fri Apr 24 01:09:11 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1202F3A6A60 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 01:09:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.223
X-Spam-Level: 
X-Spam-Status: No, score=-3.223 tagged_above=-999 required=5 tests=[AWL=-0.624, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bP20zDPA9PWo for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 01:09:10 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 2F7AC3A68C1 for <hybi@ietf.org>; Fri, 24 Apr 2009 01:09:10 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id 00AD112241A; Fri, 24 Apr 2009 01:10:28 -0700 (PDT)
Date: Fri, 24 Apr 2009 08:10:28 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Paul Prescod <paul@prescod.net>
In-Reply-To: <1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com>  <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com>  <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>  <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>  <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>  <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>  <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>  <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>  <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com>  <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com> <1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 08:09:11 -0000

On Fri, 24 Apr 2009, Paul Prescod wrote:
> 
> Libraries for UTF-8 exist. Libraries for XML exist. Libraries for EDI exist.
> Libraries for SGML exist. Libraries for TCP sockets exist. Libraries for
> every important standard under the sun EXIST. Why will WebSockets be a
> unique and special case?

It won't. It'll be just like XML, and SGML, and JSON, and URIs, and CGI, 
and HTML, and ...: despite the existence of high-quality libraries, people 
will write their own, and the more complex the language, the more of a 
mess it makes. As the poster child for this I present HTML, which 
hypothetically was an SGML application. SGML libraries, as you say, exist, 
so why did HTML processors not use them? HTML is so far removed from what 
SGML rules say it should be that even HTML4 basically admits that calling 
it SGML is a lie. Why is that, given that there are SGML libraries?

Why do some people implement their own hacky XML parsers when they 
implement BEEP, instead of reusing one of the many XML parser libraries?

Why does everyone under the sun reimplement their own URI parser, despite 
libraries existing and despite URI parsing actually being pretty 
complicated to get right?


> Why wouldn't libraries for Net::WebSocket exist? What kind of bizarro 
> universe do you anticipate where people prefer to go hunt down bits of 
> code from tutorials AND like to read protocol specs AND like to 
> implement them *themselves*, rather than just install libraries from 
> CPAN? Are these novices masochists?

Why would people do it themselves for CGI, URIs, HTML, SGML, JSON, XML, 
etc, but _not_ WebSocket? Why would, as you say, WebSocket be a unique and 
special case?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From ian@hixie.ch  Fri Apr 24 02:50:58 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9F0963A72D1 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 02:50:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.213
X-Spam-Level: 
X-Spam-Status: No, score=-3.213 tagged_above=-999 required=5 tests=[AWL=-0.614, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id p22FKrFOg7i1 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 02:50:26 -0700 (PDT)
Received: from looneymail-a2.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 928DD3A7339 for <hybi@ietf.org>; Fri, 24 Apr 2009 02:50:25 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a2.g.dreamhost.com (Postfix) with ESMTP id 532C616D415 for <hybi@ietf.org>; Fri, 24 Apr 2009 02:51:44 -0700 (PDT)
Date: Fri, 24 Apr 2009 09:51:40 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: hybi@ietf.org
In-Reply-To: <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com>  <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com>  <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 09:50:58 -0000

Sorry about my absence from this list the past few days, I've been busy 
with HTML5 work. I put aside a few e-mails to reply to on the topic of 
WebSocket, which I reply to below:

On Mon, 6 Apr 2009, Peter Saint-Andre wrote:
> On 4/6/09 2:28 AM, Ian Hickson wrote:
> > What I'd like to work on is "TCP for the Web".
> 
> That's what we tried to define with BOSH. Whether we succeeded is an 
> open question.

I think reusing TCP's semantics (e.g. the tight coupling of the two 
directions of communication into one TCP channel) is important in making a 
"TCP for the Web" -- I don't think that a TCP analogue should map to 
multiple TCP connections underneath.


On Tue, 7 Apr 2009, Mridul Muralidharan wrote:
>
> Note that browser based javascript clients are not the only consumers of 
> bi-directional http; others like mobile clients, firewalled thick 
> clients, etc are also possible (and actually exist with BOSH for 
> example).
>
> So possibly targeting specific implementation technology might not be 
> appropriate for designing the spec ?

Why wouldn't mobile clients, firewalled thick clients, etc, just use a raw 
TCP socket?

The only reason we're not looking at raw TCP sockets when it comes to what 
to provide JavaScript Web apps access to is that TCP doesn't natively 
support the browser's security model. If it wasn't for that, raw TCP is 
exactly what we would have by now -- in fact the SVG WG even specced an 
API out for exactly that in SVG 1.2 (before the security problems with 
doing that were brought up).


On Tue, 7 Apr 2009, Sylvain Hellegouarch wrote:
> 
> In this context I can understand why your draft is so short and that's 
> probably a good sign because it'll be that easier to see deployed. 
> Nonetheless this long discussion makes me wonder whether or not having a 
> new protocol is a better idea than picking one of the existing 
> alternatives (Comet, BOSH). Wouldn't it be better to focus energy on 
> best practices when it comes to long polling connections based on what's 
> working now?

Adding something new doesn't prevent things that already work from 
working.


On Wed, 8 Apr 2009, Mark Lentczner wrote:
> 
> I see the attraction of just giving something very simple to application 
> developers. Giving them a protocol that does Minimal Events is easy to 
> explain and would have a simple API. However, I have seen that over 
> time, applications often need to build most of the framework that is 
> already provided by more complex abstractions. Where the cost of those 
> abstractions is small, and the ability to use them in simpler ways easy, 
> it is often better to supply a richer abstraction to all, since the 
> uniformity of the often implemented bits means better code for all.

That's what libraries are for. Core platforms (POSIX, Win32, Cocoa, the 
Web) provide fundamental building blocks, on top of which libraries are 
built that expose many different possible abstractions.

It's better that way, as it gives authors more choice as to how to do 
things. Someone might want a comprehensive protocol with all kinds of 
metadata and channels and reconnection support and so on; someone else 
might want a different set of abstractions based more on resources with 
identifiers and types. Why should we force one to use the protocol the 
other wanted? Better to provide an infrastructure on top of which both of 
these can be built.

Having said that, if people want higher-level protocols standardised, I've 
nothing against that.


> Example: Implementing a Minimal, Event protocol with Message, Req/Resp 
> is just a matter of accepting the overhead of the framing of the simple 
> data blob, and the cost of the minimal response message.

It also probably means dealing with a framework that's more than a hundred 
lines of Perl (say), which puts a lot of burder on the developer, and 
probably the inexperienced one at that.


> On the other hand, implementing a Message, Req/Resp. protocol with 
> Minimal, Events, means building a whole framework for encoding message 
> parts (address, meta-data and data) into the minimal data units 
> (strings? binary blobs?) and coming up with a way to correlate responses 
> to requests, etc... all in the application layer.

Right, but in this case it's likely the experienced developer.


On Thu, 9 Apr 2009, Peter Saint-Andre wrote:
> 
> The same-origin model is often invoked here. Is there a good definition 
> of it? Perhaps it would be helpful for someone to write a small I-D 
> about this.

The HTML5 spec defines it formally, but I don't know of a simple 
introduction to it.

Basically it consists of partitioning untrusted code based on tuples that 
consist of hostname or IP address, port, and protocol (scheme). Code is 
then limited to dealing with resources from the same origin (with some 
historical exceptions, like the ability to invoke script from any origin 
in the context of one's own origin).

Cross-origin mechanisms work on the principle of code from one origin 
opting into communication with code from another origin, with 
communication only being possible if both sides opt in.


On Thu, 9 Apr 2009, Linner, David wrote:
>
> Looking into the hybi wiki and the use cases listed there (e.g. 
> collaborative editing, multiplayer games) I get the feeling state 
> synchronization is quite important. So, I think a shared memory 
> interaction model would be suitable for these use cases. Implementing 
> such an interaction model on top of bidirectional messaging between 
> browser and server is not a problem, but a realization as JavaScript 
> library would result in a significant loss of performance.

In practice JavaScript performance has improved by orders of magnitude in 
the recent past and shows every sign of continuing to be the focus of 
serious performance work, so I don't think that JavaScript libraries are 
likely to be a performance bottleneck in the future.


On Fri, 10 Apr 2009, Greg Wilkins wrote:
> 
> However, IF we are in the game of picking winners or designing new 
> protocols, we could do a lot worse than starting with a close inspection 
> of BEEP and the concerns that they address.

BEEP was amongst the many protocols that I examined closely when writing 
the WebSocket proposal.


> The standard HTTP security model would be used to establish a connection 
> and then the upgrade is unable to break out of that connection etc.

What is the standard HTTP security model? WebSocket wasn't able to use it, 
nor was XHR2; in both cases we had to add an extra layer to get cross- 
origin restrictions into HTTP. Are we missing something?


On Mon, 13 Apr 2009, Jonas Sicking wrote:
> 
> But the point is that the type is transferred at a specified location. 
> And IMHO that specified location should be separated from other fields 
> with other purposes. My understanding is that the path is intended to 
> allow for several websocket services from the same origin, thus I would 
> prefer if the protocol is separate from the path.

I've added an optional protocol ID to WebSocket, which is sent along with 
the handshake and (if present) must be present in the response back from 
the server.


On Tue, 14 Apr 2009, Jamie Lokier wrote:
> 
> But it would be nice if the (imho highly desirable ;-) features from my 
> protocol were available to others without them needing yet another big 
> Javascript library in future browsers, and then only perhaps with one 
> web server and application backend.

Why would it be better for it to be a big bunch of new code in the 
browser, with perhaps only one web server and application backend, rather 
than a big bunch of new code in a JS library, with perhaps only one web 
server and application backend?


> But they don't look compelling for "I want to write an attractive web 
> application with lots of updating documents and tiny widgets without 
> using a pre-existing application framework or large bit of Javascript".

Writing an attractive web application with lots of updating documents and 
tiny widgets without using a large bit of Javascript seems optimistic.


On Wed, 15 Apr 2009, Greg Wilkins wrote:
> 
> while workers might make such sharing technically possible, I think 
> there is still an issue since the websocket API is both an application 
> API and a protocol API.  Developers who just want UTF-8 datagrams and 
> don't care about the transport will use the websocket API directly and 
> will thus inadvertently make a decision about the creating a connection 
> to the server!.

Well, it's their server. Maybe they don't mind.


> I think in this regard websocket API is a step in the wrong direction as 
> it makes the developer directly request a connection.  Instead they 
> should be directly requesting a channel and it should be up to the 
> browser to decide if that is a new connection or a shared multiplexed 
> connection.

IMHO, taking control away from the author like this isn't a good idea. 
Giving authors a shared worker and low-level sockets IMHO gives them much 
more flexibility than imposing a particular multiplexing design on them.


On Thu, 16 Apr 2009, Greg Wilkins wrote:
> 
> It is my experience in deploying many Comet Ajax Push applications that 
> scalability is limited either by message rate (for apps like stock 
> tickers) or by number of connections (apps like auctions, IM and/or 
> chat). Connections are a big problem and will only become more so if 
> creating them become easier.

But the tools to fix the problem are available to the author. It's not a 
problem if the author can fix it. It's not like we're preventing authors 
from using multiplexed connections.


> If sockets are to be shared, then it will have to be as a result of 
> those applications and frameworks voluntarily and explicitly taking part 
> in some multi-plexing scheme - just like in the good old days when 
> multi-tasking was done by applications voluntarily giving back the CPU!

The difference is that the resource here is per-server anyway. Cooperative 
multitasking on a per-process level is a perfectly reasonable strategy 
(indeed it underlies "modern" techniques like fibers), it's only a problem 
when it is the mechanism used to share the CPU between different 
processes, where it allows one to take down another.


On Fri, 17 Apr 2009, Greg Wilkins wrote:
> 
> The spec really needs to have the key concepts described in introductory 
> sections rather than by pseudo code repeated several times in the body 
> of the document.

The introduction section is yet to be written, but will be in due course.


> Another problem is that the handshake is defined by explicitly saying 
> what bytes to send: [...] That's a really long winded way of saying send 
> a HTTP/1.1 request with some headers set.

There's a big difference, which is that the exact bytes matter, as the 
client will check that they are sent as described exactly.


> More over, that specification implies that the order has to be:
> 
>  GET /path HTTP/1.1
>  Upgrade: WebSocket
>  Connection: Upgrade
>  Host: acme.com
>  Origin: origin
> 
> and thus it would be illegal to send:
> 
>  GET /path HTTP/1.1
>  Origin:                origin
>  Host:
>    acme.com:80
>  Upgrade:                 WebSocket
>  Connection:      Upgrade

Yes, it is indeed illegal as a WebSocket handshake. This is required 
because if we allowed any correct HTTP request/response that happened to 
have the right semantics, it would be significantly easier to smuggle 
messages to non-HTTP servers.


On Fri, 24 Apr 2009, Greg Wilkins wrote:
> 
> It would be great if the websocket proposal could include standard 
> definitions for mime encoded datagrams.
> 
> Current frame types are:
> 
>   0x00  - sentinel framed UTF-8 message
>   0x80  - length framed binary data.
> 
> I'd like to see two additional frame types supported
> by default:
> 
>   0x01  - sentinel framed UTF-8 encoded MIME message
>   0x81  - length framed MIME message.
> 
> Both these data types would contain a data that commenced with a 
> standard mime header (RFC 2045).  The header is optional and terminated 
> by CR LF CR LF.  Thus these types have a minimal overhead of 4 bytes.

Why can't the 0x00 and 0x80 frames be used for this?


> The websocket API would need to be slightly extended to support some 
> common types of message.

Why can't the existing API be wrapped by a library to offer this?


Generally speaking I am strongly against the idea of the Web platform APIs 
doing things for convenience rather than adding actual new features. As 
noted above, I think it is important for browsers to provide a core 
feature set but not to succomb to a feature creep where they provide all 
the features that authors will ever need.

Christopher Blizzard of Mozilla said it in the context of O3D recently:

   http://www.0xdeadbeef.com/weblog/?p=1223

I think his point applies very much here.



On Thu, 16 Apr 2009, Greg Wilkins wrote:
> Ian Hickson wrote:
> > I don't understand the relevance of HTTP to this discussion.
> 
> HTTP is our starting point.

TCP is my starting point.


> It is how our browsers initiate connections

HTTP is how browsers request resources. That this is layered on top of 
connections is IMHO mostly academic; HTTP isn't a connection protocol.


> It is the protocol that is most likely to get through firewalls

SSL seems even more likely to go through reliably.


> It is a brilliant example of how an extensible flexible protocol can
> adapt to changing usage.

It certainly has been extended a lot.


> So if you want to have a bidirection full-duplex protocol, then the 
> chances are the connection will be created by a HTTP client talking to a 
> HTTP server and then doing an upgrade.

That's certainly one scenario that seems useful to support, yes.


> For that upgrade to be successful, we are going to need to engage all 
> the firewall/gateway/proxy/cache venders and developers so that they 
> will support upgrade either to a specific protocol or to a class of 
> protocols.

Or we can tunnel it over SSL.


> If websocket is just one of many possible upgrade targets, then I don't 
> think that is a problem and websocket can continue to focus on it's 
> prime use-cases. However, if websocket is to be the only widely 
> supported upgrade protocol, then I think it needs to learn some lessons 
> from HTTP as to how to keep most people mostly happy most of the time.

I don't see why WebSocket would be the only widely supported upgrade 
protocol. I do think each protocol should be considered, developed, and 
implemented on its own merits.


On Thu, 9 Apr 2009, Peter Saint-Andre wrote:
> > 
> > Sure, if you can use libraries then implementing any protocol is 
> > trivial. The point is that the protocol should be simple enough that 
> > one doesn't need to rely on libraries to get a compliant 
> > implementation.
> 
> Where does this requirement come from? I have never seen a similar 
> requirement in any RFC.

It comes from experience seeing all those RFCs get implemented in 
incredibly buggy ways over and over again.


> And are you assuming that developers should be able to implement this 
> protocol without any libraries whatsoever? Not even for TCP, UTF-8, or 
> anything else?

I wouldn't expect people to implement the UTF-8 layer themselves, though 
if they don't need to transfer user text (or if they just pass the user 
text unmodified to a database, or some such) then they wouldn't even need 
that. Both UTF-8 and TCP are core parts of most system APIs these days.


On Tue, 14 Apr 2009, Maciej Stachowiak wrote:
> 
> I think multiplexing multiple message streams, though it may seem useful 
> at the protocol level, is not a critical building block.

I agree; I think it is easily enough implemented as a layer on both sides 
(just prefix each message with a number giving the channel ID, and 
dispatch incoming messages to registered listeners based on ID).


On Tue, 14 Apr 2009, Greg Wilkins wrote:
> 
> implementing multiplexing is not the pain point as there are a million 
> ways to do it, ranging from trivial to amazingly complex.

Exactly. That's why we shouldn't bake one into the spec, IMHO.


> Websocket presents a nice simple interface that is easy for web developers
> to use directly.  So they will use it directly!

Why won't people use libraries on the client, as you argue they will to 
use WebSocket on the server?


> There is ZERO chance that all these diverse communities will come together
> and decide on how to share and multiplex connections.

Why is that a problem? We don't need everyone doing the same thing. We 
only need each site to be consistent within the site.


On Tue, 14 Apr 2009, Jamie Lokier wrote:
> 
> It's easy to write a simple HTTP server using a select() loop too, if 
> you know unix, or Perl, or Python.  There are lots of bad ones around to 
> prove it.  It's even easier with fork().

It's a lot more than 100 lines of code to write a compliant HTTP server 
in Perl.


On Thu, 16 Apr 2009, Greg Wilkins wrote:
> 
> I don't know if limiting connections on the server was a motivation for 
> the 2 connection limit, but as a developer of a HTTP server, I 
> appreciate the limit as every connection to the server consumes some 
> kernel resources and buffers.

The Web Socket protocol spec has been updated to clarify that there is no 
limit to the number of connections, but that user agents should never 
actually open more than one _new_ connection to the server at a time (so 
the most number of unexpected connections per client that a server can see 
is one).

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From martin.tyler@boo.org  Fri Apr 24 03:59:08 2009
Return-Path: <martin.tyler@boo.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0A0943A6E51 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 03:59:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.684
X-Spam-Level: 
X-Spam-Status: No, score=-0.684 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, MISSING_HEADERS=1.292]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IhHEOeb7OfX8 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 03:59:06 -0700 (PDT)
Received: from yx-out-2324.google.com (yx-out-2324.google.com [74.125.44.28]) by core3.amsl.com (Postfix) with ESMTP id 943C13A6836 for <hybi@ietf.org>; Fri, 24 Apr 2009 03:59:06 -0700 (PDT)
Received: by yx-out-2324.google.com with SMTP id 8so1309480yxb.49 for <hybi@ietf.org>; Fri, 24 Apr 2009 04:00:24 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.118.8 with SMTP id q8mr2516446agc.80.1240570824481; Fri, 24  Apr 2009 04:00:24 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
Date: Fri, 24 Apr 2009 12:00:24 +0100
Message-ID: <44abafb90904240400n7b7bfe71x796fa731b7866ac4@mail.gmail.com>
From: Martin Tyler <martin.tyler@boo.org>
Cc: hybi@ietf.org
Content-Type: multipart/alternative; boundary=0016362835a84fdbb604684ae7a1
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 10:59:08 -0000

--0016362835a84fdbb604684ae7a1
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Hi folks, I've been on holiday for a few weeks so only just managed to read
up on this list.

On Fri, Apr 24, 2009 at 10:51 AM, Ian Hickson <ian@hixie.ch> wrote:

>
> On Fri, 17 Apr 2009, Greg Wilkins wrote:
> >
> > The spec really needs to have the key concepts described in introductory
> > sections rather than by pseudo code repeated several times in the body
> > of the document.
>
> The introduction section is yet to be written, but will be in due course.
>
>
> > Another problem is that the handshake is defined by explicitly saying
> > what bytes to send: [...] That's a really long winded way of saying send
> > a HTTP/1.1 request with some headers set.
>
> There's a big difference, which is that the exact bytes matter, as the
> client will check that they are sent as described exactly.
>
>
> > More over, that specification implies that the order has to be:
> >
> >  GET /path HTTP/1.1
> >  Upgrade: WebSocket
> >  Connection: Upgrade
> >  Host: acme.com
> >  Origin: origin
> >
> > and thus it would be illegal to send:
> >
> >  GET /path HTTP/1.1
> >  Origin:                origin
> >  Host:
> >    acme.com:80
> >  Upgrade:                 WebSocket
> >  Connection:      Upgrade
>
> Yes, it is indeed illegal as a WebSocket handshake. This is required
> because if we allowed any correct HTTP request/response that happened to
> have the right semantics, it would be significantly easier to smuggle
> messages to non-HTTP servers.
>
>
I'm not sure I understand this smuggling of messages, could you explain?
Otherwise you seem to be saying the exact bytes matter without explaining
why.

