
Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5SJo3E5046603; Tue, 28 Jun 2005 12:50:03 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5SJo3GZ046602; Tue, 28 Jun 2005 12:50:03 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from newodin.ietf.org ([132.151.6.50]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5SJo2Ym046594 for <ietf-imapext@imc.org>; Tue, 28 Jun 2005 12:50:03 -0700 (PDT) (envelope-from mlee@newodin.ietf.org)
Received: from mlee by newodin.ietf.org with local (Exim 4.43) id 1DnM5q-000887-MN; Tue, 28 Jun 2005 15:50:02 -0400
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
To: i-d-announce@ietf.org
Cc: ietf-imapext@imc.org
From: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-imapext-2086upd-08.txt 
Message-Id: <E1DnM5q-000887-MN@newodin.ietf.org>
Date: Tue, 28 Jun 2005 15:50:02 -0400
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Internet Message Access Protocol Extension Working Group of the IETF.

	Title		: IMAP4 ACL extension
	Author(s)	: A. Melnikov
	Filename	: draft-ietf-imapext-2086upd-08.txt
	Pages		: 29
	Date		: 2005-6-28
	
The ACL (Access Control List) extension (RFC 2086) of the Internet
   Message Access Protocol (IMAP) permits mailbox access control lists
   to be retrieved and manipulated through the IMAP protocol.

   This document is a revision of RFC 2086.  It defines several new
   access control rights and clarifies which rights are required for
   different IMAP commands.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-imapext-2086upd-08.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-imapext-2086upd-08.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-imapext-2086upd-08.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<2005-6-28150727.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-imapext-2086upd-08.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-imapext-2086upd-08.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<2005-6-28150727.I-D@ietf.org>

--OtherAccess--

--NextPart--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFxGCa007874; Fri, 24 Jun 2005 08:59:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFxGJu007871; Fri, 24 Jun 2005 08:59:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFxF5E007794 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:59:15 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so201zzb for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Received: by 172.25.12.97 with SMTP id 17mr57zzb; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Message-ID: <5b93232d05062408593a7c0f83@mail.google.com>
Date: Fri, 24 Jun 2005 08:59:09 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Alexey Melnikov <alexey.melnikov@isode.com>
Subject: Re: user names vs. identifiers
Cc: Lisa Dusseault <lisa@osafoundation.org>, Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <42BC2D45.3080407@isode.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com> <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org> <42BC2D45.3080407@isode.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5OFxF5E007857
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/24/05, Alexey Melnikov <alexey.melnikov@isode.com> wrote:
> Lisa Dusseault wrote:
> 
> >
> > What about some wording up front (before the word is first used) that
> > explains once that "identifier" is used for access identifiers which
> > may be userids or other authorization tokens?
> 
> I would rather do what you suggest, as this minimizes the amount of
> changes to the document and thus it is less prone to introducing errors.

I agree, this seems much less invasive and still correct.

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFv0Cx005789; Fri, 24 Jun 2005 08:57:00 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFv0fH005788; Fri, 24 Jun 2005 08:57:00 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFuxRF005764 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:56:59 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.176] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Fri, 24 Jun 2005 16:56:52 +0100
Message-ID: <42BC2D45.3080407@isode.com>
Date: Fri, 24 Jun 2005 16:56:53 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Lisa Dusseault <lisa@osafoundation.org>
CC: Rob Siemborski <robsiemb@google.com>, Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: user names vs. identifiers
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com> <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
In-Reply-To: <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Lisa Dusseault wrote:

>
> What about some wording up front (before the word is first used) that 
> explains once that "identifier" is used for access identifiers which 
> may be userids or other authorization tokens?

I would rather do what you suggest, as this minimizes the amount of 
changes to the document and thus it is less prone to introducing errors.

> lisa
>
> On Jun 24, 2005, at 8:12 AM, Rob Siemborski wrote:
>
>>
>> On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
>>
>>> On Thu, 23 Jun 2005, Rob Siemborski wrote:
>>>
>>>> In the case of the ACL document, this isn't accurate, as the
>>>> identifiers can refer to groups as well as usernames/userids.
>>>
>>>
>>> Suggestion: the term "access identifier" should be used instead.  This
>>> term is used by the URLAUTH document as well, but the two documents 
>>> don't
>>> overlap in this area so the term can be reused.
>>
>>
>> Sure, this works for me.
>>
>>> "Access identifier" should certainly be defined so that one of its
>>> possible values is "userid".  AFAICT, all userids are access 
>>> identifiers
>>> for ACL purposes, but not all access identifiers are userids, right?
>>
>>
>> Yes.
>


-- 
Alexey Melnikov
__________________________________________
Isode Mbox Message Store developer

Isode Limited, http://www.isode.com

IETF standard related pages:
http://www.melnikov.ca/mel/devel/Links.html

Personal Home Page: http://www.melnikov.ca
__________________________________________



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFhU5X090987; Fri, 24 Jun 2005 08:43:30 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFhUd6090984; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFhU4L090971 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:43:30 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 367D2142264; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 28569-04; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id BBFA314225E; Fri, 24 Jun 2005 08:43:29 -0700 (PDT)
In-Reply-To: <42BBDAAA.50508@isode.com>
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org> <42BBDAAA.50508@isode.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <865da06559bc551eb0e875fa72c135c5@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Fri, 24 Jun 2005 08:43:27 -0700
To: Alexey Melnikov <alexey.melnikov@isode.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Jun 24, 2005, at 3:04 AM, Alexey Melnikov wrote:

>>
>> So, another possible sentence to add is something along the lines of 
>> "When an IMAP server is responsible for creating user names for new 
>> accounts, the server SHOULD NOT create such ambiguous user names."
>
> Regarding this sentence in particular: user management is outside the 
> scope of the 2086upd, so I don't think this sentence belongs to the 
> document.

That's just fine by me, I see your point.
Lisa



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFghBJ090156; Fri, 24 Jun 2005 08:42:43 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFghZ8090155; Fri, 24 Jun 2005 08:42:43 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFgfhQ090033 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:42:41 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 2698F142264; Fri, 24 Jun 2005 08:42:41 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 24565-08; Fri, 24 Jun 2005 08:42:40 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 6CF6914225E; Fri, 24 Jun 2005 08:42:40 -0700 (PDT)
In-Reply-To: <5b93232d050624081278553f7e@mail.google.com>
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: user names vs. identifiers
Date: Fri, 24 Jun 2005 08:42:38 -0700
To: Rob Siemborski <robsiemb@google.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

What about some wording up front (before the word is first used) that 
explains once that "identifier" is used for access identifiers which 
may be userids or other authorization tokens?

lisa

On Jun 24, 2005, at 8:12 AM, Rob Siemborski wrote:

>
> On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
>> On Thu, 23 Jun 2005, Rob Siemborski wrote:
>>> In the case of the ACL document, this isn't accurate, as the
>>> identifiers can refer to groups as well as usernames/userids.
>>
>> Suggestion: the term "access identifier" should be used instead.  This
>> term is used by the URLAUTH document as well, but the two documents 
>> don't
>> overlap in this area so the term can be reused.
>
> Sure, this works for me.
>
>> "Access identifier" should certainly be defined so that one of its
>> possible values is "userid".  AFAICT, all userids are access 
>> identifiers
>> for ACL purposes, but not all access identifiers are userids, right?
>
> Yes.
>
> -Rob
>
> -- 
> Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
> Software Engineer     |                       |            650-623-6925
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFCqR3063038; Fri, 24 Jun 2005 08:12:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFCprJ063029; Fri, 24 Jun 2005 08:12:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFCoiH062949 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:12:50 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so165zzb for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Received: by 172.25.12.188 with SMTP id 28mr49zzc; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Message-ID: <5b93232d050624081278553f7e@mail.google.com>
Date: Fri, 24 Jun 2005 08:12:45 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: user names vs. identifiers
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5OFCoiH063017
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
> On Thu, 23 Jun 2005, Rob Siemborski wrote:
> > In the case of the ACL document, this isn't accurate, as the
> > identifiers can refer to groups as well as usernames/userids.
> 
> Suggestion: the term "access identifier" should be used instead.  This
> term is used by the URLAUTH document as well, but the two documents don't
> overlap in this area so the term can be reused.

Sure, this works for me.

> "Access identifier" should certainly be defined so that one of its
> possible values is "userid".  AFAICT, all userids are access identifiers
> for ACL purposes, but not all access identifiers are userids, right?

Yes.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OA6YUQ008917; Fri, 24 Jun 2005 03:06:34 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OA4kWB006820; Fri, 24 Jun 2005 03:04:46 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OA4TRR006350 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 03:04:30 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Fri, 24 Jun 2005 11:04:26 +0100
Message-ID: <42BBDAAA.50508@isode.com>
Date: Fri, 24 Jun 2005 11:04:26 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Lisa Dusseault <lisa@osafoundation.org>
CC: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
In-Reply-To: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Lisa Dusseault wrote:

> On Jun 23, 2005, at 9:34 AM, Alexey Melnikov wrote:
>
>> ...
>>
>> So, I think this is a SASLPrep issue and should be fixed there. Two 
>> other alternatives for 2086upd I see:
>> 1). override SASLPrep rules to say that PR29 should be used (the 
>> proper procedure is described there).
>> 2). say that an 2086upd server MAY reject such sequences.
>> Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
>> sure that our document is the proper place for the fix.
>
> I would prefer to add something along the lines of #2.  Some of my 
> reasoning:
>  - Although we don't have to say that the server MAY reject such user 
> names, because there's nothing elsewhere in the document that forbids 
> it, it does make it clear what one possible, immediate, practical 
> solution to the problem is.
>  - I'd rather not have the suggested solution be merely  to wait on 
> future speculative work.
>  - I don't see any drawbacks to adding such a MAY clause.

Ok, you've convinced me that this solution is fine. My proposed text 
should still work. I just need to add a sentence about "MAY reject such 
identifiers".

> The problem is of course not entirely solvable by 2086upd, although we 
> can do our best by drawing attention to it.  If I understand 
> correctly, the software that typically creates the problem by creating 
> ambiguous user names is not IMAP clients, and not always IMAP servers, 
> but often directory and authentication services -- and the 
> implementors of these services probably have no need to read 2086bis.
>
> So, another possible sentence to add is something along the lines of 
> "When an IMAP server is responsible for creating user names for new 
> accounts, the server SHOULD NOT create such ambiguous user names."

Regarding this sentence in particular: user management is outside the 
scope of the 2086upd, so I don't think this sentence belongs to the 
document.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6oU3i044161; Thu, 23 Jun 2005 23:50:30 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O6oUFX044160; Thu, 23 Jun 2005 23:50:30 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6oT1K044148 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:50:29 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O6oRDj017601 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 23 Jun 2005 23:50:27 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O6oPwK024604 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 23 Jun 2005 23:50:27 -0700
Date: Thu, 23 Jun 2005 23:50:25 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Rob Siemborski <robsiemb@google.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: user names vs. identifiers
In-Reply-To: <5b93232d05062323352dbc2c68@mail.google.com>
Message-ID: <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 23 Jun 2005, Rob Siemborski wrote:
> In the case of the ACL document, this isn't accurate, as the
> identifiers can refer to groups as well as usernames/userids.

Suggestion: the term "access identifier" should be used instead.  This 
term is used by the URLAUTH document as well, but the two documents don't 
overlap in this area so the term can be reused.

"Access identifier" should certainly be defined so that one of its 
possible values is "userid".  AFAICT, all userids are access identifiers 
for ACL purposes, but not all access identifiers are userids, right?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6ZEUb031130; Thu, 23 Jun 2005 23:35:14 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O6ZEMs031129; Thu, 23 Jun 2005 23:35:14 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6ZDTn031065 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:35:14 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so101zzb for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Received: by 172.25.12.188 with SMTP id 28mr26zzc; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Message-ID: <5b93232d05062323352dbc2c68@mail.google.com>
Date: Thu, 23 Jun 2005 23:35:08 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: user names vs. identifiers
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5O6ZETn031122
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
> I agree that "user name" should be deprecated, although I would advocate
> "userid" or "user identifier" instead of plain "identifier".

In the case of the ACL document, this isn't accurate, as the
identifiers can refer to groups as well as usernames/userids.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O38N2D079107; Thu, 23 Jun 2005 20:08:23 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O38Not079106; Thu, 23 Jun 2005 20:08:23 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O38NOn079091 for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:23 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O38MRc004374 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:22 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O38K5x004723 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:21 -0700
Date: Thu, 23 Jun 2005 20:08:19 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: user names vs. identifiers
Message-ID: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

For what it's worth, RFC 3501 uses both "user name" and "userid"; 
"identifier" is used primarily in the context of "unique identifier" 
although once is used to refer to a charset identifier.

The URLAUTH document in Lemonade uses userid throughout, except in one 
error mesage where it uses "user name".

I agree that "user name" should be deprecated, although I would advocate 
"userid" or "user identifier" instead of plain "identifier".

Another problem is such terms as "current userid", etc. when what is 
really intended is this session's authorization identity.  I think that 
it's probably alright to use "current userid" as long as we define 
someplace that this refers to the authorization identity.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O2XPu6056963; Thu, 23 Jun 2005 19:33:25 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O2XPUJ056962; Thu, 23 Jun 2005 19:33:25 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O2XOCV056917 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 19:33:24 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so9zzb for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 19:33:19 -0700 (PDT)
Received: by 172.25.12.97 with SMTP id 17mr3zzb; Thu, 23 Jun 2005 19:33:19 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Thu, 23 Jun 2005 19:33:18 -0700 (PDT)
Message-ID: <5b93232d0506231933df1c637@mail.google.com>
Date: Thu, 23 Jun 2005 19:33:18 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Cc: Alexey Melnikov <alexey.melnikov@isode.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5O2XPCV056956
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Lisa Dusseault <lisa@osafoundation.org> wrote:
> So, another possible sentence to add is something along the lines of
> "When an IMAP server is responsible for creating user names for new
> accounts, the server SHOULD NOT create such ambiguous user names."

We should be sure that however we fix this we refer to 'identifiers'
and not 'user names'.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NHWGbc045923; Thu, 23 Jun 2005 10:32:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NHWGdG045922; Thu, 23 Jun 2005 10:32:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NHWG6d045904 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:32:16 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 1429114225D; Thu, 23 Jun 2005 10:32:39 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 15041-03; Thu, 23 Jun 2005 10:32:38 -0700 (PDT)
Received: from [192.168.101.178] (w002.z065106067.sjc-ca.dsl.cnc.net [65.106.67.2]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id CC99C14225B; Thu, 23 Jun 2005 10:32:38 -0700 (PDT)
In-Reply-To: <42BAE4A7.9080004@isode.com>
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Thu, 23 Jun 2005 10:32:12 -0700
To: Alexey Melnikov <alexey.melnikov@isode.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Jun 23, 2005, at 9:34 AM, Alexey Melnikov wrote:

> ...
>
> So, I think this is a SASLPrep issue and should be fixed there. Two 
> other alternatives for 2086upd I see:
> 1). override SASLPrep rules to say that PR29 should be used (the 
> proper procedure is described there).
> 2). say that an 2086upd server MAY reject such sequences.
> Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
> sure that our document is the proper place for the fix.