It seems odd to me that the headers need to be exact.. One of the big
headaches I can see with WebSocket or anything like it is compatibility with
existing proxies, which is why this has to be some kind of HTTP (at least to
start with) rather than a TCP socket. HTTP proxies will sometimes parse
request/headers from the client and write their own when passing the message
on to the server.. won't requiring the exact bytes above cause problems
here?


A more general point about WebSocket and proxies. I don't know what
experience the people on this list have exactly, I am from Caplin Systems
(Liberator comet server amongst other things), most of our customers are
investment banks with large inflexible web architectures. At the start of
most projects we go through the same problem of explaining that Liberator
will work best if it is Internet facing.. most of these banks setups consist
of multiple tiers of proxies/webservers/appservers etc.. usually everything
sits behind a thin reverse proxy layer. Liberator (and most Comet servers I
would imagine) can work behind these, but you can lose out big time on
performance since the proxy isnt designed for the kind of traffic a comet
server might deliver.

Is this the same in other industry sectors?

Will WebSocket work behind a reverse proxy? It doesn't look like it to me,
the reverse proxy would have to understand WebSocket. I don't have a
solution for this, just wanted to make sure people were aware of it.



Another thing that seems to be a hot topic is this goal of 12 year olds
writing WebSocket servers. I can see why it should be simple for someone to
write a server side WebSocket component (like a CGI script), but writing the
whole server doesn't really seem that important. CGI scripts are run by
webservers, they aren't listening for requests themselves. Surely it would
be better to try and separately work on a new CGI/Servlet/Whatever interface
or guidelines for servers to support that works with WebSocket like
communication rather than thinking about 12 year olds writing the servers
themselves.



Another topic (sorry for not following up on specific posts) has been how
low level the protocol is, eg people want 'channel' like sematics and HTTP
style headers or meta data. I'm not sure where I stand on this, but we
should be careful when suggesting certain features that they are general
enough - we all have certain applications in mind, eg our own projects or
other things you might be considering - but any features should be
sufficiently general concepts that are genuinely useful. Personally I think
I would be fine with the message semantics as they currently stand, but I
can understand how many things could benefit from 'channels' and headers.


Regards
Martin

--0016362835a84fdbb604684ae7a1
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hi folks, I&#39;ve been on holiday for a few weeks so only just managed to =
read up on this list. <br><br><div class=3D"gmail_quote">On Fri, Apr 24, 20=
09 at 10:51 AM, Ian Hickson <span dir=3D"ltr">&lt;ian@hixie.ch&gt;</span> w=
rote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><br>
On Fri, 17 Apr 2009, Greg Wilkins wrote:<br>
&gt;<br>
&gt; The spec really needs to have the key concepts described in introducto=
ry<br>
&gt; sections rather than by pseudo code repeated several times in the body=
<br>
&gt; of the document.<br>
<br>
The introduction section is yet to be written, but will be in due course.<b=
r>
<br>
<br>
&gt; Another problem is that the handshake is defined by explicitly saying<=
br>
&gt; what bytes to send: [...] That&#39;s a really long winded way of sayin=
g send<br>
&gt; a HTTP/1.1 request with some headers set.<br>
<br>
There&#39;s a big difference, which is that the exact bytes matter, as the<=
br>
client will check that they are sent as described exactly.<br>
<br>
<br>
&gt; More over, that specification implies that the order has to be:<br>
&gt;<br>
&gt; =A0GET /path HTTP/1.1<br>
&gt; =A0Upgrade: WebSocket<br>
&gt; =A0Connection: Upgrade<br>
&gt; =A0Host: <a href=3D"http://acme.com" target=3D"_blank">acme.com</a><br=
>
&gt; =A0Origin: origin<br>
&gt;<br>
&gt; and thus it would be illegal to send:<br>
&gt;<br>
&gt; =A0GET /path HTTP/1.1<br>
&gt; =A0Origin: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0origin<br>
&gt; =A0Host:<br>
&gt; =A0 =A0<a href=3D"http://acme.com:80" target=3D"_blank">acme.com:80</a=
><br>
&gt; =A0Upgrade: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 WebSocket<br>
&gt; =A0Connection: =A0 =A0 =A0Upgrade<br>
<br>
Yes, it is indeed illegal as a WebSocket handshake. This is required<br>
because if we allowed any correct HTTP request/response that happened to<br=
>
have the right semantics, it would be significantly easier to smuggle<br>
messages to non-HTTP servers.<br>
<br>
</blockquote></div><br>I&#39;m not sure I understand this smuggling of mess=
ages, could you explain? Otherwise you seem to be saying the exact bytes ma=
tter without explaining why.<br><br>It seems odd to me that the headers nee=
d to be exact.. One of the big headaches I can see with WebSocket or anythi=
ng like it is compatibility with existing proxies, which is why this has to=
 be some kind of HTTP (at least to start with) rather than a TCP socket. HT=
TP proxies will sometimes parse request/headers from the client and write t=
heir own when passing the message on to the server.. won&#39;t requiring th=
e exact bytes above cause problems here?<br>
<br><br>A more general point about WebSocket and proxies. I don&#39;t know =
what experience the people on this list have exactly, I am from Caplin Syst=
ems (Liberator comet server amongst other things), most of our customers ar=
e investment banks with large inflexible web architectures. At the start of=
 most projects we go through the same problem of explaining that Liberator =
will work best if it is Internet facing.. most of these banks setups consis=
t of multiple tiers of proxies/webservers/appservers etc.. usually everythi=
ng sits behind a thin reverse proxy layer. Liberator (and most Comet server=
s I would imagine) can work behind these, but you can lose out big time on =
performance since the proxy isnt designed for the kind of traffic a comet s=
erver might deliver. <br>
<br>Is this the same in other industry sectors?<br><br>Will WebSocket work =
behind a reverse proxy? It doesn&#39;t look like it to me, the reverse prox=
y would have to understand WebSocket. I don&#39;t have a solution for this,=
 just wanted to make sure people were aware of it.<br>
<br><br><br>Another thing that seems to be a hot topic is this goal of 12 y=
ear olds writing WebSocket servers. I can see why it should be simple for s=
omeone to write a server side WebSocket component (like a CGI script), but =
writing the whole server doesn&#39;t really seem that important. CGI script=
s are run by webservers, they aren&#39;t listening for requests themselves.=
 Surely it would be better to try and separately work on a new CGI/Servlet/=
Whatever interface or guidelines for servers to support that works with Web=
Socket like communication rather than thinking about 12 year olds writing t=
he servers themselves.<br>
<br><br><br>Another topic (sorry for not following up on specific posts) ha=
s been how low level the protocol is, eg people want &#39;channel&#39; like=
 sematics and HTTP style headers or meta data. I&#39;m not sure where I sta=
nd on this, but we should be careful when suggesting certain features that =
they are general enough - we all have certain applications in mind, eg our =
own projects or other things you might be considering - but any features sh=
ould be sufficiently general concepts that are genuinely useful. Personally=
 I think I would be fine with the message semantics as they currently stand=
, but I can understand how many things could benefit from &#39;channels&#39=
; and headers.<br>
<br><br>Regards<br>Martin<br>

--0016362835a84fdbb604684ae7a1--

From gregw@webtide.com  Fri Apr 24 07:12:35 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8BFE53A698F for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 07:12:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.555
X-Spam-Level: 
X-Spam-Status: No, score=-2.555 tagged_above=-999 required=5 tests=[AWL=0.044,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hNhFBS4xuacT for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 07:12:34 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.29]) by core3.amsl.com (Postfix) with ESMTP id BBD613A68F0 for <hybi@ietf.org>; Fri, 24 Apr 2009 07:12:34 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 3so991135ywj.49 for <hybi@ietf.org>; Fri, 24 Apr 2009 07:13:53 -0700 (PDT)
Received: by 10.90.118.8 with SMTP id q8mr2739300agc.40.1240582433038; Fri, 24 Apr 2009 07:13:53 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 20sm2502601agb.65.2009.04.24.07.13.50 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 24 Apr 2009 07:13:52 -0700 (PDT)
Message-ID: <49F1C918.4060007@webtide.com>
Date: Sat, 25 Apr 2009 00:13:44 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com>	<5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com>	<4936C476.1010708@gmail.com>	<5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com>	<3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 14:12:35 -0000

Wow!

For someone who is advocating against multi-plexing support you have certainly
merged a lot of threads into one.

I think you're putting up a lot of straw men in your responses.  You've assumed
that because I'm highlighting the problem of sharing a channel that I'm advocating
a one-size-fits mandatory multiplexing solution.     I think there might be some
middle ground between those extremes.

You also can't argue that users should be able to not use libraries and
write their own websocket servers, then on the other hand say that libraries
are the answer to diverse communities spontaneously coordinating their
socket sharing solutions (says Greg - guilty of building his own straw man).

For more detailed comments, I'll try to demux your thread and respond to specific
points in the original threads


cheers


From jamie@shareable.org  Fri Apr 24 08:00:03 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9409A28C24B for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 08:00:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.432
X-Spam-Level: 
X-Spam-Status: No, score=-3.432 tagged_above=-999 required=5 tests=[AWL=-0.833, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HuXYHZfa6L20 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 08:00:02 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id EF3923A6FEA for <hybi@ietf.org>; Fri, 24 Apr 2009 07:59:54 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LxMtr-0002b3-Jr; Fri, 24 Apr 2009 16:01:11 +0100
Date: Fri, 24 Apr 2009 16:01:11 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Martin Tyler <martin.tyler@boo.org>
Message-ID: <20090424150111.GB9307@shareable.org>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <44abafb90904240400n7b7bfe71x796fa731b7866ac4@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <44abafb90904240400n7b7bfe71x796fa731b7866ac4@mail.gmail.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 15:00:03 -0000

Martin Tyler wrote:
> HTTP proxies will sometimes parse request/headers
> from the client and write their own when passing the message on to the server..
> won't requiring the exact bytes above cause problems here?

In fact HTTP proxies are _required_ to do this by the HTTP spec: they
must add a Via header.  Not all do.

Also, in practice, you need to be able to use proxy authentication
sometimes - there are several ways that is done.

-- Jamie

From gregw@webtide.com  Fri Apr 24 08:46:04 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 379B83A6986 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 08:46:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.556
X-Spam-Level: 
X-Spam-Status: No, score=-2.556 tagged_above=-999 required=5 tests=[AWL=0.043,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id vC99S0jGlpfi for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 08:46:02 -0700 (PDT)
Received: from yw-out-2324.google.com (yw-out-2324.google.com [74.125.46.30]) by core3.amsl.com (Postfix) with ESMTP id 0B4B43A6DE3 for <hybi@ietf.org>; Fri, 24 Apr 2009 08:45:55 -0700 (PDT)
Received: by yw-out-2324.google.com with SMTP id 3so1042482ywj.49 for <hybi@ietf.org>; Fri, 24 Apr 2009 08:47:14 -0700 (PDT)
Received: by 10.90.56.17 with SMTP id e17mr2787077aga.61.1240588033929; Fri, 24 Apr 2009 08:47:13 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 5sm2657484agc.55.2009.04.24.08.47.12 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 24 Apr 2009 08:47:13 -0700 (PDT)
Message-ID: <49F1DEF9.2000209@webtide.com>
Date: Sat, 25 Apr 2009 01:47:05 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com>	<5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com>	<4936C476.1010708@gmail.com>	<5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com>	<3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 15:46:04 -0000

Ian Hickson wrote:
> On Fri, 17 Apr 2009, Greg Wilkins wrote:
>> More over, that specification implies that the order has to be:
>>
>>  GET /path HTTP/1.1
>>  Upgrade: WebSocket
>>  Connection: Upgrade
>>  Host: acme.com
>>  Origin: origin
>>
>> and thus it would be illegal to send:
>>
>>  GET /path HTTP/1.1
>>  Origin:                origin
>>  Host:
>>    acme.com:80
>>  Upgrade:                 WebSocket
>>  Connection:      Upgrade
> 
> Yes, it is indeed illegal as a WebSocket handshake. This is required 
> because if we allowed any correct HTTP request/response that happened to 
> have the right semantics, it would be significantly easier to smuggle 
> messages to non-HTTP servers.


Ian,

Surely a message that happens to have both correct syntax and semantics
is a correct and acceptable message!

The philosophy often associated IETF protocols is:
   "be strict in what you generate, but forgiving in what you accept."

The fact that html has not been displaced by xhtml is testament to
why such flexibility is required to achieve interoperable in
an imperfect world.




> On Fri, 24 Apr 2009, Greg Wilkins wrote:
>> It would be great if the websocket proposal could include standard 
>> definitions for mime encoded datagrams.
>>
>> Current frame types are:
>>
>>   0x00  - sentinel framed UTF-8 message
>>   0x80  - length framed binary data.
>>
>> I'd like to see two additional frame types supported
>> by default:
>>
>>   0x01  - sentinel framed UTF-8 encoded MIME message
>>   0x81  - length framed MIME message.
>>
>> Both these data types would contain a data that commenced with a 
>> standard mime header (RFC 2045).  The header is optional and terminated 
>> by CR LF CR LF.  Thus these types have a minimal overhead of 4 bytes.
> 
> Why can't the 0x00 and 0x80 frames be used for this?

I could equally say that since 0x80 packets can be be used to send
UTF-8 messages, then the 0x00 frame type is not needed.
An application could just assume the message contents are what is
expected.  Why have a frame type byte at all and both ends can
just assume the content and framing type!

Unfortunately such assumed solutions do result in robustness:
  + Clients and servers are not always implemented or upgraded
    simultaneously and sometime packets will be sent that the other
    end cannot understand.
  + Sometime specifications are interpreted differently
    interpretations can be implemented, which while semantically
    correct have slight variations in implementation.
  + Clients do connect to wrong ports.

Standardized metadata is a proven way that errors and differing
versions or interpretations can be handled, ignored or alarmed.

MIME encapsulation is the internet equivalent of the hydrogen atom
inscribed on the side of the voyager space probes.  It is a
lingua-franca on which diverse parties can base more complex and
diverse exchanges and protocols.


>> The websocket API would need to be slightly extended to support some 
>> common types of message.
> 
> Why can't the existing API be wrapped by a library to offer this?
> 
> Generally speaking I am strongly against the idea of the Web platform APIs 
> doing things for convenience rather than adding actual new features. As 
> noted above, I think it is important for browsers to provide a core 
> feature set but not to succumb to a feature creep where they provide all 
> the features that authors will ever need.

Standardized meta data is not for convenience, but for robustness
and flexibility.

I think that rather than a limit frame-type byte, MIME encapsulation
should be the core feature of websocket.

But if a frame-type is to be a core feature of websocket, then
I'm proposing use of that core feature that will 2 of the
254 currently unused values.

Surely MIME encapsulation has already proven itself enough
to warrant such a small investment?




>> It is the protocol that is most likely to get through firewalls
> 
> SSL seems even more likely to go through reliably.

SSL is a hugely expensive resource for servers, often requiring
dedicated hardware.

While I'm generally supportive of innovative ways to re-use
existing technologies, I do think the cost of using SSL for
non secure traffic is likely to be prohibitive.

It strikes me as counter productive to byte squeeze a protocol
such that it can't be carried over HTTP, only to then force
SSL encryption.



> On Thu, 16 Apr 2009, Greg Wilkins wrote:
>> I don't know if limiting connections on the server was a motivation for 
>> the 2 connection limit, but as a developer of a HTTP server, I 
>> appreciate the limit as every connection to the server consumes some 
>> kernel resources and buffers.
> 
> The Web Socket protocol spec has been updated to clarify that there is no 
> limit to the number of connections

Is there any analysis or evidence that this will not adversely affect
TCP/IP flow control or server scalability?

What are the reasons that drove the limitation of HTTP connections
that do not apply to websocket?


 > but that user agents should never
> actually open more than one _new_ connection to the server at a time (so 
> the most number of unexpected connections per client that a server can see 
> is one).

I'm not sure how this helps?

As a server implementer, I do not have any issue with the rate of new connections.
The problems that I have are with the resources allocated to the total
number of connections.


regards












From mario.balibrera@gmail.com  Fri Apr 24 11:34:02 2009
Return-Path: <mario.balibrera@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E1AD13A6A24 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 11:34:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.178
X-Spam-Level: 
X-Spam-Status: No, score=-2.178 tagged_above=-999 required=5 tests=[AWL=-0.180, BAYES_00=-2.599, HTML_MESSAGE=0.001, J_CHICKENPOX_23=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ONYDsydwUECu for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 11:33:54 -0700 (PDT)
Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.229]) by core3.amsl.com (Postfix) with ESMTP id 6068D3A6A20 for <hybi@ietf.org>; Fri, 24 Apr 2009 11:33:53 -0700 (PDT)
Received: by rv-out-0506.google.com with SMTP id f6so640330rvb.5 for <hybi@ietf.org>; Fri, 24 Apr 2009 11:35:12 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=AI/zpInAtCKaHqfSwtPP77WGp91Hd7dLcWDHPhrBSWc=; b=emsLgySNbmgjGLxcVsU4GkiK9Oj/2ljvQkX3Nu/eIFSkPzzc7feuJGxAK+HPRY1ezy NLj37q3xT4cnavUg6Yoon3ir3XViCBIz2X6zDlg6r9w1XCLDTYqNpaCprHJ62nuK+odO Kt7gi+6HYSSgy7xbGp5CnYAyQmt9faqazgmNA=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=tTX5/rRT4WjkKt39UoS2e3KMZ5jfgB1GDlFzrIZqSILoXaTCwk6xRmGErBVYLMYjvA w+1xdKjHynLrP9mCFtDL4UPjETIaLUNNGG+H4JVUu41WkxRHv2ceAn6KUm195sbrxck+ RwLeph4XdTunBQqnE6BwOj59sgJvLWOAyp7tw=
MIME-Version: 1.0
Received: by 10.114.80.18 with SMTP id d18mr1487858wab.147.1240598110999; Fri,  24 Apr 2009 11:35:10 -0700 (PDT)
In-Reply-To: <49F1DEF9.2000209@webtide.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com>
Date: Fri, 24 Apr 2009 11:35:10 -0700
Message-ID: <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>
From: Mario Balibrera <mario.balibrera@gmail.com>
To: Greg Wilkins <gregw@webtide.com>
Content-Type: multipart/alternative; boundary=0016364179e1b714fe04685141a2
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 18:34:03 -0000

--0016364179e1b714fe04685141a2
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

GW wrote:
What are the reasons that drove the limitation of HTTP connections
that do not apply to websocket?

A Frank points out in an earlier post, the 2-connection limit was conceived
in large part to prevent a browser from crashing a sh*tty synchronous web
server with 10 image requests. No one would ever, ever, ever consider
deploying a persistent-connection app on something that fragile. Ever.

MT wrote:
Another thing that seems to be a hot topic is this goal of 12 year olds
writing WebSocket servers. I can see why it should be simple for someone to
write a server side WebSocket component (like a CGI script), but writing the
whole server doesn't really seem that important.

By the same token, if a 12 year old wants to write a CGI script, she should
use existing tools that exist for that precise reason. Neither WebSocket nor
any of the protocols we're discussing here are even in the same ballpark.

Really, the 12 year old would most likely not write a WebSocket server,
because proxies and WS readers and so forth would (and do) exist. The kid
wouldn't have to implement anything resembling a server. At the most basic
level, it would be a matter of writing an onread function that parses some
plain text. Could it be any more hello world? (I realize this is basically
what you're saying, Martin. I'm just mad that CGI keeps coming up.)