I would prefer to add something along the lines of #2.  Some of my 
reasoning:
  - Although we don't have to say that the server MAY reject such user 
names, because there's nothing elsewhere in the document that forbids 
it, it does make it clear what one possible, immediate, practical 
solution to the problem is.
  - I'd rather not have the suggested solution be merely  to wait on 
future speculative work.
  - I don't see any drawbacks to adding such a MAY clause.

The problem is of course not entirely solvable by 2086upd, although we 
can do our best by drawing attention to it.  If I understand correctly, 
the software that typically creates the problem by creating ambiguous 
user names is not IMAP clients, and not always IMAP servers, but often 
directory and authentication services -- and the implementors of these 
services probably have no need to read 2086bis.

So, another possible sentence to add is something along the lines of 
"When an IMAP server is responsible for creating user names for new 
accounts, the server SHOULD NOT create such ambiguous user names."

Lisa


>
> Assuming we go with "punt the issue to SASLPrep", I suggest to add the 
> following text to the Security Considerations section:
>
>     This document relies on [SASLPrep] to describe steps required to 
> perform user name
>     canonicalization (preparation). The preparation algorithm in 
> SASLPrep was specifically
>     designed such that its output is canonical, and it is well-formed.
>     However, due to an anomaly [PR29] in the specification of Unicode
>     normalization, canonical equivalence is not guaranteed for a select
>     few character sequences.
>
>     User names prepared with SASLPrep can be stored and returned
>     by an ACL server. The anomaly affects ACL manipulation and 
> evaluation of
>     user names containing the selected character sequences. These 
> sequences, however,
>     do not appear in well-formed text. This specification was 
> published despite this
>     known technical problem. It is expected that the problem will be
>     addressed in a future revision of [Unicode] and/or [StringPrep]
>     specifications.
>
> Comments?
>
> Alexey
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NH1lFN012646; Thu, 23 Jun 2005 10:01:47 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NH1lXv012645; Thu, 23 Jun 2005 10:01:47 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from numenor.qualcomm.com (numenor.qualcomm.com [129.46.51.58]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NH1leA012638 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:47 -0700 (PDT) (envelope-from ameylan@qualcomm.com)
Received: from crowley.qualcomm.com (crowley.qualcomm.com [129.46.61.151]) by numenor.qualcomm.com (8.12.10/8.12.5/1.0) with ESMTP id j5NH1egj027659 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL) for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:40 -0700 (PDT)
Received: from AMEYLAN.qualcomm.com (ameylan.na.qualcomm.com [129.46.159.213]) by crowley.qualcomm.com (8.12.10/8.12.5/1.0) with ESMTP id j5NH1cv6003414 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:38 -0700 (PDT)
Message-Id: <6.2.2.1.2.20050623100115.03d0b0f0@unixmail.qualcomm.com>
X-Mailer: QUALCOMM Windows Eudora Version 6.2.2.1
Date: Thu, 23 Jun 2005 10:01:34 -0700
To: ietf-imapext@imc.org
From: Arnaud Meylan <ameylan@qualcomm.com>
Subject: 
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

unsubscribe ameylan@qualcomm.com



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NGZIrr086764; Thu, 23 Jun 2005 09:35:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NGZIDx086763; Thu, 23 Jun 2005 09:35:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NGZHSL086757 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 09:35:18 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.177] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 23 Jun 2005 17:34:46 +0100
Message-ID: <42BAE4A7.9080004@isode.com>
Date: Thu, 23 Jun 2005 17:34:47 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
In-Reply-To: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Scott Hollenbeck wrote:

>The IESG evaluation of draft-ietf-imapext-2086upd-07 has completed.  There's
>only one discuss comment that we started to talk about.  Revised ID needed.
>
>Lisa: as document shepherd, please work with Russ and Alexey to come up with
>appropriate text.
>
>  
>
After reading <http://www.unicode.org/review/pr-29.html> I think I 
understand the issue now (please let me know if this is not the case). 
Here is a scenario when this issue might be relevant to the document:

An 2086upd client adds permissions for user 
"<U+1100><U+0300><U+1161><U+0323>" (the value is taken from PR29 and is 
contrived) using SETACL command. The server performs normalization as 
described in SASLprep and will store permissions for the username 
"<U+AC00><U+0300><U+0323>".

The client then lists permissions (for display purposes) on the mailbox 
using GETACL. The server returns canonicalized name 
"<U+AC00><U+0300><U+0323>".

The user then decides that the permissions are not right, so the client 
will update the permissions for "<U+AC00><U+0300><U+0323>". The server 
performs canonicalization which results in username 
"<U+AC00><U+0323><U+0300>", which is not the same as 
"<U+AC00><U+0300><U+0323>". The server will fail to treat them as the 
same, so instead of replacing permissions on "<U+AC00><U+0300><U+0323>", 
it will add permissions for "<U+AC00><U+0323><U+0300>".

The problem for the user is that all 3 variants should be displayed 
exactly the same.
The problem for access calculation is that the result will depend on the 
form of identifier used. This can be unexpected at best.

The good news is that such sequences are very-very-unlikely to happen in 
real life.

So, I think this is a SASLPrep issue and should be fixed there. Two 
other alternatives for 2086upd I see:
1). override SASLPrep rules to say that PR29 should be used (the proper 
procedure is described there).
2). say that an 2086upd server MAY reject such sequences.
Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
sure that our document is the proper place for the fix.

Assuming we go with "punt the issue to SASLPrep", I suggest to add the 
following text to the Security Considerations section:

     This document relies on [SASLPrep] to describe steps required to 
perform user name
     canonicalization (preparation). The preparation algorithm in 
SASLPrep was specifically
     designed such that its output is canonical, and it is well-formed.
     However, due to an anomaly [PR29] in the specification of Unicode
     normalization, canonical equivalence is not guaranteed for a select
     few character sequences.

     User names prepared with SASLPrep can be stored and returned
     by an ACL server. The anomaly affects ACL manipulation and 
evaluation of
     user names containing the selected character sequences. These 
sequences, however,
     do not appear in well-formed text. This specification was published 
despite this
     known technical problem. It is expected that the problem will be
     addressed in a future revision of [Unicode] and/or [StringPrep]
     specifications.

Comments?

Alexey



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NG9KrD055882; Thu, 23 Jun 2005 09:09:20 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NG9KjJ055876; Thu, 23 Jun 2005 09:09:20 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NG9Heo055763 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 09:09:19 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Thu, 23 Jun 2005 12:09:12 -0400 id 0059C1E8.42BADEA8.000069DF
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: "'Alexey Melnikov'" <alexey.melnikov@isode.com>, lisa@osafoundation.org
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Thu, 23 Jun 2005 12:08:52 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
Thread-Index: AcV4DdkWELCFMBPyS2qQvyUyoUfwrQ==
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Message-ID: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

The IESG evaluation of draft-ietf-imapext-2086upd-07 has completed.  There's
only one discuss comment that we started to talk about.  Revised ID needed.

Lisa: as document shepherd, please work with Russ and Alexey to come up with
appropriate text.

-Scott-



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5LBBMgB028665; Tue, 21 Jun 2005 04:11:22 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5LBBM3E028664; Tue, 21 Jun 2005 04:11:22 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5LBBLUC028648 for <ietf-imapext@imc.org>; Tue, 21 Jun 2005 04:11:21 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Tue, 21 Jun 2005 07:11:20 -0400 id 0059C319.42B7F5D8.00003168
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: ietf-imapext@imc.org
Subject: FW: DISCUSS: draft-ietf-imapext-2086upd-07
Date: Tue, 21 Jun 2005 07:11:07 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Thread-Index: AcV13Do5bZLxzRHfSj2hlkIUvWFDJAAdWYSw
Message-ID: <courier.42B7F5D8.00003168@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

FYI -- here's an IESG evaluation comment that will need to be resolved.
More may come before Thursday.  Do NOT change the document now.  We can talk
about the needed text and talk to Russ to find something he's happy with.

-Scott-

-----Original Message-----
From: Russ Housley [mailto:housley@vigilsec.com] 
Sent: Monday, June 20, 2005 4:56 PM
To: iesg@ietf.org
Subject: DISCUSS: draft-ietf-imapext-2086upd-07

IMAP4 ACL extension (Proposed Standard)

DISCUSS

   This specification depends on SASLprep, which is defined in RFC 4013.
   The security considerations of RFC 4013 say:
   >
   > This profile is intended to prepare simple user name and password
   > strings for comparison or use in cryptographic functions (e.g.,
   > message digests).  The preparation algorithm was specifically
   > designed such that its output is canonical, and it is well-formed.
   > However, due to an anomaly [PR29] in the specification of Unicode
   > normalization, canonical equivalence is not guaranteed for a select
   > few character sequences.  These sequences, however, do not appear in
   > well-formed text.  This specification was published despite this
   > known technical problem.  It is expected that this specification will
   > be revised before further progression on the Standards Track (after
   > [Unicode] and/or [StringPrep] specifications have been updated to
   > address this problem).
   >
   The security considerations of this document need to address this point.
   How does this situation impact ACL processing?





Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5DJe89l078654; Mon, 13 Jun 2005 12:40:08 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5DJe8QK078653; Mon, 13 Jun 2005 12:40:08 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from ietf.org (odin.ietf.org [132.151.1.176]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5DJe7Em078646 for <ietf-imapext@imc.org>; Mon, 13 Jun 2005 12:40:08 -0700 (PDT) (envelope-from dinaras@cnri.reston.va.us)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15881; Mon, 13 Jun 2005 15:40:05 -0400 (EDT)
Message-Id: <200506131940.PAA15881@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: i-d-announce@ietf.org
Cc: ietf-imapext@imc.org
From: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-imapext-2086upd-07.txt
Date: Mon, 13 Jun 2005 15:40:05 -0400
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Internet Message Access Protocol Extension Working Group of the IETF.

	Title		: IMAP4 ACL extension
	Author(s)	: A. Melnikov
	Filename	: draft-ietf-imapext-2086upd-07.txt
	Pages		: 28
	Date		: 2005-6-13
	
The ACL (Access Control List) extension (RFC 2086) of the Internet
   Message Access Protocol (IMAP) permits mailbox access control lists
   to be retrieved and manipulated through the IMAP protocol.

   This document is a revision of RFC 2086.  It defines several new
   access control rights and clarifies which rights are required for
   different IMAP commands.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-imapext-2086upd-07.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-imapext-2086upd-07.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-imapext-2086upd-07.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<2005-6-13160813.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-imapext-2086upd-07.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-imapext-2086upd-07.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<2005-6-13160813.I-D@ietf.org>

--OtherAccess--

--NextPart--




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AvWhN084164; Thu, 9 Jun 2005 03:57:32 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59AvWQH084163; Thu, 9 Jun 2005 03:57:32 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AvVXZ084153 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:57:31 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.110] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 11:57:28 +0100
Message-ID: <42A82098.4010306@isode.com>
Date: Thu, 09 Jun 2005 11:57:28 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Dave Cridland <dave@cridland.net>
CC: Barry Leiba <leiba@watson.ibm.com>, IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com> <28904.1118310684.785236@localhost.localdomain>
In-Reply-To: <28904.1118310684.785236@localhost.localdomain>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Dave Cridland wrote:

> On Wed Jun  8 19:27:01 2005, Barry Leiba wrote:
>
>>>> - section 6:  should we restrict selection options to be simple atoms
>>>> as we have in this draft, or should we allow for them to be a list as
>>>> well?   I'm thinking that some future selection option might take one
>>>> or more arguments.  For example, the non-standard SCAN extension used
>>>> by Pine and UW-IMAP could be reworked to look something like this:
>>>>
>>>> x LIST ((SCAN "foo")) "" "bar/%"
>>>>
>>>> or better yet:
>>>>
>>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
>>>
>>
>> No responses.......
>> Since I think it adds too much complexity, unless there's some 
>> support of Ken's suggestion Very Soon, I will toss the idea, and buy 
>> Ken a glass of Bordeaux in Paris as compensation.
>
> I can see some uses for this - as well as SCAN, there's the discussion 
> of a CHANGEDSINCE possibility, and I think Cyrus mentioned something 
> about searching mailbox-level annotations, which could use this kind 
> of syntax.

A clarification to my previous post: as any unrecognized criterion 
should cause BAD response anyway, it is Ok not to define selection 
criteria parameters in the LISTEXT document, even though it would be 
nice to have either generic syntax or some recommendations on what 
should be used by extensions.

> However, I'm assuming a later specification could update the syntax to 
> include whatever we find we need - is there anything that would 
> prevent that?

Nothing I can think of at the moment.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AhmFN079427; Thu, 9 Jun 2005 03:43:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59AhmtW079426; Thu, 9 Jun 2005 03:43:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AhlIv079415 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:43:47 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.110] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 11:43:43 +0100
Message-ID: <42A81D60.3030503@isode.com>
Date: Thu, 09 Jun 2005 11:43:44 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Philip Guenther <guenther+imap@sendmail.com>
CC: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: comments on IMAP-ABNF extensions draft -02
References: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
In-Reply-To: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Philip Guenther wrote:

>I'm not convinced that this draft is taking the best course for defining
>the syntax for extensions.  I see this document as a place to tweak how
>the IMAP grammar is expressed to make extensions easier to express,
>*without actually actually extending it*.  That is, extensions should be
>able express their syntaxes as amending the grammar in RFC 3501 and this
>draft by incremental alternative ("=/") only and without replacing
>grammar elements defined in RFC 3501 or this draft, preferably without
>having to repeat redundant tokens.
>
>And yes, this goal is of course limited by our ability to predict how
>commands will be extended, but the key point is what it excludes:
>servers are *never* required to accept syntax for extensions that they
>don't support**.  This contrasts against the (few) places where the RFC
>3501 grammar *requires* clients to unilaterally accept extensible data
>with open-ended semantics: body-extension, flag-extension, and
>resp-text-code.  As currently written, several of the rules in this
>draft have an open-ended syntax that servers would apparently have to
>parse, only so that they can reject them with a BAD.
>  
>
I don't think that this statement is true, at least it wasn't the intent.

[...]

>IMHO, the tagged-ext syntax (or something similarly simple) is Good
>Enough to be referenced by FETCH, STORE, SELECT, and SEARCH too.  It
>certainly appears to cover the already scoped extensions for CONDSTORE,
>ANNOTATE, and several others. Using a single syntax for extensions to
>commands would make the syntax part of this draft easier to read.
>
Yes, I know. Apart from the SEARCH command extensions the grammer in the 
document is taken from other documents like CONDSTORE, ANNOTATE, etc.
What I've discovered was that all commands *were* extended differently, 
which is a pitty.