Mario Balibrera


On Fri, Apr 24, 2009 at 8:47 AM, Greg Wilkins <gregw@webtide.com> wrote:

> Ian Hickson wrote:
>
>> On Fri, 17 Apr 2009, Greg Wilkins wrote:
>>
>>> More over, that specification implies that the order has to be:
>>>
>>>  GET /path HTTP/1.1
>>>  Upgrade: WebSocket
>>>  Connection: Upgrade
>>>  Host: acme.com
>>>  Origin: origin
>>>
>>> and thus it would be illegal to send:
>>>
>>>  GET /path HTTP/1.1
>>>  Origin:                origin
>>>  Host:
>>>   acme.com:80
>>>  Upgrade:                 WebSocket
>>>  Connection:      Upgrade
>>>
>>
>> Yes, it is indeed illegal as a WebSocket handshake. This is required
>> because if we allowed any correct HTTP request/response that happened to
>> have the right semantics, it would be significantly easier to smuggle
>> messages to non-HTTP servers.
>>
>
>
> Ian,
>
> Surely a message that happens to have both correct syntax and semantics
> is a correct and acceptable message!
>
> The philosophy often associated IETF protocols is:
>  "be strict in what you generate, but forgiving in what you accept."
>
> The fact that html has not been displaced by xhtml is testament to
> why such flexibility is required to achieve interoperable in
> an imperfect world.
>
>
>
>
>  On Fri, 24 Apr 2009, Greg Wilkins wrote:
>>
>>> It would be great if the websocket proposal could include standard
>>> definitions for mime encoded datagrams.
>>>
>>> Current frame types are:
>>>
>>>  0x00  - sentinel framed UTF-8 message
>>>  0x80  - length framed binary data.
>>>
>>> I'd like to see two additional frame types supported
>>> by default:
>>>
>>>  0x01  - sentinel framed UTF-8 encoded MIME message
>>>  0x81  - length framed MIME message.
>>>
>>> Both these data types would contain a data that commenced with a standard
>>> mime header (RFC 2045).  The header is optional and terminated by CR LF CR
>>> LF.  Thus these types have a minimal overhead of 4 bytes.
>>>
>>
>> Why can't the 0x00 and 0x80 frames be used for this?
>>
>
> I could equally say that since 0x80 packets can be be used to send
> UTF-8 messages, then the 0x00 frame type is not needed.
> An application could just assume the message contents are what is
> expected.  Why have a frame type byte at all and both ends can
> just assume the content and framing type!
>
> Unfortunately such assumed solutions do result in robustness:
>  + Clients and servers are not always implemented or upgraded
>   simultaneously and sometime packets will be sent that the other
>   end cannot understand.
>  + Sometime specifications are interpreted differently
>   interpretations can be implemented, which while semantically
>   correct have slight variations in implementation.
>  + Clients do connect to wrong ports.
>
> Standardized metadata is a proven way that errors and differing
> versions or interpretations can be handled, ignored or alarmed.
>
> MIME encapsulation is the internet equivalent of the hydrogen atom
> inscribed on the side of the voyager space probes.  It is a
> lingua-franca on which diverse parties can base more complex and
> diverse exchanges and protocols.
>
>
>  The websocket API would need to be slightly extended to support some
>>> common types of message.
>>>
>>
>> Why can't the existing API be wrapped by a library to offer this?
>>
>> Generally speaking I am strongly against the idea of the Web platform APIs
>> doing things for convenience rather than adding actual new features. As
>> noted above, I think it is important for browsers to provide a core feature
>> set but not to succumb to a feature creep where they provide all the
>> features that authors will ever need.
>>
>
> Standardized meta data is not for convenience, but for robustness
> and flexibility.
>
> I think that rather than a limit frame-type byte, MIME encapsulation
> should be the core feature of websocket.
>
> But if a frame-type is to be a core feature of websocket, then
> I'm proposing use of that core feature that will 2 of the
> 254 currently unused values.
>
> Surely MIME encapsulation has already proven itself enough
> to warrant such a small investment?
>
>
>
>
>  It is the protocol that is most likely to get through firewalls
>>>
>>
>> SSL seems even more likely to go through reliably.
>>
>
> SSL is a hugely expensive resource for servers, often requiring
> dedicated hardware.
>
> While I'm generally supportive of innovative ways to re-use
> existing technologies, I do think the cost of using SSL for
> non secure traffic is likely to be prohibitive.
>
> It strikes me as counter productive to byte squeeze a protocol
> such that it can't be carried over HTTP, only to then force
> SSL encryption.
>
>
>
>  On Thu, 16 Apr 2009, Greg Wilkins wrote:
>>
>>> I don't know if limiting connections on the server was a motivation for
>>> the 2 connection limit, but as a developer of a HTTP server, I appreciate
>>> the limit as every connection to the server consumes some kernel resources
>>> and buffers.
>>>
>>
>> The Web Socket protocol spec has been updated to clarify that there is no
>> limit to the number of connections
>>
>
> Is there any analysis or evidence that this will not adversely affect
> TCP/IP flow control or server scalability?
>
> What are the reasons that drove the limitation of HTTP connections
> that do not apply to websocket?
>
>
> > but that user agents should never
>
>> actually open more than one _new_ connection to the server at a time (so
>> the most number of unexpected connections per client that a server can see
>> is one).
>>
>
> I'm not sure how this helps?
>
> As a server implementer, I do not have any issue with the rate of new
> connections.
> The problems that I have are with the resources allocated to the total
> number of connections.
>
>
> regards
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>

--0016364179e1b714fe04685141a2
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

GW wrote:<br>What are the reasons that drove the limitation of HTTP connect=
ions<br>
that do not apply to websocket?<br><br>A Frank points out in an earlier pos=
t, the 2-connection limit was conceived in large part to prevent a browser =
from crashing a sh*tty synchronous web server with 10 image requests. No on=
e would ever, ever, ever consider deploying a persistent-connection app on =
something that fragile. Ever.<br>
<br>MT wrote:<br>Another thing that seems to be a hot topic is this goal of=
 12 year olds
writing WebSocket servers. I can see why it should be simple for
someone to write a server side WebSocket component (like a CGI script),
but writing the whole server doesn&#39;t really seem that important.<br><br=
>By the same token, if a 12 year old wants to write a CGI script, she shoul=
d use existing tools that exist for that precise reason. Neither WebSocket =
nor any of the protocols we&#39;re discussing here are even in the same bal=
lpark.<br>
<br>Really, the 12 year old would most likely not write a WebSocket server,=
 because proxies and WS readers and so forth would (and do) exist. The kid =
wouldn&#39;t have to implement anything resembling a server. At the most ba=
sic level, it would be a matter of writing an onread function that parses s=
ome plain text. Could it be any more hello world? (I realize this is basica=
lly what you&#39;re saying, Martin. I&#39;m just mad that CGI keeps coming =
up.)<br>
<br>Mario Balibrera<br><br><br><div class=3D"gmail_quote">On Fri, Apr 24, 2=
009 at 8:47 AM, Greg Wilkins <span dir=3D"ltr">&lt;<a href=3D"mailto:gregw@=
webtide.com">gregw@webtide.com</a>&gt;</span> wrote:<br><blockquote class=
=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin=
: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Ian Hickson wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class=3D"im"=
>
On Fri, 17 Apr 2009, Greg Wilkins wrote:<br>
</div><div class=3D"im"><blockquote class=3D"gmail_quote" style=3D"border-l=
eft: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left:=
 1ex;">
More over, that specification implies that the order has to be:<br>
<br>
=A0GET /path HTTP/1.1<br>
=A0Upgrade: WebSocket<br>
=A0Connection: Upgrade<br>
=A0Host: <a href=3D"http://acme.com" target=3D"_blank">acme.com</a><br>
=A0Origin: origin<br>
<br>
and thus it would be illegal to send:<br>
<br>
=A0GET /path HTTP/1.1<br>
=A0Origin: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0origin<br>
=A0Host:<br>
 =A0 <a href=3D"http://acme.com:80" target=3D"_blank">acme.com:80</a><br>
=A0Upgrade: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 WebSocket<br>
=A0Connection: =A0 =A0 =A0Upgrade<br>
</blockquote>
<br>
Yes, it is indeed illegal as a WebSocket handshake. This is required becaus=
e if we allowed any correct HTTP request/response that happened to have the=
 right semantics, it would be significantly easier to smuggle messages to n=
on-HTTP servers.<br>

</div></blockquote>
<br>
<br>
Ian,<br>
<br>
Surely a message that happens to have both correct syntax and semantics<br>
is a correct and acceptable message!<br>
<br>
The philosophy often associated IETF protocols is:<br>
 =A0&quot;be strict in what you generate, but forgiving in what you accept.=
&quot;<br>
<br>
The fact that html has not been displaced by xhtml is testament to<br>
why such flexibility is required to achieve interoperable in<br>
an imperfect world.<div class=3D"im"><br>
<br>
<br>
<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Fri, 24 Apr 2009, Greg Wilkins wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
It would be great if the websocket proposal could include standard definiti=
ons for mime encoded datagrams.<br>
<br>
Current frame types are:<br>
<br>
 =A00x00 =A0- sentinel framed UTF-8 message<br>
 =A00x80 =A0- length framed binary data.<br>
<br>
I&#39;d like to see two additional frame types supported<br>
by default:<br>
<br>
 =A00x01 =A0- sentinel framed UTF-8 encoded MIME message<br>
 =A00x81 =A0- length framed MIME message.<br>
<br>
Both these data types would contain a data that commenced with a standard m=
ime header (RFC 2045). =A0The header is optional and terminated by CR LF CR=
 LF. =A0Thus these types have a minimal overhead of 4 bytes.<br>
</blockquote>
<br>
Why can&#39;t the 0x00 and 0x80 frames be used for this?<br>
</blockquote>
<br></div>
I could equally say that since 0x80 packets can be be used to send<br>
UTF-8 messages, then the 0x00 frame type is not needed.<br>
An application could just assume the message contents are what is<br>
expected. =A0Why have a frame type byte at all and both ends can<br>
just assume the content and framing type!<br>
<br>
Unfortunately such assumed solutions do result in robustness:<br>
=A0+ Clients and servers are not always implemented or upgraded<br>
 =A0 simultaneously and sometime packets will be sent that the other<br>
 =A0 end cannot understand.<br>
=A0+ Sometime specifications are interpreted differently<br>
 =A0 interpretations can be implemented, which while semantically<br>
 =A0 correct have slight variations in implementation.<br>
=A0+ Clients do connect to wrong ports.<br>
<br>
Standardized metadata is a proven way that errors and differing<br>
versions or interpretations can be handled, ignored or alarmed.<br>
<br>
MIME encapsulation is the internet equivalent of the hydrogen atom<br>
inscribed on the side of the voyager space probes. =A0It is a<br>
lingua-franca on which diverse parties can base more complex and<br>
diverse exchanges and protocols.<br>
<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class=3D"im"=
><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204,=
 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">

The websocket API would need to be slightly extended to support some common=
 types of message.<br>
</blockquote>
<br>
Why can&#39;t the existing API be wrapped by a library to offer this?<br>
<br></div>
Generally speaking I am strongly against the idea of the Web platform APIs =
doing things for convenience rather than adding actual new features. As not=
ed above, I think it is important for browsers to provide a core feature se=
t but not to succumb to a feature creep where they provide all the features=
 that authors will ever need.<br>

</blockquote>
<br>
Standardized meta data is not for convenience, but for robustness<br>
and flexibility.<br>
<br>
I think that rather than a limit frame-type byte, MIME encapsulation<br>
should be the core feature of websocket.<br>
<br>
But if a frame-type is to be a core feature of websocket, then<br>
I&#39;m proposing use of that core feature that will 2 of the<br>
254 currently unused values.<br>
<br>
Surely MIME encapsulation has already proven itself enough<br>
to warrant such a small investment?<div class=3D"im"><br>
<br>
<br>
<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><blockquote class=
=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin=
: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">

It is the protocol that is most likely to get through firewalls<br>
</blockquote>
<br>
SSL seems even more likely to go through reliably.<br>
</blockquote>
<br></div>
SSL is a hugely expensive resource for servers, often requiring<br>
dedicated hardware.<br>
<br>
While I&#39;m generally supportive of innovative ways to re-use<br>
existing technologies, I do think the cost of using SSL for<br>
non secure traffic is likely to be prohibitive.<br>
<br>
It strikes me as counter productive to byte squeeze a protocol<br>
such that it can&#39;t be carried over HTTP, only to then force<br>
SSL encryption.<div class=3D"im"><br>
<br>
<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Thu, 16 Apr 2009, Greg Wilkins wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I don&#39;t know if limiting connections on the server was a motivation for=
 the 2 connection limit, but as a developer of a HTTP server, I appreciate =
the limit as every connection to the server consumes some kernel resources =
and buffers.<br>

</blockquote>
<br>
The Web Socket protocol spec has been updated to clarify that there is no l=
imit to the number of connections<br>
</blockquote>
<br></div>
Is there any analysis or evidence that this will not adversely affect<br>
TCP/IP flow control or server scalability?<br>
<br>
What are the reasons that drove the limitation of HTTP connections<br>
that do not apply to websocket?<div class=3D"im"><br>
<br>
<br>
&gt; but that user agents should never<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
actually open more than one _new_ connection to the server at a time (so th=
e most number of unexpected connections per client that a server can see is=
 one).<br>
</blockquote>
<br></div>
I&#39;m not sure how this helps?<br>
<br>
As a server implementer, I do not have any issue with the rate of new conne=
ctions.<br>
The problems that I have are with the resources allocated to the total<br>
number of connections.<br>
<br>
<br>
regards<div><div></div><div class=3D"h5"><br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
_______________________________________________<br>
hybi mailing list<br>
<a href=3D"mailto:hybi@ietf.org" target=3D"_blank">hybi@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/hybi" target=3D"_blank">ht=
tps://www.ietf.org/mailman/listinfo/hybi</a><br>
</div></div></blockquote></div><br>

--0016364179e1b714fe04685141a2--

From ian@hixie.ch  Fri Apr 24 13:07:47 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3A5AD3A6C58 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 13:07:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.213
X-Spam-Level: 
X-Spam-Status: No, score=-3.213 tagged_above=-999 required=5 tests=[AWL=-0.614, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id euXA4i4Z5ojh for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 13:07:45 -0700 (PDT)
Received: from looneymail-a5.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id CC19B3A6966 for <hybi@ietf.org>; Fri, 24 Apr 2009 13:07:45 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a5.g.dreamhost.com (Postfix) with ESMTP id E0CA9122511 for <hybi@ietf.org>; Fri, 24 Apr 2009 13:09:03 -0700 (PDT)
Date: Fri, 24 Apr 2009 20:09:03 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: hybi@ietf.org
In-Reply-To: <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>
Message-ID: <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 20:07:47 -0000

On Sat, 25 Apr 2009, Greg Wilkins wrote:
> Ian Hickson wrote:
> > On Fri, 17 Apr 2009, Greg Wilkins wrote:
> > > More over, that specification implies that the order has to be:
> > > 
> > >  GET /path HTTP/1.1
> > >  Upgrade: WebSocket
> > >  Connection: Upgrade
> > >  Host: acme.com
> > >  Origin: origin
> > > 
> > > and thus it would be illegal to send:
> > > 
> > >  GET /path HTTP/1.1
> > >  Origin:                origin
> > >  Host:
> > >    acme.com:80
> > >  Upgrade:                 WebSocket
> > >  Connection:      Upgrade
> > 
> > Yes, it is indeed illegal as a WebSocket handshake. This is required because
> > if we allowed any correct HTTP request/response that happened to have the
> > right semantics, it would be significantly easier to smuggle messages to
> > non-HTTP servers.
> 
> Surely a message that happens to have both correct syntax and semantics 
> is a correct and acceptable message!

The problem is that if we are flexible with the handshake, then we can 
much more easily end up finding ourselves accepting a connection from a 
host that has been tricked into sending something that looks like a 
handshake but is in fact just the normal operation of another protocol. 
One of the requirements is that we be as sure as possible that the target 
host is truly willing to opt into communication with the script's origin.


> > > I'd like to see two additional frame types supported
> > > by default:
> > > 
> > >   0x01  - sentinel framed UTF-8 encoded MIME message
> > >   0x81  - length framed MIME message.
> > > 
> > > Both these data types would contain a data that commenced with a standard
> > > mime header (RFC 2045).  The header is optional and terminated by CR LF CR
> > > LF.  Thus these types have a minimal overhead of 4 bytes.
> > 
> > Why can't the 0x00 and 0x80 frames be used for this?
> 
> I could equally say that since 0x80 packets can be be used to send
> UTF-8 messages, then the 0x00 frame type is not needed.

Indeed, people have argued that. The reasons for having both of these have 
been discussed before on this list, but they boil down to th need for 
distinct APIs for binary and text data requiring different frame types, 
and a concern over implementations not being able to accurately measure 
UTF-8 but errors in measuring UTF-8 being difficult to detect.


> An application could just assume the message contents are what is 
> expected.  Why have a frame type byte at all and both ends can just 
> assume the content and framing type!

That would be ideal, yes. The protocol is (IMHO) about as close as we can 
get to that right now.


> Unfortunately such assumed solutions do result in robustness:
>  + Clients and servers are not always implemented or upgraded
>    simultaneously and sometime packets will be sent that the other
>    end cannot understand.

The protocol is designed with this in mind (in fact as currently designed, 
you'll notice that the client-side supports length-delimited binary 
packets, for future compatibility, but the server side doesn't, since it 
knows what types of packets it needs to support).


>  + Sometime specifications are interpreted differently
>    interpretations can be implemented, which while semantically
>    correct have slight variations in implementation.

As far as I can tell, there is no ambiguity allowing for different legal 
interpretations in this case. (This is why the spec uses step-by-step 
algorithms with clear conformance criteria, instead of a narrative and 
constraint-based definition.) If you find a way in which the spec is 
ambiguous, please let me know.


>  + Clients do connect to wrong ports.

This is one reason we have the handshake and why it is so strict.


> Standardized metadata is a proven way that errors and differing versions 
> or interpretations can be handled, ignored or alarmed.

It's _a_ way, indeed. IMHO not a particularly successful one, for example 
look at the many ways in which HTTP status codes are ignored (e.g. <img> 
pointing to a 404 image is treated like a 200).


> MIME encapsulation is the internet equivalent of the hydrogen atom 
> inscribed on the side of the voyager space probes. It is a lingua-franca 
> on which diverse parties can base more complex and diverse exchanges and 
> protocols.

MIME is a way to transfer resources. WebSocket isn't about resources. If 
you want to transfer resources, use something like HTTP.


> > Generally speaking I am strongly against the idea of the Web platform 
> > APIs doing things for convenience rather than adding actual new 
> > features. As noted above, I think it is important for browsers to 
> > provide a core feature set but not to succumb to a feature creep where 
> > they provide all the features that authors will ever need.
> 
> Standardized meta data is not for convenience, but for robustness and 
> flexibility.

It's trivial to include metadata in text packets, and it can be done in 
whatever way is most convenient for the type of data being sent. If a 
simple code is more effective, that can be done. If each message needs a 
URL and EXIF data transfered as an RDF block, that can be done. If 
MIME-like headers are best, that can be done.


> Surely MIME encapsulation has already proven itself enough to warrant 
> such a small investment?

It has proven itself to be adequate (though by no means outstanding, IMHO) 
for labeling resources. And it is quite possible to use MIME encapsulation 
in a WebSocket subprotocol if that is desired. But why force it on 
everyone? It's certainly not the be-all and end-all of metadata labeling, 
and it isn't a feature everyone needs.


> > On Thu, 16 Apr 2009, Greg Wilkins wrote:
> > > I don't know if limiting connections on the server was a motivation 
> > > for the 2 connection limit, but as a developer of a HTTP server, I 
> > > appreciate the limit as every connection to the server consumes some 
> > > kernel resources and buffers.
> > 
> > The Web Socket protocol spec has been updated to clarify that there is 
> > no limit to the number of connections
> 
> Is there any analysis or evidence that this will not adversely affect 
> TCP/IP flow control or server scalability?

No. If anyone is able to do such an analysis, I'd be very interested in 
seeing the results. It should be noted though that load on the server 
itself isn't a concern, since the author is the one starting the 
connections and he can quite easily just multiplex everything using a 
shared worker if so desired. Only the impact on the client and the network 
are realy a concern.


> What are the reasons that drove the limitation of HTTP connections that 
> do not apply to websocket?

The reasons that drove the limitation of HTTP connections -- the ability 
for untrusted third-party content to overload a server with many 
connections -- DO apply to WebSocket, and WebSocket has been limited to 
prevent such abuse (even more so than HTTP; WebSocket only allows one 
pending non-established connection at a time).

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From jamie@shareable.org  Fri Apr 24 14:12:09 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 765593A6BD3 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:12:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.399
X-Spam-Level: 
X-Spam-Status: No, score=-3.399 tagged_above=-999 required=5 tests=[AWL=-0.800, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZhJHyGppcbGv for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:12:08 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 613DC3A6E58 for <hybi@ietf.org>; Fri, 24 Apr 2009 14:11:57 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LxShu-0004vq-2K; Fri, 24 Apr 2009 22:13:14 +0100
Date: Fri, 24 Apr 2009 22:13:14 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090424211314.GA18260@shareable.org>
References: <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 21:12:09 -0000

Ian Hickson wrote:
> > An application could just assume the message contents are what is 
> > expected.  Why have a frame type byte at all and both ends can just 
> > assume the content and framing type!
> 
> That would be ideal, yes. The protocol is (IMHO) about as close as we can 
> get to that right now.

Can you say a bit about why that isn't possible?

I asked myself the same thing a while ago: since every application has
to know the format of what it's going to receive _anyway_, in
considerable byte/character-level detail, and application-defined
sub-framing metadata (such as it's own sequence numbers, target object
ids or whatever), what is the purpose of the framing byte?

Why is it not enough for the receiving API to offer
receiveBinaryFixed(length), receiveBinaryFrame() and
receiveTextUTF8Frame() methods?

If it's just to catch a class of errors, where what was sent didn't
match the expected format, why is it limited to just the two types?

> > Unfortunately such assumed solutions do result in robustness:
> >  + Clients and servers are not always implemented or upgraded
> >    simultaneously and sometime packets will be sent that the other
> >    end cannot understand.
> 
> The protocol is designed with this in mind (in fact as currently designed, 
> you'll notice that the client-side supports length-delimited binary 
> packets, for future compatibility, but the server side doesn't, since it 
> knows what types of packets it needs to support).

That's interesting.  If the protocol reaches mainstream browsers, I
imagine using the length-delimited binary packets if they work.