>If it *isn't* sufficient, then I suspect the Right Thing might be to simply
>say that extensions should consist of a atom as a tag, followed by
>exactly one argument with balanced parenthesis.
>
Yes, this can be done for the newly defined things.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59ALpiX072082; Thu, 9 Jun 2005 03:21:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59ALp2Q072081; Thu, 9 Jun 2005 03:21:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout5.cac.washington.edu (mxout5.cac.washington.edu [140.142.32.135]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59ALo7G072073 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:21:50 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout5.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j59ALfrM019510 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 9 Jun 2005 03:21:42 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j59ALc6u027146 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 9 Jun 2005 03:21:41 -0700
Date: Thu, 9 Jun 2005 03:21:37 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118309793.629223@localhost.localdomain>
Message-ID: <Pine.OSX.4.63.0506090317440.12669@pangtzu.panda.com>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>   <28904.1118220817.700674@localhost.localdomain>   <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>  <28904.1118258976.392523@localhost.localdomain>  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118309793.629223@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 9 Jun 2005, Dave Cridland wrote:
> Of course, you're right in that we require the ability to base the value on 
> stat() time_t values for some implementations, and I agree we should provide 
> recommendations for implementors choosing this method, but it is and should 
> remain an implementation issue, and confined to an appendix.

I have no problem with this, as long as the word "require" above remains 
operative.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59A01L1063913; Thu, 9 Jun 2005 03:00:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59A01qp063912; Thu, 9 Jun 2005 03:00:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xxKs063889 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:00:00 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:59:52 +0100
Message-ID: <42A8130E.2060301@isode.com>
Date: Thu, 09 Jun 2005 10:59:42 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Barry Leiba <leiba@watson.ibm.com>
CC: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba wrote:

>I've made the editorial changes; thanks.
>  
>
>>- section 6:  should we restrict selection options to be simple atoms
>>as we have in this draft, or should we allow for them to be a list as
>>well?   I'm thinking that some future selection option might take one
>>or more arguments.  For example, the non-standard SCAN extension used
>>by Pine and UW-IMAP could be reworked to look something like this:
>>
>>x LIST ((SCAN "foo")) "" "bar/%"
>>
>>or better yet:
>>
>>x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
>>    
>>
>
>I have no objection, myself, to extending things further in this way,
>but I point out that it *is* a significant additional extension, not
>just a simple issue of grammar.  Now there may be many more things to
>test in the selection mechanism, for one thing -- it's not just
>inspecting whether the mailbox matches some "attribute" (like REMOTE or
>SUBSCRIBED), but also whether it meets some other open-ended criteria.
>
Yes, and I believe several people have proposed this already. So 
ignoring the open ended syntax issue: this is something that should be 
allowed by the spec.

>The grammar may also be open-ended (as in your second example; why stop
>at two astrings?), which can cause us problems.
>
>Also also, we now have a complication with boolean conditions.  What
>might "(SUBSCRIBED (SCAN "foo") (SCAN "bar"))" mean?
>
>We should do this right, while we're doing it.  I'm just not sure, for
>this situation, what "right" is.  My sense, from talking with some of
>the client authors (mostly with Cyrus), is that this goes beyond what
>they need this for.
>  
>
Actually Dave Cridland wanted to have parameters to selection criteria, 
for example to list all mailboxes changed since MODSEQ. So, if we can 
nail the basic syntax now, that would be good.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xuDw063874; Thu, 9 Jun 2005 02:59:56 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599xtWo063872; Thu, 9 Jun 2005 02:59:55 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xtpG063824 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:59:55 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 5881B154009; Thu,  9 Jun 2005 09:59:47 +0000 (UTC)
Date: Thu, 09 Jun 2005 10:51:24 +0100
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
MIME-Version: 1.0
Message-Id: <28904.1118310684.785236@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Barry Leiba <leiba@watson.ibm.com>
Cc: IMAP extensions Mailing List <ietf-imapext@imc.org>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 19:27:01 2005, Barry Leiba wrote:
> 
>>> - section 6:  should we restrict selection options to be simple 
>>> atoms
>>> as we have in this draft, or should we allow for them to be a 
>>> list as
>>> well?   I'm thinking that some future selection option might take 
>>> one
>>> or more arguments.  For example, the non-standard SCAN extension 
>>> used
>>> by Pine and UW-IMAP could be reworked to look something like this:
>>> 
>>> x LIST ((SCAN "foo")) "" "bar/%"
>>> 
>>> or better yet:
>>> 
>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
> 
> No responses.......
> Since I think it adds too much complexity, unless there's some 
> support of Ken's suggestion Very Soon, I will toss the idea, and 
> buy Ken a glass of Bordeaux in Paris as compensation.

I can see some uses for this - as well as SCAN, there's the 
discussion of a CHANGEDSINCE possibility, and I think Cyrus mentioned 
something about searching mailbox-level annotations, which could use 
this kind of syntax.

However, I'm assuming a later specification could update the syntax 
to include whatever we find we need - is there anything that would 
prevent that?

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599oqTW060829; Thu, 9 Jun 2005 02:50:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599oqka060828; Thu, 9 Jun 2005 02:50:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599oqFw060816 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:50:52 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:50:50 +0100
Message-ID: <42A810FA.2060304@isode.com>
Date: Thu, 09 Jun 2005 10:50:50 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
CC: ietf-imapext@imc.org, Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, Dave Cridland <dave@cridland.net>
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <1118233852.29860.17.camel@localhost.localdomain> <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
In-Reply-To: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Arnt Gulbrandsen wrote:

> Timo Sirainen writes:
>
>> If we assume that timestamp can be increased after each modification, 
>> can't we just assume that timestamp is one implementation of 
>> CONDSTORE's MODSEQ? In that case we don't need a whole new timestamp 
>> thing, just a possibility to get latest MODSEQ from LIST reply.
>
> Now there's a good idea. How much baggage does it drag along - how 
> much of CONDSTORE has to be provided to provide MODSEQ?

Server code needs to update MODSEQs on each STORE operation. That should 
be sufficient.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599j387058874; Thu, 9 Jun 2005 02:45:03 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599j3Pm058873; Thu, 9 Jun 2005 02:45:03 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599j2Y0058831 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:45:02 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id D0F63154009; Thu,  9 Jun 2005 09:44:55 +0000 (UTC)
Date: Thu, 09 Jun 2005 10:36:33 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>   <28904.1118220817.700674@localhost.localdomain>   <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>  <28904.1118258976.392523@localhost.localdomain>  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To:  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
MIME-Version: 1.0
Message-Id: <28904.1118309793.629223@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <MRC@CAC.Washington.EDU>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu Jun  9 00:20:57 2005, Mark Crispin wrote:
> On Wed, 8 Jun 2005, Dave Cridland wrote:
>> b) Nothing can show the actual time anyway, at least with the 
>> technology we had back in 2005.
> 
> Is this really true?  Suppose the client wants to show a "human 
> readable" last date/time?
> 
> 
I believe it is true. I also believe it's an orthogonal problem. Any 
direct, protocol-specified use of timestamps as modification 
indicators (as opposed to their use as timestamps) would inevitably 
cause problems. Not least, didn't those idiots back in 2005 realise 
that we'd be updating mailboxes faster than once per millisecond?


>> If ever we need to know the system time of a particular update, 
>> we'll need to do that in a distinct way anyway - system time can, 
>> under some circumstances, go backwards. (Or it could in 2005, 
>> anyway.)
> 
> Leaving aside the question of drifting clocks, there's still the 
> question of a "human readable" last date/time for user display.
> 
> 
. STATUS INBOX (TIMESTAMP)
* STATUS INBOX (TIMESTAMP "09-Jun-2005 09:32:00+0000"
. OK STATUS completed.
. FETCH 1 TIMESTAMP
* 1 FETCH (TIMESTAMP "09-Jun-2005 09:32:00+0000")
. OK FETCH completed.

This does assume that humans aren't interested in millisecond 
accuracy, of course.

But that's no use to test for modifications, purely for use to 
display to 2005-era humans.


>>> I like the idea of using a CONDSTORE type mechanism, as long as 
>>> it can be implemented using a file date from the operating system 
>>> (e.g. from the stat() system call on UNIX).  Maybe not a "best" 
>>> implementation (in that there may be requirements for 
>>> synchronization delays if the value's resolution is too low) but 
>>> nonetheless a viable and compliant implementation.
>> Your earlier description of taking a timestamp, multiplying it to 
>> emulate an increased resolution, and "making up" the extra 
>> resolution to fit the specification, ought to work fine.
> 
> We also need some sort of stipulation that this value is not 
> allowed to go backwards, and recommendations for preventing that 
> from happening, such as detecting an old time-based value that is 
> in the future, and monotonically increasing it on update rather 
> than setting it to time.

Yes, we need this, but it damages the relationship between the 64-bit 
modseq and the actual time, which in turn means that clients cannot 
rely on it to provide a human-readable, accurate time.

I really believe that timestamps are irrelevant to the actual 
protocol here - all we need is some kind of data type which is 
comparable, either less than, greater than, or equal to. A 64-bit 
integer is fast, and has precendence in CONDSTORE.

Of course, you're right in that we require the ability to base the 
value on stat() time_t values for some implementations, and I agree 
we should provide recommendations for implementors choosing this 
method, but it is and should remain an implementation issue, and 
confined to an appendix.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599WtXe054888; Thu, 9 Jun 2005 02:32:55 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599WtMH054887; Thu, 9 Jun 2005 02:32:55 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599WrkC054868 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:32:54 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:32:49 +0100
Message-ID: <42A80CC1.8080001@isode.com>
Date: Thu, 09 Jun 2005 10:32:49 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Mark Crispin <mrc@cac.washington.edu>
CC: Dave Cridland <dave@cridland.net>, IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain> <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin wrote:

> On Wed, 8 Jun 2005, Dave Cridland wrote:
>
>> The client only needs to know that a change has happened, never the 
>> actual times involved.
>
> The question is if a client ever will care about the time.  I don't 
> want to have to revisit this issue some years from now due to a "what 
> you did in 2005 sucks since it doesn't show the actual time, and 
> that's what we need now."
>
>>> I happen to think that a one-second resolution is probably good 
>>> enough for most practical purposes.
>>
>> I worry about the "probably" and "most" in that sentence. :-) 
>> Moreover there's no need to have a time-based granularity at all.
>
> Words such as "probably" and "most" are the earmarks of handwaving; 
> and that statement was intended as a handwave (and an opinion at 
> that), not a statement of fact.  :-)
>
> I like the idea of using a CONDSTORE type mechanism, as long as it can 
> be implemented using a file date from the operating system (e.g. from 
> the stat() system call on UNIX).  Maybe not a "best" implementation 
> (in that there may be requirements for synchronization delays if the 
> value's resolution is too low) but nonetheless a viable and compliant 
> implementation.

As long as OS time doesn't go back, that should be possible.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j591CuaG020098; Wed, 8 Jun 2005 18:12:56 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j591CuIu020097; Wed, 8 Jun 2005 18:12:56 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from eagle.oceana.com (eagle.oceana.com [208.17.123.12]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j591CtRU020087 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 18:12:55 -0700 (PDT) (envelope-from ken@oceana.com)
Received: from [192.168.137.140] (69-161-65-27.bflony.adelphia.net [69.161.65.27]) (authenticated bits=0) by eagle.oceana.com (8.13.2/8.13.2) with ESMTP id j591ClNP005617 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 21:12:48 -0400 (EDT)
Message-ID: <42A7978A.9070605@oceana.com>
Date: Wed, 08 Jun 2005 21:12:42 -0400
From: Ken Murchison <ken@oceana.com>
Organization: Oceana Matrix Ltd.
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040514
X-Accept-Language: en,pdf
MIME-Version: 1.0
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com> <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0 (BAYES_00)
X-Scanned-By: MIMEDefang 2.39
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba wrote:

> 
>>> - section 6:  should we restrict selection options to be simple atoms
>>> as we have in this draft, or should we allow for them to be a list as
>>> well?   I'm thinking that some future selection option might take one
>>> or more arguments.  For example, the non-standard SCAN extension used
>>> by Pine and UW-IMAP could be reworked to look something like this:
>>>
>>> x LIST ((SCAN "foo")) "" "bar/%"
>>>
>>> or better yet:
>>>
>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
> 
> 
> No responses.......
> Since I think it adds too much complexity, unless there's some support 
> of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a glass 
> of Bordeaux in Paris as compensation.


It was just a thought, since I was thinking about Mark's SCAN extension 
at the time.  I won't be devastated if we leave things alone.

Unfortunately, I won't be in Paris, and the compensation wouldn't be 
necessary anyways.  :)

-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58NNGrw012465; Wed, 8 Jun 2005 16:23:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58NNGiD012464; Wed, 8 Jun 2005 16:23:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58NNCJZ012455 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 16:23:12 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58NMxtL004968 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 8 Jun 2005 16:22:59 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58NMxTA022133 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Wed, 8 Jun 2005 16:22:59 -0700
Date: Wed, 8 Jun 2005 16:20:57 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118258976.392523@localhost.localdomain>
Message-ID: <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com> <28904.1118258976.392523@localhost.localdomain>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 8 Jun 2005, Dave Cridland wrote:
> b) Nothing can show the actual time anyway, at least with the technology we 
> had back in 2005.

Is this really true?  Suppose the client wants to show a "human readable" 
last date/time?

A 32-bit ulong (and all IMAP 32-bit values *are* ulongs) since 2005-Jan-01 
00:00:00 UTC will work until February 2106.  The much talked-about 
doomsday at 2038-Jan-19 03:14:08 is because UNIX uses longs, not ulongs.

> If ever we need to know the system time of a particular update, we'll need to 
> do that in a distinct way anyway - system time can, under some circumstances, 
> go backwards. (Or it could in 2005, anyway.)

Leaving aside the question of drifting clocks, there's still the question 
of a "human readable" last date/time for user display.

>> I like the idea of using a CONDSTORE type mechanism, as long as it can be 
>> implemented using a file date from the operating system (e.g. from the 
>> stat() system call on UNIX).  Maybe not a "best" implementation (in that 
>> there may be requirements for synchronization delays if the value's 
>> resolution is too low) but nonetheless a viable and compliant 
>> implementation.
> Your earlier description of taking a timestamp, multiplying it to emulate an 
> increased resolution, and "making up" the extra resolution to fit the 
> specification, ought to work fine.

We also need some sort of stipulation that this value is not allowed to go 
backwards, and recommendations for preventing that from happening, such as 
detecting an old time-based value that is in the future, and monotonically 
increasing it on update rather than setting it to time.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jlq5C096261; Wed, 8 Jun 2005 12:47:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58JlqAg096260; Wed, 8 Jun 2005 12:47:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58JlpCf096251 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 12:47:51 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id C2ED9142271; Wed,  8 Jun 2005 12:46:00 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 06673-08; Wed, 8 Jun 2005 12:46:00 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 264E514226E; Wed,  8 Jun 2005 12:46:00 -0700 (PDT)
Message-ID: <42A74B60.9040102@osafoundation.org>
Date: Wed, 08 Jun 2005 12:47:44 -0700
From: Lisa Dusseault <lisa@osafoundation.org>
User-Agent: Mozilla Thunderbird 1.0 (Macintosh/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Cc: ietf-imapext@imc.org, Barry Leiba <leiba@watson.ibm.com>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com> <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
In-Reply-To: <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Yes, let's have a Friends of IMAP supper in Paris!  I've asked for a 
meeting slot but I don't know what day it will be on yet.