It will be a shame if it's "specified" but half the client
implementations are broken because that part was never tested.  I'd
have to have to encode binary data into UTF-8 *and then decode UTF-8
back into binary in Javascript* just to make it work on real
deployments. :-)

> >  + Clients do connect to wrong ports.
> 
> This is one reason we have the handshake and why it is so strict.

It sounds like the reason could be explained more clearly in the
document, as the question shouldn't have come up.

I'm a bit concerned if it's so strict that proxies (including
transparent ones) will break it though.

> > Standardized metadata is a proven way that errors and differing versions 
> > or interpretations can be handled, ignored or alarmed.
> 
> It's _a_ way, indeed. IMHO not a particularly successful one, for example 
> look at the many ways in which HTTP status codes are ignored (e.g. <img> 
> pointing to a 404 image is treated like a 200).

Eh?  Is it?  I've always seen a broken image icon, I'm pretty sure.

> > MIME encapsulation is the internet equivalent of the hydrogen atom 
> > inscribed on the side of the voyager space probes. It is a lingua-franca 
> > on which diverse parties can base more complex and diverse exchanges and 
> > protocols.
> 
> MIME is a way to transfer resources. WebSocket isn't about resources. If 
> you want to transfer resources, use something like HTTP.

WebSocket is about messages.  HTTP (as a protocol, not a data model)
is about messages with some metadata and a request-response model.
I'm not seeing that much difference.

I understand that WebSockets needs to be simpler, but I do thing the
"not about resources" argument is spurious.  Of coures we will
transfer small HTTP-style resource messages over WebSockets.  I
foresee that being a major use of it.  Even if it ends up looking like
a Bayeux-style JSON frame in practice.

> > > Generally speaking I am strongly against the idea of the Web platform 
> > > APIs doing things for convenience rather than adding actual new 
> > > features. As noted above, I think it is important for browsers to 
> > > provide a core feature set but not to succumb to a feature creep where 
> > > they provide all the features that authors will ever need.
> > 
> > Standardized meta data is not for convenience, but for robustness and 
> > flexibility.
> 
> It's trivial to include metadata in text packets, and it can be done in 
> whatever way is most convenient for the type of data being sent. If a 
> simple code is more effective, that can be done. If each message needs a 
> URL and EXIF data transfered as an RDF block, that can be done. If 
> MIME-like headers are best, that can be done.

What do you recommend for a binary blob with metadata?  Two frames,
one for text metadata and one binary?

> > Surely MIME encapsulation has already proven itself enough to warrant 
> > such a small investment?
> 
> It has proven itself to be adequate (though by no means outstanding, IMHO) 
> for labeling resources. And it is quite possible to use MIME encapsulation 
> in a WebSocket subprotocol if that is desired. But why force it on 
> everyone? It's certainly not the be-all and end-all of metadata labeling, 
> and it isn't a feature everyone needs.

Given that, I'm still wondering why we're looking at a WebSocket
message framing that isn't a feature everyone needs.  Why doesn't it
go the connection handshake then provide a TCP stream?  The message
framing doesn't even provide anything useful that fits with future
transport protocols such as SCTP.

> > > On Thu, 16 Apr 2009, Greg Wilkins wrote:
> > > > I don't know if limiting connections on the server was a motivation 
> > > > for the 2 connection limit, but as a developer of a HTTP server, I 
> > > > appreciate the limit as every connection to the server consumes some 
> > > > kernel resources and buffers.
> > > 
> > > The Web Socket protocol spec has been updated to clarify that there is 
> > > no limit to the number of connections
> > 
> > Is there any analysis or evidence that this will not adversely affect 
> > TCP/IP flow control or server scalability?
> 
> No. If anyone is able to do such an analysis, I'd be very interested in 
> seeing the results. It should be noted though that load on the server 
> itself isn't a concern, since the author is the one starting the 
> connections and he can quite easily just multiplex everything using a 
> shared worker if so desired. Only the impact on the client and the network 
> are realy a concern.

Without a doubt, you get better end-to-end TCP/IP with fewer
connections provided both ends are consuming all the data sent to
them, both ends are multiplexing different message sources fairly, and
they have similar QoS requirements.  It's a simple matter of coherence
vs. incoherence in the packet stream, and of small messages batching
into larger packets.  Batching also means better throughput.

But we should be looking beyond TCP/IP, and into what SCTP/IP
provides.  If not now, then anticipating how to map onto it in future.
SCTP's main feature is the ability to have separate message streams
multiplexed over a single aggregated packet stream, handled properly
at the transport level - avoiding many of the problems with that over TCP.

> > What are the reasons that drove the limitation of HTTP connections that 
> > do not apply to websocket?
> 
> The reasons that drove the limitation of HTTP connections -- the ability 
> for untrusted third-party content to overload a server with many 
> connections -- DO apply to WebSocket, and WebSocket has been limited to 
> prevent such abuse (even more so than HTTP; WebSocket only allows one 
> pending non-established connection at a time).

I suspect that limitation will have to be ignored in practice.

If you have a web page with many active parts authored by different
people, and a connection setup round-trip time of 2 seconds (but you
could do 100 in parallel), you simply will not be happy when the page
takes 1 minute for all of it to appear.

-- Jamie

From fielding@gbiv.com  Fri Apr 24 14:42:37 2009
Return-Path: <fielding@gbiv.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DC2683A6C1A for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:42:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.37
X-Spam-Level: 
X-Spam-Status: No, score=-5.37 tagged_above=-999 required=5 tests=[AWL=-2.771,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GDzRvZGjVl4M for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:42:36 -0700 (PDT)
Received: from spaceymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 9A79A3A6B79 for <hybi@ietf.org>; Fri, 24 Apr 2009 14:41:11 -0700 (PDT)
Received: from [10.2.8.56] (wsip-70-183-62-251.oc.oc.cox.net [70.183.62.251]) by spaceymail-a3.g.dreamhost.com (Postfix) with ESMTP id B7738194E7B; Fri, 24 Apr 2009 14:42:30 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <49369C5C.1070609@gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>
Mime-Version: 1.0 (Apple Message framework v753.1)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <E9E990FF-C358-4E68-AAEF-DE0E20083AF8@gbiv.com>
Content-Transfer-Encoding: 7bit
From: "Roy T. Fielding" <fielding@gbiv.com>
Date: Fri, 24 Apr 2009 14:42:52 -0700
To: Ian Hickson <ian@hixie.ch>
X-Mailer: Apple Mail (2.753.1)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 21:42:38 -0000

On Apr 24, 2009, at 1:09 PM, Ian Hickson wrote:
> On Sat, 25 Apr 2009, Greg Wilkins wrote:
>> What are the reasons that drove the limitation of HTTP connections  
>> that
>> do not apply to websocket?
>
> The reasons that drove the limitation of HTTP connections -- the  
> ability
> for untrusted third-party content to overload a server with many
> connections -- DO apply to WebSocket, and WebSocket has been  
> limited to
> prevent such abuse (even more so than HTTP; WebSocket only allows one
> pending non-established connection at a time).

Umm, that's very creative history.

FTR, the limitation exists because of social engineering at the IETF.
Netscape introduced four parallel connections in late 1994.  It was
observed that the TCP overhead from four simultaneous connection
initializations, followed by often-aborted requests for images, were
causing the cross-Atlantic network links to become congested, and
Netscape's own servers were encountering TCP connection-queue limits.
Therefore, keep-alive and persistent connections were added to HTTP.

However, browser developers are notoriously user-centric (i.e.,  
selfish).
Instead of reducing the number of keep-alive connections, they deployed
four parallel long-lived connections in spite of the fact that two
connections performed better over congested or slow links.  Since that
behavior was considered harmful to the global commons of TCP, the
HTTP WG added a social requirement that no more than 2 persistent
connections be used to the same host.

All of the significant HTTP servers had internal connection
management of some form (hunt group, server pool, event-based, etc.)
by mid 1995, long before keep-alive and persistent connections existed.
Flash-crowds were a problem even before Navigator was released.

I don't know if the connection limit is still necessary.
My guess, judging from the amount poured into sites like hulu that
transmit HD over the net, is that the connection limit is no longer
needed by the backbone carriers.  Websites certainly don't care.

....Roy

From ian@hixie.ch  Fri Apr 24 14:48:03 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 931F23A67EB for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:48:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.583
X-Spam-Level: 
X-Spam-Status: No, score=-2.583 tagged_above=-999 required=5 tests=[AWL=-1.224, BAYES_00=-2.599, SARE_LWSHORTT=1.24]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 26P-L4Px8W9i for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 14:48:01 -0700 (PDT)
Received: from looneymail-a2.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 8FA8A3A68A8 for <hybi@ietf.org>; Fri, 24 Apr 2009 14:47:59 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a2.g.dreamhost.com (Postfix) with ESMTP id BD0DF16D40C; Fri, 24 Apr 2009 14:49:18 -0700 (PDT)
Date: Fri, 24 Apr 2009 21:49:18 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Jamie Lokier <jamie@shareable.org>
In-Reply-To: <20090424211314.GA18260@shareable.org>
Message-ID: <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>
References: <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 21:48:03 -0000

On Fri, 24 Apr 2009, Jamie Lokier wrote:
> 
> I asked myself the same thing a while ago: since every application has 
> to know the format of what it's going to receive _anyway_, in 
> considerable byte/character-level detail, and application-defined 
> sub-framing metadata (such as it's own sequence numbers, target object 
> ids or whatever), what is the purpose of the framing byte?

The DOM Events model is a push mechanism, so to get predictable 
interoperable behaviour, we have to have a way of defining the end of a 
block of content (either by length declaration or by a sentinel byte).

Right now we can only support Unicode data, as JS doesn't have binary data 
support yet, so on the short term we need a way to transfer text. On the 
longer term we need to be able to extend this to support binary data. Thus 
the two kinds of data. We use different framing for both because it is 
easier to support strings using a termination byte than a length 
declaration (measuring UTF-8 data is easy to get wrong in undetected ways, 
especially when testing only with ASCII), but it is easier to support 
length declarations than sentinels for binary data.


> Why is it not enough for the receiving API to offer 
> receiveBinaryFixed(length), receiveBinaryFrame() and 
> receiveTextUTF8Frame() methods?

It doesn't really fit the DOM model.


> > The protocol is designed with this in mind (in fact as currently 
> > designed, you'll notice that the client-side supports length-delimited 
> > binary packets, for future compatibility, but the server side doesn't, 
> > since it knows what types of packets it needs to support).
> 
> That's interesting.  If the protocol reaches mainstream browsers, I 
> imagine using the length-delimited binary packets if they work.
> 
> It will be a shame if it's "specified" but half the client 
> implementations are broken because that part was never tested.  I'd have 
> to have to encode binary data into UTF-8 *and then decode UTF-8 back 
> into binary in Javascript* just to make it work on real deployments. :-)

Browsers don't currently support binary data at the JS level. Once they 
do, we'll add the binary support to the API.


> > >  + Clients do connect to wrong ports.
> > 
> > This is one reason we have the handshake and why it is so strict.
> 
> It sounds like the reason could be explained more clearly in the 
> document, as the question shouldn't have come up.

Yeah, I've yet to write the intro section.


> I'm a bit concerned if it's so strict that proxies (including 
> transparent ones) will break it though.

It would be useful to do some research to see whether there are specific 
ways in which we can relax the handshake to allow for proxies, without 
substantially opening up the attack surface.


> > > Standardized metadata is a proven way that errors and differing versions 
> > > or interpretations can be handled, ignored or alarmed.
> > 
> > It's _a_ way, indeed. IMHO not a particularly successful one, for example 
> > look at the many ways in which HTTP status codes are ignored (e.g. <img> 
> > pointing to a 404 image is treated like a 200).
> 
> Eh?  Is it?  I've always seen a broken image icon, I'm pretty sure.

Not if the 404 response is an image.


> WebSocket is about messages.  HTTP (as a protocol, not a data model) is 
> about messages with some metadata and a request-response model. I'm not 
> seeing that much difference.

While there may be superficial similarities, I think one has to practice 
architectural astronautics to get the HTTP and WebSocket models to appear 
similar. It's like saying SMTP and FTP are the both "about messages".


> > It's trivial to include metadata in text packets, and it can be done 
> > in whatever way is most convenient for the type of data being sent. If 
> > a simple code is more effective, that can be done. If each message 
> > needs a URL and EXIF data transfered as an RDF block, that can be 
> > done. If MIME-like headers are best, that can be done.
> 
> What do you recommend for a binary blob with metadata?  Two frames, one 
> for text metadata and one binary?

Once we have binary data in JS, we will presumably also have a library of 
functions to convert text to binary and vice versa, so it would be simpler 
to just put the headers before the blob, as Greg suggested.


> > The reasons that drove the limitation of HTTP connections -- the 
> > ability for untrusted third-party content to overload a server with 
> > many connections -- DO apply to WebSocket, and WebSocket has been 
> > limited to prevent such abuse (even more so than HTTP; WebSocket only 
> > allows one pending non-established connection at a time).
> 
> I suspect that limitation will have to be ignored in practice.
> 
> If you have a web page with many active parts authored by different 
> people, and a connection setup round-trip time of 2 seconds (but you 
> could do 100 in parallel), you simply will not be happy when the page 
> takes 1 minute for all of it to appear.

That's a good thing, it will encourage multiplexing.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From gregw@webtide.com  Fri Apr 24 16:23:34 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E59343A6C83 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 16:23:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.557
X-Spam-Level: 
X-Spam-Status: No, score=-2.557 tagged_above=-999 required=5 tests=[AWL=0.042,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lzeFYNZwuYhH for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 16:23:34 -0700 (PDT)
Received: from mail-gx0-f166.google.com (mail-gx0-f166.google.com [209.85.217.166]) by core3.amsl.com (Postfix) with ESMTP id 09C963A697F for <hybi@ietf.org>; Fri, 24 Apr 2009 16:23:33 -0700 (PDT)
Received: by gxk10 with SMTP id 10so2847692gxk.13 for <hybi@ietf.org>; Fri, 24 Apr 2009 16:24:52 -0700 (PDT)
Received: by 10.90.52.1 with SMTP id z1mr3156361agz.38.1240615492333; Fri, 24 Apr 2009 16:24:52 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 6sm3403096agb.28.2009.04.24.16.24.49 (version=TLSv1/SSLv3 cipher=RC4-MD5); Fri, 24 Apr 2009 16:24:51 -0700 (PDT)
Message-ID: <49F24A3B.8080405@webtide.com>
Date: Sat, 25 Apr 2009 09:24:43 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Jamie Lokier <jamie@shareable.org>
References: <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com>	<4936C476.1010708@gmail.com>	<5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com>	<3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com>	<c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>	<Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>	<49F1DEF9.2000209@webtide.com>	<79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>	<Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org>
In-Reply-To: <20090424211314.GA18260@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2009 23:23:35 -0000

Jamie Lokier wrote:
> Ian Hickson wrote:
>>> An application could just assume the message contents are what is 
>>> expected.  Why have a frame type byte at all and both ends can just 
>>> assume the content and framing type!
>> That would be ideal, yes. The protocol is (IMHO) about as close as we can 
>> get to that right now.
> 
> Can you say a bit about why that isn't possible?
> 
> I asked myself the same thing a while ago: since every application has
> to know the format of what it's going to receive _anyway_, in
> considerable byte/character-level detail, and application-defined
> sub-framing metadata (such as it's own sequence numbers, target object
> ids or whatever), what is the purpose of the framing byte?
> 
> Why is it not enough for the receiving API to offer
> receiveBinaryFixed(length), receiveBinaryFrame() and
> receiveTextUTF8Frame() methods?


Indeed!

If websocket limited itself to being a minimal data framing protocol,
then it could be well suited to be the standard adopted by
firewalls,gateways and proxies to support upgraded HTTP connections.

But web socket also is providing support for a very specific
application - DOM based applications written in javascript
running withing a browser and communicating via untyped UTF-8
text datagrams.

Now this specific application may be suitable for a large
number of uses, but it is certainly not suitable for all.

Thus those that want to write by directional protocols using
a different base message type (eg gzipped utf-8 text
datagrams or arbitrary MIME encapsulated resources) are going
to ask - "why do a specific class of javascript DOM users get
favour favoured-son status?" and "why can't my packet type be
supported?"

Ian obviously feels that he does not want to require compliant
implementations to have to have support for MIME encapsulation
I assume because he can see a class of application that can be
written without it.   I likewise don't want to have my compliant
implementation support sentinel framed UTF-8 strings when the
class of protocol/application that I want to write would not
use such a limiting datagram.

It would seam that support for UTF-8 text datagrams has opened
indeed created the feature creep demand that Ian has previously
alluded to.   This is because by supporting 2 framing types,
websocket violates the 0, 1 or infinity principal and there
is demand for the infinity option.

It would be far better if websocket restricted itself to
1 or provided a standardized way to negotiate the infinite.

regards







From jamie@shareable.org  Fri Apr 24 17:59:57 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 90B143A6A55 for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 17:59:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.591
X-Spam-Level: 
X-Spam-Status: No, score=-2.591 tagged_above=-999 required=5 tests=[AWL=-1.547, BAYES_00=-2.599, SARE_LWSHORTT=1.24, SARE_MILLIONSOF=0.315]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id YWepQmqsKL0a for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 17:59:56 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 3DF013A6954 for <hybi@ietf.org>; Fri, 24 Apr 2009 17:59:56 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LxWGU-0006Ds-Ij; Sat, 25 Apr 2009 02:01:10 +0100
Date: Sat, 25 Apr 2009 02:01:10 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090425010110.GB23106@shareable.org>
References: <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 25 Apr 2009 00:59:57 -0000

Ian Hickson wrote:
> On Fri, 24 Apr 2009, Jamie Lokier wrote:
> > 
> > I asked myself the same thing a while ago: since every application has 
> > to know the format of what it's going to receive _anyway_, in 
> > considerable byte/character-level detail, and application-defined 
> > sub-framing metadata (such as it's own sequence numbers, target object 
> > ids or whatever), what is the purpose of the framing byte?
> 
> The DOM Events model is a push mechanism, so to get predictable 
> interoperable behaviour, we have to have a way of defining the end of a 
> block of content (either by length declaration or by a sentinel byte).

That makes no sense.

If I call ws.getTextUTF8Frame() then of course you can read a UTF-8
sentinel-terminated string and send me the event when it's done.

If I call ws.getBinary(1024) then you receive 1024 bytes and send me
the event when that much is received.

Nothing about "push mechanism" or DOM Events requires the base
protocol to have framing.

> Right now we can only support Unicode data, as JS doesn't have binary data 
> support yet, so on the short term we need a way to transfer text. On the 
> longer term we need to be able to extend this to support binary data. Thus 
> the two kinds of data.

Fine, I accept the need for a text option for Javascript compatibility.

(But I've got a horrible feeling non-JS client apps are going to be
sending binary packets encoded with gzip and then transcoded into a
UTF-8 compatible octet sequence to run over JS-compatible infrastructure.)

> We use different framing for both because it is 
> easier to support strings using a termination byte than a length 
> declaration (measuring UTF-8 data is easy to get wrong in undetected ways, 
> especially when testing only with ASCII), but it is easier to support 
> length declarations than sentinels for binary data.

But you haven't given a good reason why the protocol has to be
self-describing in that way (with a frame type byte).  The app can say
what it's expecting, and you can read the framing the app asks for.
It's easy to use, and efficient.

That's what all other programming languages do when reading from sockets!

> > Why is it not enough for the receiving API to offer 
> > receiveBinaryFixed(length), receiveBinaryFrame() and 
> > receiveTextUTF8Frame() methods?
> 
> It doesn't really fit the DOM model.

I think (forget..) the DOM model for XmlHttpRequest is that you get an
event telling you something has happened, then you read whatever state
is available from the XHR object.

If WebSockets did the same, you'd get an event when data is available
to read, then you'd read as much as you can with those functions until
you got an error saying "not enough available", then you'd get another
event when more data is available.

That sounds both simpler to use in Javascript and more efficient with
rapid streams of small messages than an event per message.

> > I'm a bit concerned if it's so strict that proxies (including 
> > transparent ones) will break it though.
> 
> It would be useful to do some research to see whether there are specific 
> ways in which we can relax the handshake to allow for proxies, without 
> substantially opening up the attack surface.

One way that's very likely to work is over SSL.

Another is to embed the stream inside a pair of streaming HTTP
requests, as AJAX/Comet apps do.  Another is to use long-GET, for
maximum compatibility.

If you embed in a way which sufficiently advanced proxies are
officially allowed to "unembed", you can be compatible with existing
infrastructure at the same time as allowing an upgrade path to getting
rid of the embedding.  (That's what I'm looking at for bidirectional
multiplexed cached HTTP over existing infrastructure).

That's all very ugly, I know.  But very compatible, in a way that I'm
not convinced the WebSockets handshake will be for practical
internet-sized deployments.

> > WebSocket is about messages.  HTTP (as a protocol, not a data model) is 
> > about messages with some metadata and a request-response model. I'm not 
> > seeing that much difference.
> 
> While there may be superficial similarities, I think one has to practice 
> architectural astronautics to get the HTTP and WebSocket models to appear 
> similar. It's like saying SMTP and FTP are the both "about messages".

Can't entirely disagree :-)

> > > The reasons that drove the limitation of HTTP connections -- the 
> > > ability for untrusted third-party content to overload a server with 
> > > many connections -- DO apply to WebSocket, and WebSocket has been 
> > > limited to prevent such abuse (even more so than HTTP; WebSocket only 
> > > allows one pending non-established connection at a time).
> > 
> > I suspect that limitation will have to be ignored in practice.
> > 
> > If you have a web page with many active parts authored by different 
> > people, and a connection setup round-trip time of 2 seconds (but you 
> > could do 100 in parallel), you simply will not be happy when the page 
> > takes 1 minute for all of it to appear.
> 
> That's a good thing, it will encourage multiplexing.

If multiplexing is good, why aren't we making it the trivial default option?

I.e. if there was an API which does multiplexing by default, but where
you could ask for a dedicated connection easily and get identical
behaviour except over a dedicated connection - wouldn't that encourage
multiplexing even better than non-obvious networking characteristics?

You must have noticed that most web developers don't know much about
network performance and aggregate behaviour beyond keeping the size of
pages small enough for some target audience, and which browsers to test.

I'm not saying we should optimise everything or bake in lots of
complexity.  I like that WebSockets aims for simplicity.

But if it were as trivial as a flag to the WS constructor meaning
"give me a multiplexed connection" vs. "give me a dedicated
connection", that would go a long way to improving overall system
behaviour in the presence of millions of web developers who don't know
much about networking, I suspect.

I know that's an unproven data point, and would need measurements
before committing to it, but please at least consider the possibility,
then we can talk about what it would look like in detail.

-- Jamie

From jamie@shareable.org  Fri Apr 24 18:18:36 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CFB0E3A6BFF for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 18:18:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.311
X-Spam-Level: 
X-Spam-Status: No, score=-3.311 tagged_above=-999 required=5 tests=[AWL=-0.712, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CiqLNG857gMK for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 18:18:36 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id D555A3A6A60 for <hybi@ietf.org>; Fri, 24 Apr 2009 18:18:35 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LxWYZ-0006I4-5n; Sat, 25 Apr 2009 02:19:51 +0100
Date: Sat, 25 Apr 2009 02:19:51 +0100
From: Jamie Lokier <jamie@shareable.org>
To: "Roy T. Fielding" <fielding@gbiv.com>
Message-ID: <20090425011951.GC23106@shareable.org>
References: <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <E9E990FF-C358-4E68-AAEF-DE0E20083AF8@gbiv.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <E9E990FF-C358-4E68-AAEF-DE0E20083AF8@gbiv.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 25 Apr 2009 01:18:36 -0000

Roy T. Fielding wrote:
> they deployed
> four parallel long-lived connections in spite of the fact that two
> connections performed better over congested or slow links.  Since that
> behavior was considered harmful to the global commons of TCP, the
> HTTP WG added a social requirement that no more than 2 persistent
> connections be used to the same host.

That right there points at a reason to be cautious, at least, about
numbers of connections over congested or slow links.

I'll add that the "many connections but maximum one outstanding new
connection" rule from WebSockets will behave poorly with high latency
links in addition to congested and slow ones.

Quite simply, many connections to the same site is poor use of TCP.
This might not be obvious to anyone who sees TCP as a simple byte
pipe, and isn't aware of what TCP has to do with packets and what
makes TCP perform better or worse.

If a site works well over high-speed broadband with just one user, and
sucks over dialup, that's not our problem.

But web app developers can't be expected to know the finer points of
transport protocol behaviour.  As infrastructure engineers we can make
sites suck more, or less, by our choices at that level.

> I don't know if the connection limit is still necessary.
> My guess, judging from the amount poured into sites like hulu that
> transmit HD over the net, is that the connection limit is no longer
> needed by the backbone carriers.  Websites certainly don't care.

I think you're right about the backbone.

Large numbers of long-lived (long-GET) connections are a problem for
load-balancing reverse proxies at not-so-high-profile web sites,
according to another posting about Comet deployment.

I've seen quite a lot of posts around the net about how Comet
(i.e. long-GET or streaming-GET) is very bad for web servers, and
cannot ever be as efficient as polling every few seconds with Ajax
requests.

Of course this is nonsense to anyone with half a computer science
degree, but the belief seems to come from a lot of server-side
infrastructure spawning threads, processes and other expensive things
per connection.

At the client end, those slow or congested links are still quite common,
and multiple computers congesting a shared home link are getting more
common (I suspect).

-- Jamie

From ian@hixie.ch  Fri Apr 24 18:22:09 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BFBE03A6C0A for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 18:22:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.183
X-Spam-Level: 
X-Spam-Status: No, score=-3.183 tagged_above=-999 required=5 tests=[AWL=-0.584, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4SXhr8DnOT2Z for <hybi@core3.amsl.com>; Fri, 24 Apr 2009 18:22:08 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id A33953A6DAB for <hybi@ietf.org>; Fri, 24 Apr 2009 18:21:34 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id A305027B3E; Fri, 24 Apr 2009 18:22:53 -0700 (PDT)
Date: Sat, 25 Apr 2009 01:22:53 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Jamie Lokier <jamie@shareable.org>
In-Reply-To: <20090425010110.GB23106@shareable.org>
Message-ID: <Pine.LNX.4.62.0904250113521.12381@hixie.dreamhostps.com>
References: <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 25 Apr 2009 01:22:09 -0000

On Sat, 25 Apr 2009, Jamie Lokier wrote:
> 
> If I call ws.getTextUTF8Frame() then of course you can read a UTF-8 
> sentinel-terminated string and send me the event when it's done.
> 
> If I call ws.getBinary(1024) then you receive 1024 bytes and send me the 
> event when that much is received.
> 
> Nothing about "push mechanism" or DOM Events requires the base protocol 
> to have framing.

I guess I don't understand what you're proposing. Could you describe your 
proposal in more detail?


> (But I've got a horrible feeling non-JS client apps are going to be 
> sending binary packets encoded with gzip and then transcoded into a 
> UTF-8 compatible octet sequence to run over JS-compatible 
> infrastructure.)

Non-JS client apps would presumably not use WebSocket. The whole point of 
WebSocket is to provide something optimised for Web apps only. A native IM 
client, for example, would just use XMPP.


> > It would be useful to do some research to see whether there are 
> > specific ways in which we can relax the handshake to allow for 
> > proxies, without substantially opening up the attack surface.
> 
> One way that's very likely to work is over SSL.

Indeed.


> Another is to embed the stream inside a pair of streaming HTTP requests, 
> as AJAX/Comet apps do.  Another is to use long-GET, for maximum 
> compatibility.

Those techniques are already possible, and WebSocket isn't attempting to 
remove that. The idea if just to provide another option that uses just a 
single TCP connection ("TCP for the Web").


> > > If you have a web page with many active parts authored by different 
> > > people, and a connection setup round-trip time of 2 seconds (but you 
> > > could do 100 in parallel), you simply will not be happy when the 
> > > page takes 1 minute for all of it to appear.
> >
> > That's a good thing, it will encourage multiplexing.
> 
> If multiplexing is good, why aren't we making it the trivial default 
> option?

Generally speaking I am strongly against the idea of the Web platform APIs 
doing things for convenience rather than adding actual new features. As 
noted in earlier e-mails, I think it is important for browsers to provide 
a core feature set but not to succomb to a feature creep where they 
provide all the features that authors will ever need.

Christopher Blizzard of Mozilla said it in the context of O3D recently:

   http://www.0xdeadbeef.com/weblog/?p=1223

I think his point applies very much here.


> I.e. if there was an API which does multiplexing by default, but where 
> you could ask for a dedicated connection easily and get identical 
> behaviour except over a dedicated connection

I would be interested in seeing a concrete proposal. My impression is that 
this would be a non-trivial addition to the core protocol and API, and so 
I would prefer to not add it to the core building blocks.


> - wouldn't that encourage multiplexing even better than non-obvious 
> networking characteristics?

One minute of latency is pretty obvious.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From sh@defuze.org  Sat Apr 25 11:25:36 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C40943A6A33 for <hybi@core3.amsl.com>; Sat, 25 Apr 2009 11:25:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.215
X-Spam-Level: 
X-Spam-Status: No, score=-2.215 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IgiXpk7CLCLo for <hybi@core3.amsl.com>; Sat, 25 Apr 2009 11:25:35 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 9A2283A6B5C for <hybi@ietf.org>; Sat, 25 Apr 2009 11:25:35 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3PIQphJ005299; Sat, 25 Apr 2009 13:26:52 -0500
Message-ID: <49F355EA.3010101@defuze.org>
Date: Sat, 25 Apr 2009 20:26:50 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Jamie Lokier <jamie@shareable.org>
References: <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com>	<3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com>	<c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>	<Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>	<49F1DEF9.2000209@webtide.com>	<79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>	<Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>	<20090424211314.GA18260@shareable.org>	<Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org>
In-Reply-To: <20090425010110.GB23106@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 25 Apr 2009 18:25:36 -0000

Jamie Lokier wrote:
> That makes no sense.
>
> If I call ws.getTextUTF8Frame() then of course you can read a UTF-8
> sentinel-terminated string and send me the event when it's done.
>
> If I call ws.getBinary(1024) then you receive 1024 bytes and send me
> the event when that much is received.
>
> Nothing about "push mechanism" or DOM Events requires the base
> protocol to have framing.
>   
>
> But you haven't given a good reason why the protocol has to be
> self-describing in that way (with a frame type byte).  The app can say
> what it's expecting, and you can read the framing the app asks for.
> It's easy to use, and efficient.
>
> That's what all other programming languages do when reading from sockets!
>
>   

In that case, do we even need a new protocol?

- Sylvain

From gregw@webtide.com  Sun Apr 26 07:11:07 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2105B3A6F00 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 07:11:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.912
X-Spam-Level: 
X-Spam-Status: No, score=-1.912 tagged_above=-999 required=5 tests=[AWL=-0.605, BAYES_00=-2.599, MISSING_HEADERS=1.292]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Rch+AiAsAbPW for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 07:11:06 -0700 (PDT)
Received: from mail-qy0-f132.google.com (mail-qy0-f132.google.com [209.85.221.132]) by core3.amsl.com (Postfix) with ESMTP id 2DA083A6A9B for <hybi@ietf.org>; Sun, 26 Apr 2009 07:10:42 -0700 (PDT)
Received: by qyk38 with SMTP id 38so1357305qyk.29 for <hybi@ietf.org>; Sun, 26 Apr 2009 07:12:00 -0700 (PDT)
Received: by 10.224.20.77 with SMTP id e13mr4706740qab.186.1240755120688; Sun, 26 Apr 2009 07:12:00 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 10sm4105924qyk.57.2009.04.26.07.11.58 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sun, 26 Apr 2009 07:12:00 -0700 (PDT)
Message-ID: <49F46BA5.6040506@webtide.com>
Date: Mon, 27 Apr 2009 00:11:49 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
CC: hybi@ietf.org
References: <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com>	<3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com>	<c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>	<Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>	<49F1DEF9.2000209@webtide.com>	<79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>	<Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>	<20090424211314.GA18260@shareable.org>	<Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>	<20090425010110.GB23106@shareable.org> <Pine.LNX.4.62.0904250113521.12381@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904250113521.12381@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 14:11:07 -0000

Ian Hickson wrote:

> Non-JS client apps would presumably not use WebSocket. 

Why?

It seams really strange to propose an IETF protocol that
is so heavily  influenced by one programming language used
in one particular application.


 > The whole point of WebSocket is to provide something optimised
 > for Web apps only.

Which is why it might not be a protocol that all can get behind and
give it the support that it needs to be accepted as a standard and
then implemented widely.

 > A native IM  client, for example, would just use XMPP.

A native IM client still has the problem of tunneling through
firewalls, proxies and gateways.   Thus they have to tunnel
XMPP over BOSH, which uses all the long polling techniques
over HTTP.

I'm sure XMPP client developers would prefer to replace BOSH
with an HTTP upgrade protocol designed for bidirection usage over
a single socket that can legally tunnel through firewalls.
If the don't, that implies that HTTP is sufficient for there
needs thus it is also likely that it is sufficient for web
applications.

Indeed, cometd/bayeux is moderately happy with long polling
over HTTP and websocket as it is currently formulated
does little or nothing to address the issues that we have.

But websocket is very close to something that could help
and could be made more widely applicable with some small
changes.

I really think those that are proposing websocket should
consider widening the scope of the proposal that it includes
the interests and concerns that have been expressed on this list.

It might be perfectly suited to the webapplications that some
want to write, but unless it can attract support from a wider
audience I doubt that will receive the support it needs to
be adopted.


regards


From ian@hixie.ch  Sun Apr 26 11:00:05 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AC5693A6D0F for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 11:00:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.174
X-Spam-Level: 
X-Spam-Status: No, score=-3.174 tagged_above=-999 required=5 tests=[AWL=-0.575, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tYXDp99+W02S for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 10:59:58 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id B15393A6B24 for <hybi@ietf.org>; Sun, 26 Apr 2009 10:59:58 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 6576227B68; Sun, 26 Apr 2009 11:01:18 -0700 (PDT)
Date: Sun, 26 Apr 2009 18:01:17 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Greg Wilkins <gregw@webtide.com>
In-Reply-To: <49F46BA5.6040506@webtide.com>
Message-ID: <Pine.LNX.4.62.0904261701440.12381@hixie.dreamhostps.com>
References: <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <Pine.LNX.4.62.0904250113521.12381@hixie.dreamhostps.com> <49F46BA5.6040506@webtide.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 18:00:05 -0000

On Mon, 27 Apr 2009, Greg Wilkins wrote:
> Ian Hickson wrote:
> 
> > Non-JS client apps would presumably not use WebSocket.
> 
> Why?

Because the whole point is to provide a protocol that is safe for Web 
pages. If you have native access to TCP, I don't see any reason why you'd 
using WebSocket instead of just using a more appropriate protocol or using 
a custom protocol over TCP. The security aspects of WebSocket make no 
sense in a situation where the program has control over the connection.


> It seams really strange to propose an IETF protocol that is so heavily 
> influenced by one programming language used in one particular 
> application.

I agree. This wasn't originally going to be proposed to the IETF, it was 
just a subsection of the WebSocket API as part of HTML5. The IETF 
requested that it be split out into a dedicated RFC. If this group doesn't 
think we should develop the protocol separate from the API, I'm happy to 
move them back into the same document. What is the group's opinion?


> But websocket is very close to something that could help and could be 
> made more widely applicable with some small changes.

IMHO it doesn't make sense to try to abstract things out to solve two 
completely unrelated use cases (native IM apps and a protocol for JS 
running in browsers). I think it would be much better for dedicated 
solutions to be designed for each case. Trying to design abstract 
protocols that apply to unrelated use cases usually results in compromises 
that don't work well for any of the supposed use cases.

Having said that, I have no objection to us developing a generic protocol 
for tunneling other protocols over HTTP or upgrading from HTTP to other 
protocols, separate from WebSocket. Is anyone interested in editing such a 
specification? If so, I say we go for it.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From jamie@shareable.org  Sun Apr 26 12:37:17 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0F70F3A6E4A for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 12:37:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.286
X-Spam-Level: 
X-Spam-Status: No, score=-3.286 tagged_above=-999 required=5 tests=[AWL=-0.687, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xwJmTvAHa6HX for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 12:37:16 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 1DCA53A67CC for <hybi@ietf.org>; Sun, 26 Apr 2009 12:37:15 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LyABM-0002tm-QO; Sun, 26 Apr 2009 20:38:32 +0100
Date: Sun, 26 Apr 2009 20:38:32 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090426193832.GA10627@shareable.org>
References: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <Pine.LNX.4.62.0904250113521.12381@hixie.dreamhostps.com> <49F46BA5.6040506@webtide.com> <Pine.LNX.4.62.0904261701440.12381@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904261701440.12381@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 19:37:17 -0000

Ian Hickson wrote:
> On Mon, 27 Apr 2009, Greg Wilkins wrote:
> > Ian Hickson wrote:
> > 
> > > Non-JS client apps would presumably not use WebSocket.
> > 
> > Why?
> 
> Because the whole point is to provide a protocol that is safe for Web 
> pages. If you have native access to TCP, I don't see any reason why you'd 
> using WebSocket instead of just using a more appropriate protocol or using 
> a custom protocol over TCP. The security aspects of WebSocket make no 
> sense in a situation where the program has control over the connection.

Once there is substantial deployment, the reason you'll see WebSocket
clients in other languages is that will be the only way to access some
web-based services that you don't control.

> > It seams really strange to propose an IETF protocol that is so heavily 
> > influenced by one programming language used in one particular 
> > application.
> 
> I agree. This wasn't originally going to be proposed to the IETF, it was 
> just a subsection of the WebSocket API as part of HTML5. The IETF 
> requested that it be split out into a dedicated RFC. If this group doesn't 
> think we should develop the protocol separate from the API, I'm happy to 
> move them back into the same document. What is the group's opinion?

I can't speak for the group.  My opinion is the wire protocol should
be kept very separate from the Javascript API for talking it.

The API should be designed entirely around whatever works best in
Javascript: be simple and easy to use for Javascript programmers.
If it's good, the API will survive a few changes of wire protocol.

And the wire protocol should take note of input from network engineers ;-)

> Having said that, I have no objection to us developing a generic protocol 
> for tunneling other protocols over HTTP or upgrading from HTTP to other 
> protocols, separate from WebSocket. Is anyone interested in editing such a 
> specification? If so, I say we go for it.

There are a number of ways, as many here know, to tunnel one-way and
two-way messages over HTTP.  Some are more reliable than others.  Some
work over HTTP protocol, but poorly or not at all over realistic
infrastructure.

I do think it would be good to put the known (useful) methods of
tunnelling into a concise, RFC-style document, and from there propose
best-practice methods for tunnelling over the existing and future web.

WebSockets could reference that, and WebSockets itself could be
standardised and implemented while the tunnelling strategy continues
to be refined based on real experience.  The WebSockets wire protocol
need not change, but strategies for tunnelling it might need to.

So count me as voting for separate documents on:

    - Javascript API for WebSockets
    - WebSockets wire protocol
    - Generic two-way message/stream tunnelling over HTTP

Each can remain small and focused on its domain.

-- Jamie

From jamie@shareable.org  Sun Apr 26 12:46:10 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CFA003A6C31 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 12:46:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.262
X-Spam-Level: 
X-Spam-Status: No, score=-3.262 tagged_above=-999 required=5 tests=[AWL=-0.663, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nmR1Qo8RTdNy for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 12:46:10 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 07B9B3A67CC for <hybi@ietf.org>; Sun, 26 Apr 2009 12:45:44 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LyAJZ-0002y1-F9; Sun, 26 Apr 2009 20:47:01 +0100
Date: Sun, 26 Apr 2009 20:47:01 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Sylvain Hellegouarch <sh@defuze.org>
Message-ID: <20090426194701.GB10627@shareable.org>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49F355EA.3010101@defuze.org>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 19:46:10 -0000

Sylvain Hellegouarch wrote:
> Jamie Lokier wrote:
> >That makes no sense.
> >
> >If I call ws.getTextUTF8Frame() then of course you can read a UTF-8
> >sentinel-terminated string and send me the event when it's done.
> >
> >If I call ws.getBinary(1024) then you receive 1024 bytes and send me
> >the event when that much is received.
> >
> >Nothing about "push mechanism" or DOM Events requires the base
> >protocol to have framing.
> >  
> >
> >But you haven't given a good reason why the protocol has to be
> >self-describing in that way (with a frame type byte).  The app can say
> >what it's expecting, and you can read the framing the app asks for.
> >It's easy to use, and efficient.
> >
> >That's what all other programming languages do when reading from sockets!
> >
> >  
> 
> In that case, do we even need a new protocol?

No we don't while it's running.  TCP would be fine from this
perspective, and (imho) preferable as you could use it more easily
with any TCP services, just relaying after the connection setup.

But there are issues with security, service addressing, and operating
over port 80 (http) though, which WebSockets deals with in the
connection setup.  Those may be important for web pages establishing
sockets.

Flash in web pages can use plain TCP.  I'm not sure what the
restrictions are.

WebSockets uses port 80 to work over real infrastructure where other
ports may be firewalled, but unfortunately I suspect the "Upgrade:"
header isn't adequate for a significant minority of real
infrastructure in the way, nor the 101 response.  It's hard to know
without real measurements...  (Has anyone done such measurements?)