Lisa

Arnt Gulbrandsen wrote:

>
> Barry Leiba writes:
>
>> No responses.......
>> Since I think it adds too much complexity, unless there's some 
>> support of Ken's suggestion Very Soon, I will toss the idea, and buy 
>> Ken a glass of Bordeaux in Paris as compensation.
>
>
> I didn't like it, sounded like a bug magnet to me, but I decided to 
> say nothing in case someone came up with justification.
>
> How about not having an IMAP lunch, btw? How about a proper dinner at 
> a marginally expensive place instead? (You can buy Ken a glass of 
> vintage armagnac afterwards, that's sure to make him forgive you.)
>
> Arnt
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jbk8O095712; Wed, 8 Jun 2005 12:37:46 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58Jbkn2095711; Wed, 8 Jun 2005 12:37:46 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jbi1x095703 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 12:37:45 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id C5C58154009; Wed,  8 Jun 2005 19:37:38 +0000 (UTC)
Date: Wed, 08 Jun 2005 20:29:36 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
MIME-Version: 1.0
Message-Id: <28904.1118258976.392523@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <mrc@CAC.Washington.EDU>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 16:18:27 2005, Mark Crispin wrote:
> On Wed, 8 Jun 2005, Dave Cridland wrote:
>> The client only needs to know that a change has happened, never 
>> the actual times involved.
> 
> The question is if a client ever will care about the time.  I don't 
> want to have to revisit this issue some years from now due to a 
> "what you did in 2005 sucks since it doesn't show the actual time, 
> and that's what we need now."
> 
> 
I'd argue then that:
a) CONDSTORE had already prepackaged that can of worms for us, by 
using an essentially opaque 64-bit int.
b) Nothing can show the actual time anyway, at least with the 
technology we had back in 2005.

If ever we need to know the system time of a particular update, we'll 
need to do that in a distinct way anyway - system time can, under 
some circumstances, go backwards. (Or it could in 2005, anyway.)

> I like the idea of using a CONDSTORE type mechanism, as long as it 
> can be implemented using a file date from the operating system 
> (e.g. from the stat() system call on UNIX).  Maybe not a "best" 
> implementation (in that there may be requirements for 
> synchronization delays if the value's resolution is too low) but 
> nonetheless a viable and compliant implementation.

Your earlier description of taking a timestamp, multiplying it to 
emulate an increased resolution, and "making up" the extra resolution 
to fit the specification, ought to work fine.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Ig4CC091741; Wed, 8 Jun 2005 11:42:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58Ig4UK091740; Wed, 8 Jun 2005 11:42:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j58Ig3dY091734 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:42:04 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id DCEC81A754; Wed,  8 Jun 2005 20:42:01 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 5694F1A6E1; Wed,  8 Jun 2005 20:41:55 +0200 (CEST)
Message-Id: <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
Date: Wed, 8 Jun 2005 20:40:55 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: draft-ietf-imapext-list-extensions-13
Cc: Barry Leiba <leiba@watson.ibm.com>
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com> <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba writes:
> No responses.......
> Since I think it adds too much complexity, unless there's some support 
> of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a 
> glass of Bordeaux in Paris as compensation.

I didn't like it, sounded like a bug magnet to me, but I decided to say 
nothing in case someone came up with justification.

How about not having an IMAP lunch, btw? How about a proper dinner at a 
marginally expensive place instead? (You can buy Ken a glass of vintage 
armagnac afterwards, that's sure to make him forgive you.)

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58IR9QZ090492; Wed, 8 Jun 2005 11:27:09 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58IR9wZ090491; Wed, 8 Jun 2005 11:27:09 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58IR8ud090483 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:27:08 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j58ISMLV005282 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:28:22 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j58IR2k123308 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:27:02 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j58IR1d109726 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:27:01 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118255221.284; Wed, 08 Jun 2005 14:27:01 -0400
Message-ID: <42A73875.3040501@watson.ibm.com>
Date: Wed, 08 Jun 2005 14:27:01 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

>>- section 6:  should we restrict selection options to be simple atoms
>>as we have in this draft, or should we allow for them to be a list as
>>well?   I'm thinking that some future selection option might take one
>>or more arguments.  For example, the non-standard SCAN extension used
>>by Pine and UW-IMAP could be reworked to look something like this:
>>
>>x LIST ((SCAN "foo")) "" "bar/%"
>>
>>or better yet:
>>
>>x LIST ((SCAN "utf-8" "foo")) "" "bar/%"

No responses.......
Since I think it adds too much complexity, unless there's some support 
of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a glass 
of Bordeaux in Paris as compensation.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58INZR6090317; Wed, 8 Jun 2005 11:23:35 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58INZI1090316; Wed, 8 Jun 2005 11:23:35 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58INYnh090305 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:23:34 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j58IOlmT004737 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:24:48 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j58INSk101542 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:23:28 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j58INRd102098 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:23:27 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118255007.283; Wed, 08 Jun 2005 14:23:27 -0400
Message-ID: <42A7379F.2080007@watson.ibm.com>
Date: Wed, 08 Jun 2005 14:23:27 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com> <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To: <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Folks, really... this has strayed widely from what I think Simon's 
original issue was.  If I have this right, Simon was talking about 
synchronizing client cached state when coming back online.  What Timo 
has been arguing about is, as Mark noted, notification.  These are 
vastly different issues, and people have already proposed models for 
multi-mailbox notification schemes.

If we're concerned about notification, we should work on a proposal that 
fits that.  Then a client can have one socket open to the server to be 
notified on, and can provide the server with a list of mailboxes to 
monitor.  Sorta like a multi-mailbox IDLE.  This is waaaay better than 
polling, regardless of how you poll.

If that doesn't work because you can't keep a socket open, then look at 
the Lemonade work, and at Stephane Maes's "push IMAP" spec -- which has 
a lot of complexity, but the widely useful parts of it are being plucked 
out separately.

Now, back to the original question:  Can Simon (and others) resynch his 
client cache effectively without doing STATUS on every mailbox?  Or 
should we have an extension that enables STATUS options on an extended 
LIST command, to make his work more efficient on the servers that 
support it?

Note that however that last question is answered, I do NOT think this 
should affect the LIST extensions spec we're working on now.  It would, 
if anything, be another extension, to be supported by servers or not, 
independently of whether they support the base LIST extensions.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58FIv5N070862; Wed, 8 Jun 2005 08:18:57 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58FIvoE070861; Wed, 8 Jun 2005 08:18:57 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58FIrj3070853 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 08:18:53 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58FIhea015050 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 8 Jun 2005 08:18:43 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58FISWI018977 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Wed, 8 Jun 2005 08:18:42 -0700
Date: Wed, 8 Jun 2005 08:18:27 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118220817.700674@localhost.localdomain>
Message-ID: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 8 Jun 2005, Dave Cridland wrote:
> The client only needs to know that a change has happened, never the actual 
> times involved.

The question is if a client ever will care about the time.  I don't want 
to have to revisit this issue some years from now due to a "what you did 
in 2005 sucks since it doesn't show the actual time, and that's what we 
need now."

>> I happen to think that a one-second resolution is probably good enough for 
>> most practical purposes.
> I worry about the "probably" and "most" in that sentence. :-) Moreover 
> there's no need to have a time-based granularity at all.