-- Jamie

From sh@defuze.org  Sun Apr 26 13:56:09 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5D79B28C0F3 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 13:56:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.219
X-Spam-Level: 
X-Spam-Status: No, score=-2.219 tagged_above=-999 required=5 tests=[AWL=0.046,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fBEPI85cd4hT for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 13:56:08 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id C593428C0FA for <hybi@ietf.org>; Sun, 26 Apr 2009 13:55:55 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3QKv8FR003785; Sun, 26 Apr 2009 15:57:13 -0500
Message-ID: <49F4CAA4.6020602@defuze.org>
Date: Sun, 26 Apr 2009 22:57:08 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Jamie Lokier <jamie@shareable.org>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org>
In-Reply-To: <20090426194701.GB10627@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 20:56:09 -0000

Jamie Lokier a écrit :
> Sylvain Hellegouarch wrote:
>   
>> Jamie Lokier wrote:
>>     
>>> That makes no sense.
>>>
>>> If I call ws.getTextUTF8Frame() then of course you can read a UTF-8
>>> sentinel-terminated string and send me the event when it's done.
>>>
>>> If I call ws.getBinary(1024) then you receive 1024 bytes and send me
>>> the event when that much is received.
>>>
>>> Nothing about "push mechanism" or DOM Events requires the base
>>> protocol to have framing.
>>>  
>>>
>>> But you haven't given a good reason why the protocol has to be
>>> self-describing in that way (with a frame type byte).  The app can say
>>> what it's expecting, and you can read the framing the app asks for.
>>> It's easy to use, and efficient.
>>>
>>> That's what all other programming languages do when reading from sockets!
>>>
>>>  
>>>       
>> In that case, do we even need a new protocol?
>>     
>
> No we don't while it's running.  TCP would be fine from this
> perspective, and (imho) preferable as you could use it more easily
> with any TCP services, just relaying after the connection setup.
>
> But there are issues with security, service addressing, and operating
> over port 80 (http) though, which WebSockets deals with in the
> connection setup.  Those may be important for web pages establishing
> sockets.
>   
Right, so I second you when you suggest we should clearly separate the 
transport protocol from the application's level one and indeed maybe 
from the transport protocol's perspective, all WebSockets has to define 
is the security measure to take to secure a connection and once it's 
done, let a socket be just that, a socket.

In other words, why do we need the framing part of the protocol? Ian?

- Sylvain

From ian@hixie.ch  Sun Apr 26 14:04:01 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E2D213A69D3 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:04:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.165
X-Spam-Level: 
X-Spam-Status: No, score=-3.165 tagged_above=-999 required=5 tests=[AWL=-0.566, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PB9EzE257wp9 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:04:01 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id 20C203A69E4 for <hybi@ietf.org>; Sun, 26 Apr 2009 14:04:01 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 8D477125C84; Sun, 26 Apr 2009 14:05:21 -0700 (PDT)
Date: Sun, 26 Apr 2009 21:05:21 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Sylvain Hellegouarch <sh@defuze.org>
In-Reply-To: <49F4CAA4.6020602@defuze.org>
Message-ID: <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 21:04:02 -0000

On Sun, 26 Apr 2009, Sylvain Hellegouarch wrote:
>
> Right, so I second you when you suggest we should clearly separate the 
> transport protocol from the application's level one and indeed maybe 
> from the transport protocol's perspective, all WebSockets has to define 
> is the security measure to take to secure a connection and once it's 
> done, let a socket be just that, a socket.
> 
> In other words, why do we need the framing part of the protocol? Ian?

Do these e-mails answer your question?:

   http://www.ietf.org/mail-archive/web/hybi/current/msg00248.html
   http://www.ietf.org/mail-archive/web/hybi/current/msg00252.html

I'm open to other mechanisms, but it's not clear to me what they would be. 
Concrete API proposals would be good here.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From jamie@shareable.org  Sun Apr 26 14:12:06 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D1EEB3A6868 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:12:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.24
X-Spam-Level: 
X-Spam-Status: No, score=-3.24 tagged_above=-999 required=5 tests=[AWL=-0.641,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FvGeLBW7FD1O for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:12:06 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id CB9F83A69E4 for <hybi@ietf.org>; Sun, 26 Apr 2009 14:12:05 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LyBf8-0003fB-IU; Sun, 26 Apr 2009 22:13:22 +0100
Date: Sun, 26 Apr 2009 22:13:22 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Sylvain Hellegouarch <sh@defuze.org>
Message-ID: <20090426211322.GE10627@shareable.org>
References: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <49F4CAA4.6020602@defuze.org>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 21:12:06 -0000

Sylvain Hellegouarch wrote:
> Jamie Lokier a écrit :
> >Sylvain Hellegouarch wrote:
> >  
> >>Jamie Lokier wrote:
> >>    
> >>>That makes no sense.
> >>>
> >>>If I call ws.getTextUTF8Frame() then of course you can read a UTF-8
> >>>sentinel-terminated string and send me the event when it's done.
> >>>
> >>>If I call ws.getBinary(1024) then you receive 1024 bytes and send me
> >>>the event when that much is received.
> >>>
> >>>Nothing about "push mechanism" or DOM Events requires the base
> >>>protocol to have framing.
> >>> 
> >>>
> >>>But you haven't given a good reason why the protocol has to be
> >>>self-describing in that way (with a frame type byte).  The app can say
> >>>what it's expecting, and you can read the framing the app asks for.
> >>>It's easy to use, and efficient.
> >>>
> >>>That's what all other programming languages do when reading from sockets!
> >>>
> >>> 
> >>>      
> >>In that case, do we even need a new protocol?
> >>    
> >
> >No we don't while it's running.  TCP would be fine from this
> >perspective, and (imho) preferable as you could use it more easily
> >with any TCP services, just relaying after the connection setup.
> >
> >But there are issues with security, service addressing, and operating
> >over port 80 (http) though, which WebSockets deals with in the
> >connection setup.  Those may be important for web pages establishing
> >sockets.
> >  
> Right, so I second you when you suggest we should clearly separate the 
> transport protocol from the application's level one and indeed maybe 
> from the transport protocol's perspective, all WebSockets has to define 
> is the security measure to take to secure a connection and once it's 
> done, let a socket be just that, a socket.
> 
> In other words, why do we need the framing part of the protocol? Ian?

As far as I can tell, the current WebSockets Javascript/DOM API
requires the protocol to have self-describing frames, because that's
how it knows when a message is received, then it delivers the message.

So it's a property of the API requiring that from the protocol.

Fine, if it really does make the API work better.  But as I said
earlier, there are other approaches to the API which still work with
the DOM "push mechanism" and don't have the protocol framing
requirement, and will then work with general TCP services more easily.

To be honest I do think framing is a good idea - splitting bytes in
Javascript would be messy (it doesn't have byte arrays) and message
frames is useful.  But I don't see the point of the type byte except
for the "future extension" of byte frames in Javascript.  All the
_real_ type information has to be application defined inside each frame.

So if you're going to have a future extension which is not to be
implemented in the first version and so cannot be used except when
explicitly requested, you might as well make it more general - plan
for accepting any byte stream, when/if byte handling is added to
Javascript.

-- Jamie

From jamie@shareable.org  Sun Apr 26 14:12:50 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6A64B3A68ED for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:12:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.219
X-Spam-Level: 
X-Spam-Status: No, score=-3.219 tagged_above=-999 required=5 tests=[AWL=-0.620, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hzJqmfKPzZBy for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:12:49 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id ACA5A3A6868 for <hybi@ietf.org>; Sun, 26 Apr 2009 14:12:49 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LyBfr-0003fe-TJ; Sun, 26 Apr 2009 22:14:07 +0100
Date: Sun, 26 Apr 2009 22:14:07 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090426211407.GF10627@shareable.org>
References: <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org> <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 21:12:50 -0000

Ian Hickson wrote:
> Concrete API proposals would be good here.

I agree they'd help a lot, and I'll take a go at it when I have time,
in a few days.

-- Jamie

From sh@defuze.org  Sun Apr 26 14:46:40 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A10383A6951 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:46:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.222
X-Spam-Level: 
X-Spam-Status: No, score=-2.222 tagged_above=-999 required=5 tests=[AWL=0.043,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3IJjgrWwHcN4 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 14:46:39 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id C3D013A6AB4 for <hybi@ietf.org>; Sun, 26 Apr 2009 14:46:39 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3QLlvDC018803; Sun, 26 Apr 2009 16:47:58 -0500
Message-ID: <49F4D68D.9060100@defuze.org>
Date: Sun, 26 Apr 2009 23:47:57 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org> <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 21:46:40 -0000

Ian Hickson a écrit :
> On Sun, 26 Apr 2009, Sylvain Hellegouarch wrote:
>   
>> Right, so I second you when you suggest we should clearly separate the 
>> transport protocol from the application's level one and indeed maybe 
>> from the transport protocol's perspective, all WebSockets has to define 
>> is the security measure to take to secure a connection and once it's 
>> done, let a socket be just that, a socket.
>>
>> In other words, why do we need the framing part of the protocol? Ian?
>>     
>
> Do these e-mails answer your question?:
>
>    http://www.ietf.org/mail-archive/web/hybi/current/msg00248.html
>    http://www.ietf.org/mail-archive/web/hybi/current/msg00252.html
>
> I'm open to other mechanisms, but it's not clear to me what they would be. 
> Concrete API proposals would be good here.
>
>   
Well they triggered my question in fact (as per Jamie's emails too). If 
I recap, as I understand it, there's been a requirement to describe and 
define the best solution for long lived connections from browsers. That 
requirement has two main constraints and a feature.

The first constraint is to have a web oriented security model.
The second constraint is to be able to re-use the existing web 
infrastructure hence the HTTP bootstrap.

The feature is to allow for an API that plays well with both DOM and 
Javascript.


There are currently two drafts that deal each with one side of the 
overall problem at hand. One at the transport level describing the 
bootstrap mechanism and how to frame data. Another one at the browser 
level explaining how to consume and produce data frames.

I'm still wondering then why not keeping the bootstrap mechanism to 
preserve the security policy but strip out the framing part and redesign 
the Javascript API so that it acts like any other socket library in any 
existing programming language? In other words, let developers directly 
create Javascripts API that implement any given protocol.

You mentioned the push aspect of the DOM event model as well as the fact 
Javascript currently doesn't have binary support yet. I think it's 
taking the problem the wrong way, "we want to have this JS API because 
it fits us well so we need a protocol that maps to that API". Why not 
saying instead, "now that we have a secured connection, let's create a 
Javascript socket API and let developers play with it". I can't offer 
better API than:

open()
read(size)
write(data)
close()

Maybe Javascript lacks some aspects of other modern languages like 
binary support, is that rationale enough for creating a new framing 
protocol?

- Sylvain

From sh@defuze.org  Sun Apr 26 15:01:15 2009
Return-Path: <sh@defuze.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 052C43A6D79 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 15:01:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.225
X-Spam-Level: 
X-Spam-Status: No, score=-2.225 tagged_above=-999 required=5 tests=[AWL=0.040,  BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MPf56zYdDEo3 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 15:01:14 -0700 (PDT)
Received: from mail1.webfaction.com (mail1.webfaction.com [67.15.2.85]) by core3.amsl.com (Postfix) with ESMTP id 477A93A6D0E for <hybi@ietf.org>; Sun, 26 Apr 2009 15:01:14 -0700 (PDT)
Received: from [82.229.61.197] (localhost.localdomain [127.0.0.1]) (authenticated bits=0) by mail1.webfaction.com (8.12.11.20060308/8.13.3) with ESMTP id n3QM2S3L022576; Sun, 26 Apr 2009 17:02:30 -0500
Message-ID: <49F4D9F4.8070100@defuze.org>
Date: Mon, 27 Apr 2009 00:02:28 +0200
From: Sylvain Hellegouarch <sh@defuze.org>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: Ian Hickson <ian@hixie.ch>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com>	<3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com>	<Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>	<49F1DEF9.2000209@webtide.com>	<79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>	<Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>	<20090424211314.GA18260@shareable.org>	<Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>	<20090425010110.GB23106@shareable.org>	<49F355EA.3010101@defuze.org>	<20090426194701.GB10627@shareable.org>	<49F4CAA4.6020602@defuze.org>	<Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com> <49F4D68D.9060100@defuze.org>
In-Reply-To: <49F4D68D.9060100@defuze.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 22:01:15 -0000

> I can't offer better API than:
>
> open()
> read(size)
> write(data)
> close()

Just to clarify:

open() would establish the secure connection using the HTTP bootstrap 
mechanism.

Then, as a developer, I shouldn't need the framing protocol described 
because I already  know what I'm expecting. I'm not opening a connection 
without knowing in advance what will be sent though. Once I have read 
whatever amount of data I need I would create a DOM event and push it 
through. I don't feel like I need the framing overhead to do that.

Now if my connection is used to transport several distinct protocols, 
that is different, but as Roy suggested, is there still a need to limit 
the number of connections opened by the browser these days? Do we need 
multiplexing?

- Sylvain

From ian@hixie.ch  Sun Apr 26 15:57:32 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8D8B83A6EC9 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 15:57:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.156
X-Spam-Level: 
X-Spam-Status: No, score=-3.156 tagged_above=-999 required=5 tests=[AWL=-0.557, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Ayh5y6cZ+rT6 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 15:57:31 -0700 (PDT)
Received: from looneymail-a4.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id A050E3A6E09 for <hybi@ietf.org>; Sun, 26 Apr 2009 15:57:31 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a4.g.dreamhost.com (Postfix) with ESMTP id 73F49125C83; Sun, 26 Apr 2009 15:58:51 -0700 (PDT)
Date: Sun, 26 Apr 2009 22:58:50 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Jamie Lokier <jamie@shareable.org>, Sylvain Hellegouarch <sh@defuze.org>
In-Reply-To: <49F4D9F4.8070100@defuze.org>
Message-ID: <Pine.LNX.4.62.0904262252420.12381@hixie.dreamhostps.com>
References: <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <49F1DEF9.2000209@webtide.com> <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org> <Pine.LNX.4.62.0904262103230.12381@hixie.dreamhostps.com> <49F4D68D.9060100@defuze.org> <49F4D9F4.8070100@defuze.org>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Content-ID: <Pine.LNX.4.62.0904262254131.12381@hixie.dreamhostps.com>
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 22:57:32 -0000

On Sun, 26 Apr 2009, Jamie Lokier wrote:
> 
> As far as I can tell, the current WebSockets Javascript/DOM API requires 
> the protocol to have self-describing frames, because that's how it knows 
> when a message is received, then it delivers the message.
> 
> So it's a property of the API requiring that from the protocol.

Right.


> Fine, if it really does make the API work better.  But as I said 
> earlier, there are other approaches to the API which still work with the 
> DOM "push mechanism" and don't have the protocol framing requirement, 
> and will then work with general TCP services more easily.

> > Concrete API proposals would be good here.
>
> I agree they'd help a lot, and I'll take a go at it when I have time, in 
> a few days.

I look forward to seeing concrete proposals for how this would work.


> [...] I don't see the point of the type byte except for the "future 
> extension" of byte frames in Javascript.

That's the only reason for the type byte, yes.


> So if you're going to have a future extension which is not to be 
> implemented in the first version and so cannot be used except when 
> explicitly requested, you might as well make it more general - plan for 
> accepting any byte stream, when/if byte handling is added to Javascript.

I'm not really sure I see a way to do that. Concrete proposals would be 
very useful here.


On Sun, 26 Apr 2009, Sylvain Hellegouarch wrote:
>
> If I recap, as I understand it, there's been a requirement to describe 
> and define the best solution for long lived connections from browsers. 
> That requirement has two main constraints and a feature.
> 
> The first constraint is to have a web oriented security model.
>
> The second constraint is to be able to re-use the existing web 
> infrastructure hence the HTTP bootstrap.

The requirements that brought about WebSocket are a little more elaborate; 
they are documented here:

   http://www.ietf.org/mail-archive/web/hybi/current/msg00007.html


> I can't offer better API than:
> 
> open()
> read(size)
> write(data)
> close()

Could you elaborate on the semantics of the read() and write() methods? 
e.g. do they block, how does the user know when to call them, what do they 
return, how does the user handle TCP packets that split UTF-8 bytes, etc.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From Martin.Thomson@andrew.com  Sun Apr 26 16:46:58 2009
Return-Path: <Martin.Thomson@andrew.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4C08A3A6DC3 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 16:46:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.523
X-Spam-Level: 
X-Spam-Status: No, score=-2.523 tagged_above=-999 required=5 tests=[AWL=0.076,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id e5ViadDhioxt for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 16:46:57 -0700 (PDT)
Received: from andrew.com (smtp3.andrew.com [198.135.207.235]) by core3.amsl.com (Postfix) with ESMTP id 64B963A6A3F for <hybi@ietf.org>; Sun, 26 Apr 2009 16:46:57 -0700 (PDT)
X-SEF-Processed: 5_0_0_910__2009_04_26_19_09_01
X-SEF-16EBA1E9-99E8-4E1D-A1CA-4971F5510AF: 1
Received: from acdcexbh1.andrew.com [10.86.20.91] by smtp3.andrew.com - SurfControl E-mail Filter (5.2.1); Sun, 26 Apr 2009 19:09:01 -0500
Received: from AHQEX1.andrew.com ([10.86.20.21]) by acdcexbh1.andrew.com with Microsoft SMTPSVC(6.0.3790.3959); Sun, 26 Apr 2009 18:48:16 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Date: Sun, 26 Apr 2009 18:48:14 -0500
Message-ID: <E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com>
In-Reply-To: <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [hybi] Is there a traffic jam?
Thread-Index: AcnEtCMLFv5NJ+2rSQ+SjgQ5XFzz0wCFFK0A
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com><1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com> <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>
From: "Thomson, Martin" <Martin.Thomson@andrew.com>
To: "Ian Hickson" <ian@hixie.ch>, "Paul Prescod" <paul@prescod.net>
X-OriginalArrivalTime: 26 Apr 2009 23:48:16.0967 (UTC) FILETIME=[7887ED70:01C9C6C9]
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Apr 2009 23:46:58 -0000
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From gregw@webtide.com  Sun Apr 26 17:00:19 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 273783A6ABE for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 17:00:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.545
X-Spam-Level: 
X-Spam-Status: No, score=-2.545 tagged_above=-999 required=5 tests=[AWL=0.054,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KAp392+cvSjT for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 17:00:18 -0700 (PDT)
Received: from qw-out-2122.google.com (qw-out-2122.google.com [74.125.92.27]) by core3.amsl.com (Postfix) with ESMTP id 4E6E23A6AB2 for <hybi@ietf.org>; Sun, 26 Apr 2009 17:00:18 -0700 (PDT)
Received: by qw-out-2122.google.com with SMTP id 3so2243899qwe.31 for <hybi@ietf.org>; Sun, 26 Apr 2009 17:01:38 -0700 (PDT)
Received: by 10.220.92.21 with SMTP id p21mr9226831vcm.47.1240790498159; Sun, 26 Apr 2009 17:01:38 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 4sm12721131yxj.57.2009.04.26.17.01.36 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sun, 26 Apr 2009 17:01:37 -0700 (PDT)
Message-ID: <49F4F5D5.7080900@webtide.com>
Date: Mon, 27 Apr 2009 10:01:25 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com>	<49F1DEF9.2000209@webtide.com>	<79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com>	<Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com>	<20090424211314.GA18260@shareable.org>	<Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com>	<20090425010110.GB23106@shareable.org>	<49F355EA.3010101@defuze.org>	<20090426194701.GB10627@shareable.org>	<49F4CAA4.6020602@defuze.org> <20090426211322.GE10627@shareable.org>
In-Reply-To: <20090426211322.GE10627@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 00:00:19 -0000

Jamie Lokier wrote:

>> In other words, why do we need the framing part of the protocol? Ian?
> 
> As far as I can tell, the current WebSockets Javascript/DOM API
> requires the protocol to have self-describing frames, because that's
> how it knows when a message is received, then it delivers the message.


What I think is also an unstated reason for having framing is that it
can be used to limit/define the buffering done by any intermediary (proxy/
gateway/firewall).

The minimal behaviour for intermediaries will need to be specified and
while it could be done in terms of immediate forwarding of every byte
received, that would significantly reduce the options for implementing
an intermediary and handling differing MTU sizes etc.

It would be more reasonable minimum requirement that an intermediary
must forward only when a complete frame is received.

Of course this would not prevent an implementation that did forward
every byte and/or packet received, as that would also meet this
minimum requirement.


regards

From ian@hixie.ch  Sun Apr 26 18:00:25 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8C19F3A6996 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 18:00:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.148
X-Spam-Level: 
X-Spam-Status: No, score=-3.148 tagged_above=-999 required=5 tests=[AWL=-0.549, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2GHH88voCRVR for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 18:00:24 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id C06833A686A for <hybi@ietf.org>; Sun, 26 Apr 2009 18:00:24 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id 2769827B77; Sun, 26 Apr 2009 18:01:45 -0700 (PDT)
Date: Mon, 27 Apr 2009 01:01:44 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: "Thomson, Martin" <Martin.Thomson@andrew.com>
In-Reply-To: <E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com>
Message-ID: <Pine.LNX.4.62.0904270033570.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <Pine.LNX.4.62.0904140002360.10339@hixie.dreamhostps.com> <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com><1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com> <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 01:00:25 -0000

On Sun, 26 Apr 2009, Thomson, Martin wrote:
> > 
> > Why does everyone under the sun reimplement their own URI parser, 
> > despite libraries existing and despite URI parsing actually being 
> > pretty complicated to get right?
> 
> I think that I can answer that - because the protocol _appears_ so 
> simple, that folks think they can do better.  This might be pure ego, or 
> it could be something as basic as the library API not meshing well with 
> their application design.  By making the protocol simple, you've lowered 
> the bar; by lowering the bar, you've made people write their own.

If the decision is between the following:

 - A complex-looking protocol that will be implemented a few times only

 - A simple-looking protocol that is actually rather complex that will be 
   implemented many times but often incorrectly

 - A simple-looking protocol that is actually simple that will be 
   implemented many times but mostly correctly

...then I guess I'll pick the third one. :-)


> So why is that we heap derision on those who roll their own XML parser, 
> then advocate the same behaviour for another protocol?

You can't write a fully-conforming XML parser in 100 lines of (normally- 
indented) Perl. It falls in the second category above (presumably; I 
would have thought it looked pretty damn complex personally, but I'm 
going with your theory here).


> Is there some complexity line that is crossed where one is acceptable, 
> but the other not?

I don't think it's a line, more of a continuum. IMHO reading a few bytes 
off the wire like in the WebSocket protocol is way easier than 
implementing a multi-state parser with the complexity of XML's.

But that's mostly academic -- according to your hypothesis, if the 
protocol looks simple, whether it is or not, people will implement it. 
Since making the protocol complex arbitrarily is bad language design, that 
just leaves us in the position of doing our best to make it as simple as 
we possibly can.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From jamie@shareable.org  Sun Apr 26 18:07:35 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4C7033A6860 for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 18:07:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.2
X-Spam-Level: 
X-Spam-Status: No, score=-3.2 tagged_above=-999 required=5 tests=[AWL=-0.601,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2zdmpJDTDIbX for <hybi@core3.amsl.com>; Sun, 26 Apr 2009 18:07:34 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id 30CD33A67D2 for <hybi@ietf.org>; Sun, 26 Apr 2009 18:07:34 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LyFL3-00052A-6F; Mon, 27 Apr 2009 02:08:53 +0100
Date: Mon, 27 Apr 2009 02:08:53 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Greg Wilkins <gregw@webtide.com>
Message-ID: <20090427010853.GA18569@shareable.org>
References: <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org> <20090426211322.GE10627@shareable.org> <49F4F5D5.7080900@webtide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <49F4F5D5.7080900@webtide.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 01:07:35 -0000

Greg Wilkins wrote:
> Jamie Lokier wrote:
> 
> >>In other words, why do we need the framing part of the protocol? Ian?
> >
> >As far as I can tell, the current WebSockets Javascript/DOM API
> >requires the protocol to have self-describing frames, because that's
> >how it knows when a message is received, then it delivers the message.
> 
> What I think is also an unstated reason for having framing is that it
> can be used to limit/define the buffering done by any intermediary (proxy/
> gateway/firewall).

It can, but is there anything gained by that?

It doesn't limit the buffering: I can send a 100MB frame - why not?
Will that break some intermediaries?  Am I out of spec sending that
from a server/client?  Does the spec give a maximum frame length,
which all frame-buffering intermediaries guarantee to support, and
will applications then need to work around that if they happen to have
a message which is larger?

Sending a large chunk of HTML to replace in a page is a likely use
case; you know how big those get.

My point is it's hard enough testing things on servers and clients.
Whatever deployed intermediaries there are, it's much harder for
specific web site authors to test against them, or even know what they
are.

So it's really important that intermediaries don't behave in strange
ways with arbitrary limits, especially if those limits cannot be
tested for and negotiated around easily later.

> The minimal behaviour for intermediaries will need to be specified and
> while it could be done in terms of immediate forwarding of every byte
> received, that would significantly reduce the options for implementing
> an intermediary and handling differing MTU sizes etc.

TCP/IP provides pure byte streams without any boundaries.  There's no
connection between WebSockets frame boundaries and packet boundaries,
except by chance some of the time.

Intermediaries forward between TCP byte streams.  They only see byte
streams.  They can _parse_ the byte stream to get WebSockets frames,
but that has no connection with MTU which is an IP packet concept
below TCP, and hidden completely from intermediary applications.

Differing MTU is handled by the two TCPs.

Did you have something in mind where an intermediary bypasses TCP and
operates on IP packets directly?  (That would be quite different from
what's been proposed).

> It would be more reasonable minimum requirement that an intermediary
> must forward only when a complete frame is received.

It's very important that this is not misunderstood as _requiring_ an
intermediary to buffer a frame until it's received.  That would be
tragic :-)

I'm not sure what benefit it brings, if the protocol isn't
multiplexing different connections.

For an intermediary to be _allowed_ to delay data until a whole frame
is parsed limits how the protocol may be used.

The most stringent is that you can't use frame types that any
intermediary doesn't recognise, because it'll break their parsing.

Since only UTF-8 sentinel-delimited text frames will actually be
deployed at first, that means the first generation of intermediaries
will _only_ be able to forward that, if they take up your
implementation advice.

Should the spec say what happens when an intermediary cannot parse the
byte stream because it has an unrecognised frame type?  Is it
something newer clients/servers will be able to detect and workaround
(e.g. by tunnelling other frame types into UTF-8 encoded frames), or
will newer clients/servers have difficulty detecting and working
around older intermediaries which cannot handle other frame types?

By the way, this is quite different from HTTP chunking in an
interesting way.  HTTP chunk boundaries are parsed by intermediaries
for the useful ability to _change_ the boundaries to match
transmission characteristics.  Now _that's_ a usefully relaxed framing
constraint :-)

But that's not allowed in WebSockets because the boundaries are
part of the data transmitted.

I think a most reasonable requirement is:

   - Intermediaries don't edit the byte stream (unless they have
     specific reason to do so perhaps).
   - Intermediaries forward all of it in a reasonable time without
     attempting to buffer whole frames, which they might fail to do.

That will ensure intermediaries work with any frame size, and work
with any frame types (or length formats) which are added later to
clients and servers.

Now if multiplexing of any kind is added :-)  I'd still keep the
requirement to not attempt to buffer whole frames, for the same
reasons, but of course parsing and editing of some kind is implied.

-- Jamie

From gregw@webtide.com  Mon Apr 27 05:45:11 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D03843A6929 for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 05:45:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.546
X-Spam-Level: 
X-Spam-Status: No, score=-2.546 tagged_above=-999 required=5 tests=[AWL=0.053,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GkJyPMpfwBTY for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 05:45:11 -0700 (PDT)
Received: from mail-gx0-f166.google.com (mail-gx0-f166.google.com [209.85.217.166]) by core3.amsl.com (Postfix) with ESMTP id 05B9D3A6D61 for <hybi@ietf.org>; Mon, 27 Apr 2009 05:45:10 -0700 (PDT)
Received: by gxk10 with SMTP id 10so4679822gxk.13 for <hybi@ietf.org>; Mon, 27 Apr 2009 05:46:31 -0700 (PDT)
Received: by 10.90.66.14 with SMTP id o14mr6963174aga.106.1240836391203; Mon, 27 Apr 2009 05:46:31 -0700 (PDT)
Received: from ?10.10.1.13? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 36sm1272359agc.60.2009.04.27.05.46.29 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 27 Apr 2009 05:46:30 -0700 (PDT)
Message-ID: <49F5A921.7080102@webtide.com>
Date: Mon, 27 Apr 2009 22:46:25 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <79ea848f0904241135p49a7a310sf2717070165b8ae8@mail.gmail.com> <Pine.LNX.4.62.0904241945030.12381@hixie.dreamhostps.com> <20090424211314.GA18260@shareable.org> <Pine.LNX.4.62.0904242129110.12381@hixie.dreamhostps.com> <20090425010110.GB23106@shareable.org> <49F355EA.3010101@defuze.org> <20090426194701.GB10627@shareable.org> <49F4CAA4.6020602@defuze.org> <20090426211322.GE10627@shareable.org> <49F4F5D5.7080900@webtide.com> <20090427010853.GA18569@shareable.org>
In-Reply-To: <20090427010853.GA18569@shareable.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 12:45:11 -0000

Jamie Lokier wrote:
> It doesn't limit the buffering: I can send a 100MB frame - why not?

Jamie,

I wasn't proposing that intermediaries MUST buffer whole frames.
I'm proposing that intermediaries MUST NOT buffer beyond a frame boundary.
ie the end of a frame must always trigger a flush of any buffers that
may or may not exist.

This is specifically different to HTTP chunking which is point to point
and an intermediary is perfectly entitled to buffer up all chunks before
forwarding an unchunked message.

> I think a most reasonable requirement is:
> 
>    - Intermediaries don't edit the byte stream (unless they have
>      specific reason to do so perhaps).
>    - Intermediaries forward all of it in a reasonable time without
>      attempting to buffer whole frames, which they might fail to do.

Something like would also work.  However it might be harder to
rigorously define "all of it in a reasonable time".   It might also
discourage some reasonable amount of buffering.


> Now if multiplexing of any kind is added :-)  I'd still keep the
> requirement to not attempt to buffer whole frames

Agreed.

Indeed multiplexing might require transport level chunking of application
defined frames, so that large content does not delay other channels.

The trick is to find a way to support such complexity in a way that
still allows a fallback to simplicity if the complex features are
not required.

cheers





From martin.tyler@boo.org  Mon Apr 27 06:37:02 2009
Return-Path: <martin.tyler@boo.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4E8F23A6F85 for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 06:37:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.976
X-Spam-Level: 
X-Spam-Status: No, score=-1.976 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rLlFHEXVjbjo for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 06:37:01 -0700 (PDT)
Received: from mail-gx0-f166.google.com (mail-gx0-f166.google.com [209.85.217.166]) by core3.amsl.com (Postfix) with ESMTP id 16EF028C14E for <hybi@ietf.org>; Mon, 27 Apr 2009 06:35:55 -0700 (PDT)
Received: by gxk10 with SMTP id 10so4735189gxk.13 for <hybi@ietf.org>; Mon, 27 Apr 2009 06:37:15 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.90.65.14 with SMTP id n14mr7091057aga.29.1240839435528; Mon,  27 Apr 2009 06:37:15 -0700 (PDT)
In-Reply-To: <49F5AF2C.9000409@ericsson.com>
References: <5dd474260812030437i57287df8l164006ea2859d203@mail.gmail.com> <5dd474260812030734s3389aec3y5f63c7dd572ac82e@mail.gmail.com> <4936C476.1010708@gmail.com> <5dd474260812151553g23e3d9e5i478282da561b75cd@mail.gmail.com> <3a05072b0904080200v4f824b29s21925e2809c93ae@mail.gmail.com> <c7c14730904080356s2df228d1id324d3e42cf1fa3e@mail.gmail.com> <3a05072b0904170945n56b14d28o30f578c96a372323@mail.gmail.com> <Pine.LNX.4.62.0904240611350.10370@hixie.dreamhostps.com> <44abafb90904240400n7b7bfe71x796fa731b7866ac4@mail.gmail.com> <49F5AF2C.9000409@ericsson.com>
Date: Mon, 27 Apr 2009 14:37:15 +0100
Message-ID: <44abafb90904270637y4044aa54sa5db4ebcf6a7e190@mail.gmail.com>
From: Martin Tyler <martin.tyler@boo.org>
To: Salvatore Loreto <salvatore.loreto@ericsson.com>, hybi@ietf.org
Content-Type: multipart/alternative; boundary=001636283b74c736e50468897154
Subject: Re: [hybi] WebSocket feedback
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 13:37:02 -0000

--001636283b74c736e50468897154
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Hi

I don't have any links or anything.

A reverse proxy basically pretends to be the actual server (as far as any
client or client proxy is concerned), but it passes on the request to the
actual server that is usually sitting behind another layer of firewalls.

Any polling or long polling comet solution should be fine with this, as will
most streaming comet connections. The main downside is just the performance
- since most proxies are designed to hold lots of open connections like a
dedicated comet server is.

However, when it comes to WebSocket, which isnt a standard use of HTTP
(unlike Comet, which is HTTP, just a little hacky), the reserve proxy isnt
going to know what to do with subsequent client->server messages in the
bidirectional communication - that's if it gets passed the handshake part in
the first place.


On Mon, Apr 27, 2009 at 2:12 PM, Salvatore Loreto <
salvatore.loreto@ericsson.com> wrote:

> Martin Tyler wrote:
>
>>
>> I'm not sure I understand this smuggling of messages, could you explain?
>> Otherwise you seem to be saying the exact bytes matter without explaining
>> why.
>>
>> It seems odd to me that the headers need to be exact.. One of the big
>> headaches I can see with WebSocket or anything like it is compatibility with
>> existing proxies, which is why this has to be some kind of HTTP (at least to
>> start with) rather than a TCP socket. HTTP proxies will sometimes parse
>> request/headers from the client and write their own when passing the message
>> on to the server.. won't requiring the exact bytes above cause problems
>> here?
>>
>>
>> A more general point about WebSocket and proxies. I don't know what
>> experience the people on this list have exactly, I am from Caplin Systems
>> (Liberator comet server amongst other things), most of our customers are
>> investment banks with large inflexible web architectures. At the start of
>> most projects we go through the same problem of explaining that Liberator
>> will work best if it is Internet facing.. most of these banks setups consist
>> of multiple tiers of proxies/webservers/appservers etc.. usually everything
>> sits behind a thin reverse proxy layer. Liberator (and most Comet servers I
>> would imagine) can work behind these, but you can lose out big time on
>> performance since the proxy isnt designed for the kind of traffic a comet
>> server might deliver.
>>
> Hi Martin,
>
> this is an interesting problem, how Comet behave with "Reverse Proxy"
> do you have any more information : document or link about this problem?
>
> thanks
> Sal
>

--001636283b74c736e50468897154
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hi<br><br>I don&#39;t have any links or anything.<br><br>A reverse proxy ba=
sically pretends to be the actual server (as far as any client or client pr=
oxy is concerned), but it passes on the request to the actual server that i=
s usually sitting behind another layer of firewalls.<br>
<br>Any polling or long polling comet solution should be fine with this, as=
 will most streaming comet connections. The main downside is just the perfo=
rmance - since most proxies are designed to hold lots of open connections l=
ike a dedicated comet server is.<br>
<br>However, when it comes to WebSocket, which isnt a standard use of HTTP =
(unlike Comet, which is HTTP, just a little hacky), the reserve proxy isnt =
going to know what to do with subsequent client-&gt;server messages in the =
bidirectional communication - that&#39;s if it gets passed the handshake pa=
rt in the first place.<br>
<br><br><div class=3D"gmail_quote">On Mon, Apr 27, 2009 at 2:12 PM, Salvato=
re Loreto <span dir=3D"ltr">&lt;<a href=3D"mailto:salvatore.loreto@ericsson=
.com">salvatore.loreto@ericsson.com</a>&gt;</span> wrote:<br><blockquote cl=
ass=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); mar=
gin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"im">Martin Tyler wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, =
204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
I&#39;m not sure I understand this smuggling of messages, could you explain=
? Otherwise you seem to be saying the exact bytes matter without explaining=
 why.<br>
<br>
It seems odd to me that the headers need to be exact.. One of the big heada=
ches I can see with WebSocket or anything like it is compatibility with exi=
sting proxies, which is why this has to be some kind of HTTP (at least to s=
tart with) rather than a TCP socket. HTTP proxies will sometimes parse requ=
est/headers from the client and write their own when passing the message on=
 to the server.. won&#39;t requiring the exact bytes above cause problems h=
ere?<br>

<br>
<br>
A more general point about WebSocket and proxies. I don&#39;t know what exp=
erience the people on this list have exactly, I am from Caplin Systems (Lib=
erator comet server amongst other things), most of our customers are invest=
ment banks with large inflexible web architectures. At the start of most pr=
ojects we go through the same problem of explaining that Liberator will wor=
k best if it is Internet facing.. most of these banks setups consist of mul=
tiple tiers of proxies/webservers/appservers etc.. usually everything sits =
behind a thin reverse proxy layer. Liberator (and most Comet servers I woul=
d imagine) can work behind these, but you can lose out big time on performa=
nce since the proxy isnt designed for the kind of traffic a comet server mi=
ght deliver.<br>

</blockquote></div>
Hi Martin,<br>
<br>
this is an interesting problem, how Comet behave with &quot;Reverse Proxy&q=
uot;<br>
do you have any more information : document or link about this problem?<br>
<br>
thanks<br>
Sal<br>
</blockquote></div><br>

--001636283b74c736e50468897154--

From jamie@shareable.org  Mon Apr 27 08:03:16 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 917633A6FA1 for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 08:03:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.182
X-Spam-Level: 
X-Spam-Status: No, score=-3.182 tagged_above=-999 required=5 tests=[AWL=-0.583, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7rq0JSeWmhSJ for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 08:03:15 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id BC8C13A6F90 for <hybi@ietf.org>; Mon, 27 Apr 2009 08:03:15 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LySNh-0001qE-S1; Mon, 27 Apr 2009 16:04:29 +0100
Date: Mon, 27 Apr 2009 16:04:29 +0100
From: Jamie Lokier <jamie@shareable.org>
To: Ian Hickson <ian@hixie.ch>
Message-ID: <20090427150429.GE4885@shareable.org>
References: <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com> <Pine.LNX.4.62.0904270033570.12381@hixie.dreamhostps.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.62.0904270033570.12381@hixie.dreamhostps.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org, "Thomson, Martin" <Martin.Thomson@andrew.com>
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 15:03:16 -0000

Ian Hickson wrote:
>  - A complex-looking protocol that will be implemented a few times only

==> HTML

>  - A simple-looking protocol that is actually rather complex that will be 
>    implemented many times but often incorrectly

==> HTTP

>  - A simple-looking protocol that is actually simple that will be 
>    implemented many times but mostly correctly

==> :-)

> But that's mostly academic -- according to your hypothesis, if the 
> protocol looks simple, whether it is or not, people will implement it.

I think that's correct.

> Since making the protocol complex arbitrarily is bad language design, that 
> just leaves us in the position of doing our best to make it as simple as 
> we possibly can.

I think WebSockets is in the unusual position that client browser
implementations won't be implemented often, but servers will be and so
will any Javascript layer that people need to put on top.

I guess that's why I push for more complexity in the client browser
implementation, in exchange for less complexity at the Javascript
level when you can't avoid putting it somewhere, for a broad class of
applications.

You can bet whatever people have to write in Javascript, will be
written many times over and badly.

-- Jamie

From ian@hixie.ch  Mon Apr 27 11:16:23 2009
Return-Path: <ian@hixie.ch>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EBA303A6F31 for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 11:16:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.139
X-Spam-Level: 
X-Spam-Status: No, score=-3.139 tagged_above=-999 required=5 tests=[AWL=-0.540, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jPMuyA4yr5UC for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 11:16:23 -0700 (PDT)
Received: from looneymail-a3.g.dreamhost.com (caibbdcaaaaf.dreamhost.com [208.113.200.5]) by core3.amsl.com (Postfix) with ESMTP id BDD6D3A6B13 for <hybi@ietf.org>; Mon, 27 Apr 2009 11:16:05 -0700 (PDT)
Received: from hixie.dreamhostps.com (hixie.dreamhost.com [208.113.210.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by looneymail-a3.g.dreamhost.com (Postfix) with ESMTP id AF07627B7D; Mon, 27 Apr 2009 11:17:26 -0700 (PDT)
Date: Mon, 27 Apr 2009 18:17:26 +0000 (UTC)
From: Ian Hickson <ian@hixie.ch>
To: Jamie Lokier <jamie@shareable.org>
In-Reply-To: <20090427150429.GE4885@shareable.org>
Message-ID: <Pine.LNX.4.62.0904271812110.12381@hixie.dreamhostps.com>
References: <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com> <Pine.LNX.4.62.0904270033570.12381@hixie.dreamhostps.com> <20090427150429.GE4885@shareable.org>
Content-Language: en-GB-hixie
Content-Style-Type: text/css
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 18:16:24 -0000

On Mon, 27 Apr 2009, Jamie Lokier wrote:
> > 
> > Since making the protocol complex arbitrarily is bad language design, 
> > that just leaves us in the position of doing our best to make it as 
> > simple as we possibly can.
> 
> I think WebSockets is in the unusual position that client browser 
> implementations won't be implemented often, but servers will be and so 
> will any Javascript layer that people need to put on top.

Agreed.


> I guess that's why I push for more complexity in the client browser 
> implementation, in exchange for less complexity at the Javascript level 
> when you can't avoid putting it somewhere, for a broad class of 
> applications.

I'm happy to add more complexity (so long as it is justified) to the 
client implementation, with the caveat that it must not also increase the 
complexity of the server-side implementations.

The client component has to deal with any server, including non-conforming 
servers sending back garbage, future servers that support an extended 
version of the protocol, etc. Server components only have to deal with the 
output possible from their own scripts driving a conforming implementation 
of the API. We can leverage that asymmetry.

This is why currently the spec for the client side defines the entire 
protocol (including binary frames which might be sent by servers in the 
future), but the spec for the server side only defines the rules required 
for handling content sent by the current client API.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

From prescod@gmail.com  Mon Apr 27 11:46:50 2009
Return-Path: <prescod@gmail.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E76EB3A6ACC for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 11:46:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.744
X-Spam-Level: 
X-Spam-Status: No, score=-1.744 tagged_above=-999 required=5 tests=[AWL=0.232,  BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FSvvB004Pnxv for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 11:46:50 -0700 (PDT)
Received: from mail-gx0-f166.google.com (mail-gx0-f166.google.com [209.85.217.166]) by core3.amsl.com (Postfix) with ESMTP id C6E533A6FF3 for <hybi@ietf.org>; Mon, 27 Apr 2009 11:46:49 -0700 (PDT)
Received: by gxk10 with SMTP id 10so180948gxk.13 for <hybi@ietf.org>; Mon, 27 Apr 2009 11:48:10 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to:cc :content-type; bh=Yl27ypyWHppT928ojQyRTQEa3ci20d4haeTfNpd/95Y=; b=wYQlam6eLtOairfQfWkBVEVilcdm+zIfy5nHXafrlEh80htPhinmljUSvKRvnqR+2d vKo5oaMDS7hA0ao9qP9n+/6riUr7jVASno/ittGojs7XiY2AG2R17GnSkB7sxgOLNd4/ 5DKJZXxXxmRgQggg38nrzNwZlsgeEhX1ieJAk=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; b=JK/KDvmuZtkfEOMvub1sRMvimzqP0QXJfBMByN5k5Ss3D9wtlFpXt1irEVSzAWdhvM mGpUAbg6MfqkwgXciTgYO1m5M0ccLUCEmb56OBSufCQq4f97+vgpaJm9TjMMhRQaAiR3 dXdkDDZn26BgghqYxnoeMBDiAOtm50i+VsmAY=
MIME-Version: 1.0
Sender: prescod@gmail.com
Received: by 10.150.136.12 with SMTP id j12mr10637778ybd.46.1240858090167;  Mon, 27 Apr 2009 11:48:10 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>
References: <03BCE29D-7AA5-4128-9F61-446E0229479A@lindenlab.com> <Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com> <1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com> <Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com> <1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com> <Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com> <E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com> <Pine.LNX.4.62.0904240652180.12381@hixie.dreamhostps.com> <1cb725390904240018n20e73324q955c349723b45a45@mail.gmail.com> <Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>
Date: Mon, 27 Apr 2009 11:48:10 -0700
X-Google-Sender-Auth: ba7d649b1f6b4135
Message-ID: <1cb725390904271148u34cb7e29we1b9713442644208@mail.gmail.com>
From: Paul Prescod <paul@prescod.net>
To: Ian Hickson <ian@hixie.ch>
Content-Type: multipart/alternative; boundary=000e0cd5d000ae5fdf04688dc9e0
Cc: hybi@ietf.org
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2009 18:46:51 -0000

--000e0cd5d000ae5fdf04688dc9e0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On Fri, Apr 24, 2009 at 1:10 AM, Ian Hickson <ian@hixie.ch> wrote:

> On Fri, 24 Apr 2009, Paul Prescod wrote:
> >
> > Libraries for UTF-8 exist. Libraries for XML exist. Libraries for EDI
> exist.
> > Libraries for SGML exist. Libraries for TCP sockets exist. Libraries for
> > every important standard under the sun EXIST. Why will WebSockets be a
> > unique and special case?
>
> It won't. It'll be just like XML, and SGML, and JSON, and URIs, and CGI,
> and HTML, and ...: despite the existence of high-quality libraries, people
> will write their own, and the more complex the language, the more of a
> mess it makes.


Not one of the things you listed is a networking protocol. CGI is the
closest thing to what we are talking about, and actually I see very, very
few hand-written CGI parsers. In fact, I'd venture that only a tiny
percentage of all "inexperienced" web programmers know that CGI exists. The
vast majority of them know about PHP or ASP and depend on those tools to
handle CGI (or server-internal equivalent).

That's the right model for network server software development. Let
security-aware experts handle the networking interface. Let them simplify
the protocol to simple parameter values or global variables. The vast
majority of people should understand the protocol at the wire-trace level
but they don't need to implement it.

 Paul Prescod

--000e0cd5d000ae5fdf04688dc9e0
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Fri, Apr 24, 2009 at 1:10 AM, Ian Hickson <span dir=3D"ltr">&lt;ian@hixi=
e.ch&gt;</span> wrote:<br><div class=3D"gmail_quote"><blockquote class=3D"g=
mail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt=
 0pt 0pt 0.8ex; padding-left: 1ex;">

<div>On Fri, 24 Apr 2009, Paul Prescod wrote:<br>
&gt;<br>
&gt; Libraries for UTF-8 exist. Libraries for XML exist. Libraries for EDI =
exist.<br>
&gt; Libraries for SGML exist. Libraries for TCP sockets exist. Libraries f=
or<br>
&gt; every important standard under the sun EXIST. Why will WebSockets be a=
<br>
&gt; unique and special case?<br>
<br>
</div>It won&#39;t. It&#39;ll be just like XML, and SGML, and JSON, and URI=
s, and CGI,<br>
and HTML, and ...: despite the existence of high-quality libraries, people<=
br>
will write their own, and the more complex the language, the more of a<br>
mess it makes. </blockquote><div><br>Not one of the things you listed is a =
networking protocol. CGI is the closest thing to what we are talking about,=
 and actually I see very, very few hand-written CGI parsers. In fact, I&#39=
;d venture that only a tiny percentage of all &quot;inexperienced&quot; web=
 programmers know that CGI exists. The vast majority of them know about PHP=
 or ASP and depend on those tools to handle CGI (or server-internal equival=
ent).<br>
<br>That&#39;s the right model for network server software development. Let=
 security-aware experts handle the networking interface. Let them simplify =
the protocol to simple parameter values or global variables. The vast major=
ity of people should understand the protocol at the wire-trace level but th=
ey don&#39;t need to implement it.<br>
<br>=C2=A0Paul Prescod<br><br></div></div>

--000e0cd5d000ae5fdf04688dc9e0--

From gregw@webtide.com  Mon Apr 27 19:01:57 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A84C03A6B22 for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 19:01:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.547
X-Spam-Level: 
X-Spam-Status: No, score=-2.547 tagged_above=-999 required=5 tests=[AWL=0.052,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QnVSh+LDtj7x for <hybi@core3.amsl.com>; Mon, 27 Apr 2009 19:01:56 -0700 (PDT)
Received: from mail-gx0-f166.google.com (mail-gx0-f166.google.com [209.85.217.166]) by core3.amsl.com (Postfix) with ESMTP id A4AD43A6856 for <hybi@ietf.org>; Mon, 27 Apr 2009 19:01:56 -0700 (PDT)
Received: by gxk10 with SMTP id 10so613848gxk.13 for <hybi@ietf.org>; Mon, 27 Apr 2009 19:03:17 -0700 (PDT)
Received: by 10.90.82.17 with SMTP id f17mr7829755agb.41.1240884197433; Mon, 27 Apr 2009 19:03:17 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 8sm10805035agd.77.2009.04.27.19.03.15 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 27 Apr 2009 19:03:16 -0700 (PDT)
Message-ID: <49F663DF.7030300@webtide.com>
Date: Tue, 28 Apr 2009 12:03:11 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: hybi@ietf.org
References: <1cb725390904131712k292a4860pbd078bb251d3855b@mail.gmail.com>	<Pine.LNX.4.62.0904140031040.10339@hixie.dreamhostps.com>	<1cb725390904131752u5842c039wb3d75602c479fa45@mail.gmail.com>	<Pine.LNX.4.62.0904140053050.10339@hixie.dreamhostps.com>	<1cb725390904131814o6040a8d3t637069a344d561bd@mail.gmail.com>	<Pine.LNX.4.62.0904240546100.10370@hixie.dreamhostps.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105AD1ADC@AHQEX1.andrew.com>	<Pine.LNX.4.62.0904240805380.12381@hixie.dreamhostps.com>	<E51D5B15BFDEFD448F90BDD17D41CFF105AD1E13@AHQEX1.andrew.com>	<Pine.LNX.4.62.0904270033570.12381@hixie.dreamhostps.com>	<20090427150429.GE4885@shareable.org> <Pine.LNX.4.62.0904271812110.12381@hixie.dreamhostps.com>
In-Reply-To: <Pine.LNX.4.62.0904271812110.12381@hixie.dreamhostps.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [hybi] Is there a traffic jam?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2009 02:01:57 -0000

Ian Hickson wrote:
> I'm happy to add more complexity (so long as it is justified) to the 
> client implementation, with the caveat that it must not also increase the 
> complexity of the server-side implementations.

Would it be sufficient to allow optional support of complex features.
So a server could implement a either a simple core of websocket or
a richer more complex feature set?

Also, you have argued against additional frame-type that could better
support complex behaviour on the grounds that the base frame types
can carry that complex protocol with assumptions.   If a mime-encapsulating
frame type is not worth a dedicated frame-type byte, then what
kinds of extensions do you see additional frame-type being defined for?




> The client component has to deal with any server, including non-conforming 
> servers sending back garbage, future servers that support an extended 
> version of the protocol, etc.

Welcome to the wonderful world of web protocols!

> Server components only have to deal with the 
> output possible from their own scripts driving a conforming implementation 
> of the API. We can leverage that asymmetry.

I do not think this is generally true.

Server that will implement the websocket protocol are more than likely
going to be done by entirely different sets of developers than those
that write the javascript content that is served from them.

More over, the javascript served will probably be from 2, 3 or even
more distinct sets of developers as frameworks, mashups and other
such things will bring together the work of many into a single page.

There is also the likely hood of more and more cross-site support
and websockets will have to support cross site in some way, so a server
may receive connections from code it didn't actually serve.

Furthermore, intermediary proxy's etc. may alter what is sent/received.
Even if the initial proxies/firewalls don't,  I'm sure somebody will
develop some smart websocket aware proxy that does some

So Server components also have to deal wit non-conforming clients
sending back garbage etc. etc.


> This is why currently the spec for the client side defines the entire 
> protocol (including binary frames which might be sent by servers in the 
> future), but the spec for the server side only defines the rules required 
> for handling content sent by the current client API.

I think the specification should entirely define the wire protocol without
any assumptions about who/what/where is implementing the client or the
server.    If we wait for non-js-in-browser clients before fully specifying
the server side of the protocol, then we have a chicken and the egg problem.


cheers




From john.fallows@kaazing.com  Wed Apr 29 12:06:42 2009
Return-Path: <john.fallows@kaazing.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0F29E3A69FB for <hybi@core3.amsl.com>; Wed, 29 Apr 2009 12:06:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.677
X-Spam-Level: 
X-Spam-Status: No, score=-0.677 tagged_above=-999 required=5 tests=[AWL=-1.300, BAYES_50=0.001, FM_FORGED_GMAIL=0.622]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 67GJKCToBur9 for <hybi@core3.amsl.com>; Wed, 29 Apr 2009 12:06:41 -0700 (PDT)
Received: from wf-out-1314.google.com (wf-out-1314.google.com [209.85.200.172]) by core3.amsl.com (Postfix) with ESMTP id 9D48228C225 for <hybi@ietf.org>; Wed, 29 Apr 2009 12:06:01 -0700 (PDT)
Received: by wf-out-1314.google.com with SMTP id 29so1049606wff.31 for <hybi@ietf.org>; Wed, 29 Apr 2009 12:07:24 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.142.68.2 with SMTP id q2mr161236wfa.223.1241032044171; Wed, 29  Apr 2009 12:07:24 -0700 (PDT)
Date: Wed, 29 Apr 2009 12:07:24 -0700
Message-ID: <c5b3a7130904291207g51201c23q156ac76bce9efaf4@mail.gmail.com>
From: John Fallows <john.fallows@kaazing.com>
To: hybi@ietf.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: [hybi] Divide and conquer?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Apr 2009 19:06:42 -0000

Folks,

In an effort to help organize our thoughts and discussion, I propose
we divide the problem into chunks and tackle each one separately.

  o Connection strategy - eg. crossdomain.xml + raw TCP, HTTP 101
Switching Protocols, etc
  o Proxy traversal strategy - eg. transparent proxies,
non-transparent proxies, reverse proxies, etc
  o Wire protocol - eg. raw TCP, text/binary frames, higher-level
messaging protocols, etc
  o JavaScript APIs - eg. event-driven data, notify then read,
higher-level messaging protocols, etc

Does this seem like a reasonable separation of concerns that would
help us to make useful progress?

Based on the separation of responsibilities between the IETF and the
W3C, it seems that connection strategy, proxy traversal strategy and
wire protocol would fall under the IETF whereas JavaScript APIs would
fall under the W3C.

Therefore, assuming that the above separation of topics is useful, I'm
planning to wait a day or so to give folks a chance to respond, then
start a thread on each of the topics that pertain to the IETF.

Kind Regards,
John Fallows.

Disclaimer:  Kaazing currently offers (emulated) WebSocket-based
products that implement both the wire protocol (when possible,
otherwise fallback to HTTP) and JavaScript API from a draft of the
specification(s).  Our goal in participating with the IETF and the W3C
is offer insights gained from customer feedback to help standardize
the future of real-time communications on the Web.

-- 
>|< Kaazing Corporation >|<
John Fallows | CTO | +1.650.960.8148
888 Villa St, Ste 410 | Mountain View, CA 94041, USA

From jamie@shareable.org  Wed Apr 29 13:32:23 2009
Return-Path: <jamie@shareable.org>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 46AEF3A6EDE for <hybi@core3.amsl.com>; Wed, 29 Apr 2009 13:32:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.164
X-Spam-Level: 
X-Spam-Status: No, score=-3.164 tagged_above=-999 required=5 tests=[AWL=-0.565, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id JW32CwTCqRWy for <hybi@core3.amsl.com>; Wed, 29 Apr 2009 13:32:22 -0700 (PDT)
Received: from mail2.shareable.org (mail2.shareable.org [80.68.89.115]) by core3.amsl.com (Postfix) with ESMTP id C85433A7182 for <hybi@ietf.org>; Wed, 29 Apr 2009 13:30:59 -0700 (PDT)
Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from <jamie@shareable.org>) id 1LzGS3-0006yN-UM; Wed, 29 Apr 2009 21:32:19 +0100
Date: Wed, 29 Apr 2009 21:32:19 +0100
From: Jamie Lokier <jamie@shareable.org>
To: John Fallows <john.fallows@kaazing.com>
Message-ID: <20090429203219.GA20993@shareable.org>
References: <c5b3a7130904291207g51201c23q156ac76bce9efaf4@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <c5b3a7130904291207g51201c23q156ac76bce9efaf4@mail.gmail.com>
User-Agent: Mutt/1.5.13 (2006-08-11)
Cc: hybi@ietf.org
Subject: Re: [hybi] Divide and conquer?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Apr 2009 20:32:23 -0000

John Fallows wrote:
> Folks,
> 
> In an effort to help organize our thoughts and discussion, I propose
> we divide the problem into chunks and tackle each one separately.
> 
>   o Connection strategy - eg. crossdomain.xml + raw TCP, HTTP 101
> Switching Protocols, etc
>   o Proxy traversal strategy - eg. transparent proxies,
> non-transparent proxies, reverse proxies, etc

These seem to be completely independent of the Javascript APIs as far
as I can tell.  As long as they can deliver a two-way message or byte
stream.  Both look tremendously useful for any other protcols being
tunnelled over HTTP too, not just WebSockets.

>   o Wire protocol - eg. raw TCP, text/binary frames, higher-level
> messaging protocols, etc
>   o JavaScript APIs - eg. event-driven data, notify then read,
> higher-level messaging protocols, etc

These are mutually dependent because the current Javascript API
requires the protocol to have delimited message boundaries, so general
"TCP style" byte stream is not an option with the current API.

> Does this seem like a reasonable separation of concerns that would
> help us to make useful progress?

I think it's an excellent separation.

I'm not sure if the following is subsumed under wire protocol.  It's
an issue with practical implications, though interestingly needn't
affect the API, but may affect users of the API:

   o Connection merging - wire-level multiplexing or many separate
     connections, aggregation at proxies, framing and flow control.

> Based on the separation of responsibilities between the IETF and the
> W3C, it seems that connection strategy, proxy traversal strategy and
> wire protocol would fall under the IETF whereas JavaScript APIs would
> fall under the W3C.

The Javascript APIs cannot be developed entirely independently of the
wire protocol for the reason given above: the _current_ Javascript API
may/may not require the protocol to have self-describing frames, and
hence be incompatible with a general "TCP style" byte stream which
does not have message boundaries.

Maybe we'll settle on an extensible protocol with a well-defined
"profile" of capabilities which is accessible to the Javascript API.

I think the IETF people are ideal for addressing the networking
issues, and W3C people are ideal for designing a good Javascript API
for ordinary web developers; that looks like a good plan.

-- Jamie

From gregw@webtide.com  Thu Apr 30 01:38:41 2009
Return-Path: <gregw@webtide.com>
X-Original-To: hybi@core3.amsl.com
Delivered-To: hybi@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DE6863A6923 for <hybi@core3.amsl.com>; Thu, 30 Apr 2009 01:38:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.548
X-Spam-Level: 
X-Spam-Status: No, score=-2.548 tagged_above=-999 required=5 tests=[AWL=0.051,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4PUIdetSk5me for <hybi@core3.amsl.com>; Thu, 30 Apr 2009 01:38:41 -0700 (PDT)
Received: from yx-out-2324.google.com (yx-out-2324.google.com [74.125.44.28]) by core3.amsl.com (Postfix) with ESMTP id F1A893A67D2 for <hybi@ietf.org>; Thu, 30 Apr 2009 01:38:40 -0700 (PDT)
Received: by yx-out-2324.google.com with SMTP id 8so1861255yxb.49 for <hybi@ietf.org>; Thu, 30 Apr 2009 01:40:03 -0700 (PDT)
Received: by 10.100.140.15 with SMTP id n15mr2751027and.83.1241080803365; Thu, 30 Apr 2009 01:40:03 -0700 (PDT)
Received: from ?10.10.1.11? (60-242-119-126.tpgi.com.au [60.242.119.126]) by mx.google.com with ESMTPS id 11sm3169616aga.70.2009.04.30.01.40.01 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 30 Apr 2009 01:40:02 -0700 (PDT)
Message-ID: <49F963D9.7020009@webtide.com>
Date: Thu, 30 Apr 2009 18:39:53 +1000
From: Greg Wilkins <gregw@webtide.com>
User-Agent: Thunderbird 2.0.0.21 (X11/20090409)
MIME-Version: 1.0
To: John Fallows <john.fallows@kaazing.com>
References: <c5b3a7130904291207g51201c23q156ac76bce9efaf4@mail.gmail.com>
In-Reply-To: <c5b3a7130904291207g51201c23q156ac76bce9efaf4@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc: hybi@ietf.org
Subject: Re: [hybi] Divide and conquer?
X-BeenThere: hybi@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Server-Initiated HTTP <hybi.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/hybi>
List-Post: <mailto:hybi@ietf.org>
List-Help: <mailto:hybi-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/hybi>, <mailto:hybi-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 30 Apr 2009 08:38:41 -0000

John,

I think this is a good plan, but with some caveats.

The current websocket API is described in very network related
way. Eg it is about opening connections and connection state etc.

If the separation of concerns that you describe is to be successful
then the websocket API is going need to be recast in less
network centric language - ie saying channel/conduit/pipe or similar
instead of connection.     From the API point of view, it should not
matter if that channel/conduit/pipe is carried over a dedicated
socket or some other possibly shared transport.

There would need to be a good dialog between the API concerns
and the network concerns.    Perhaps that could start by the
API writing up some requirements that they would like the
network concerns to provide.

cheers







John Fallows wrote:
> Folks,
> 
> In an effort to help organize our thoughts and discussion, I propose
> we divide the problem into chunks and tackle each one separately.
> 
>   o Connection strategy - eg. crossdomain.xml + raw TCP, HTTP 101
> Switching Protocols, etc
>   o Proxy traversal strategy - eg. transparent proxies,
> non-transparent proxies, reverse proxies, etc
>   o Wire protocol - eg. raw TCP, text/binary frames, higher-level
> messaging protocols, etc
>   o JavaScript APIs - eg. event-driven data, notify then read,
> higher-level messaging protocols, etc
> 
> Does this seem like a reasonable separation of concerns that would
> help us to make useful progress?
> 
> Based on the separation of responsibilities between the IETF and the
> W3C, it seems that connection strategy, proxy traversal strategy and
> wire protocol would fall under the IETF whereas JavaScript APIs would
> fall under the W3C.
> 
> Therefore, assuming that the above separation of topics is useful, I'm
> planning to wait a day or so to give folks a chance to respond, then
> start a thread on each of the topics that pertain to the IETF.
> 
> Kind Regards,
> John Fallows.
> 
> Disclaimer:  Kaazing currently offers (emulated) WebSocket-based
> products that implement both the wire protocol (when possible,
> otherwise fallback to HTTP) and JavaScript API from a draft of the
> specification(s).  Our goal in participating with the IETF and the W3C
> is offer insights gained from customer feedback to help standardize
> the future of real-time communications on the Web.
> 