Words such as "probably" and "most" are the earmarks of handwaving; and 
that statement was intended as a handwave (and an opinion at that), not a 
statement of fact.  :-)

I like the idea of using a CONDSTORE type mechanism, as long as it can be 
implemented using a file date from the operating system (e.g. from the 
stat() system call on UNIX).  Maybe not a "best" implementation (in that 
there may be requirements for synchronization delays if the value's 
resolution is too low) but nonetheless a viable and compliant 
implementation.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58EdJTu066910; Wed, 8 Jun 2005 07:39:19 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58EdJej066909; Wed, 8 Jun 2005 07:39:19 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58EdIAA066900 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 07:39:19 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 72D30154009; Wed,  8 Jun 2005 14:39:12 +0000 (UTC)
Date: Wed, 08 Jun 2005 15:31:17 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <1118233852.29860.17.camel@localhost.localdomain>  <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
In-Reply-To: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
MIME-Version: 1.0
Message-Id: <28904.1118241077.035520@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, <ietf-imapext@imc.org>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 13:38:58 2005, Arnt Gulbrandsen wrote:
> Timo Sirainen writes:
>> If we assume that timestamp can be increased after each 
>> modification, can't we just assume that timestamp is one 
>> implementation of CONDSTORE's MODSEQ? In that case we don't need a 
>> whole new timestamp thing, just a possibility to get latest MODSEQ 
>> from LIST reply.
> 
> Now there's a good idea. How much baggage does it drag along - how 
> much of CONDSTORE has to be provided to provide MODSEQ?

Well, there's two things here.

Firstly, even if CONDSTORE is implemented, not every mailbox has a 
HIGHESTMODSEQ.

Secondly, although any increase in HIGHESTMODSEQ must - assuming 
we're using Alexey's STATUS-inna-LIST concept - also increase the 
LIST's magic sequence, HIGHESTMODSEQ is per-mailbox, while the 
LISTS's sequence is per-logical-server (ie, per server or per 
cluster, whichever applies). This means that when we wrap the LIST 
sequence, we have to wrap all the mailboxes' sequences as well.

The second issue is pretty minor - I'd fully expect any server 
implementing both CONDSTORE and CHANGEDSINCE LIST to use the same 
sequence number for both purposes, however, it's not essential. The 
first is more crucial - it seems fairly logical, to me, that certain 
types of mailbox can be made to support CONDSTORE easier than others, 
so an IMAP implementation with multiple backends (which, IIRC, 
includes both Mark and Timo's implementations) may well only support 
CONDSTORE on some.

So while I'd expect that in many cases, servers implementing 
CONDSTORE would havea  single sequence, I wouldn't want to mandate 
that.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CeCQO045166; Wed, 8 Jun 2005 05:40:12 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58CeCdh045165; Wed, 8 Jun 2005 05:40:12 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j58CeApY045150 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 05:40:11 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id B93F01A750; Wed,  8 Jun 2005 14:40:09 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 987271A710; Wed,  8 Jun 2005 14:39:56 +0200 (CEST)
Message-Id: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
Date: Wed, 8 Jun 2005 14:38:58 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, Dave Cridland <dave@cridland.net>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi> <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain> <1118233852.29860.17.camel@localhost.localdomain>
In-Reply-To: <1118233852.29860.17.camel@localhost.localdomain>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Timo Sirainen writes:
> If we assume that timestamp can be increased after each modification, 
> can't we just assume that timestamp is one implementation of 
> CONDSTORE's MODSEQ? In that case we don't need a whole new timestamp 
> thing, just a possibility to get latest MODSEQ from LIST reply.

Now there's a good idea. How much baggage does it drag along - how much 
of CONDSTORE has to be provided to provide MODSEQ?

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CV1aM041998; Wed, 8 Jun 2005 05:31:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58CV1Bs041997; Wed, 8 Jun 2005 05:31:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from dmgw.movial.fi (dmgw.movial.fi [62.236.91.5]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CV0Mq041952 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 05:31:00 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from movial.fi (hellmail.movial.fi [62.236.91.6]) by dmgw.movial.fi (Postfix) with ESMTP id D1A53EC12B; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Received: from localhost (hellmail.movial.fi [127.0.0.1]) by movial.fi (Postfix) with ESMTP id AB2538005D; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Received: from movial.fi ([127.0.0.1]) by localhost (hellmail [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 06628-05; Wed, 8 Jun 2005 15:30:53 +0300 (EEST)
Received: from sd022.hel.movial.fi (sd022.hel.movial.fi [172.17.49.22]) (using SSLv3 with cipher RC4-MD5 (128/128 bits)) (Client did not present a certificate) by movial.fi (Postfix) with ESMTP id 79F2780052; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Dave Cridland <dave@cridland.net>
Cc: Mark Crispin <MRC@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
In-Reply-To: <28904.1118220817.700674@localhost.localdomain>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi> <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain>
Content-Type: text/plain
Date: Wed, 08 Jun 2005 12:30:52 +0000
Message-Id: <1118233852.29860.17.camel@localhost.localdomain>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: by amavisd-new-20030616-p10 (Debian) at movial.fi
X-Amavis-Alert: BAD HEADER Improper use of control character (char 0D hex) in message header 'References' References: ...46801216@PD.gordano.com>\n\t \r <Pine.WNT.4.6... ^
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 2005-06-08 at 09:53 +0100, Dave Cridland wrote:
> > On Tue, 7 Jun 2005, Timo Sirainen wrote:
> >>> We haven't defined what the timestamp interval should be.
> >> Isn't it going to be seconds if you want it to be supported by 
> >> legacy stores?
> > 
> > Not necessarily.
> > 
> > A store which only resolves to a second can comply very simply.  It 
> > just multiplies the second-resolution values by whatever the 
> > necessary scale factor is, and guarantees that new message(s) 
> > delivery always updates the value.
> > 
> Exactly - that's essentially what an ACAP modtime is. Or you could 
> use a system similar to CONDSTORE's MODSEQ - an always-increasing 
> 64-bit integer. The client only needs to know that a change has 
> happened, never the actual times involved.

If we assume that timestamp can be increased after each modification,
can't we just assume that timestamp is one implementation of CONDSTORE's
MODSEQ? In that case we don't need a whole new timestamp thing, just a
possibility to get latest MODSEQ from LIST reply.




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5891gqT069179; Wed, 8 Jun 2005 02:01:42 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5891gTP069178; Wed, 8 Jun 2005 02:01:42 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5891Zrb069103 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 02:01:36 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 43E9A154009; Wed,  8 Jun 2005 09:01:25 +0000 (UTC)
Date: Wed, 08 Jun 2005 09:53:37 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To:  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
MIME-Version: 1.0
Message-Id: <28904.1118220817.700674@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <MRC@cac.washington.edu>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon Jun  6 23:17:09 2005, Mark Crispin wrote:
> 
> On Tue, 7 Jun 2005, Timo Sirainen wrote:
>>> We haven't defined what the timestamp interval should be.
>> Isn't it going to be seconds if you want it to be supported by 
>> legacy stores?
> 
> Not necessarily.
> 
> A store which only resolves to a second can comply very simply.  It 
> just multiplies the second-resolution values by whatever the 
> necessary scale factor is, and guarantees that new message(s) 
> delivery always updates the value.
> 
> 
Exactly - that's essentially what an ACAP modtime is. Or you could 
use a system similar to CONDSTORE's MODSEQ - an always-increasing 
64-bit integer. The client only needs to know that a change has 
happened, never the actual times involved.

I do like ACAP's modtimes, because they lend a human-readability to 
it, but the human semantics aren't required for the client at all. 
(Indeed, it makes no sense to try - the modtimes my ACAP 
implementation uses are only "usually" accurate to within a second, 
and may be wildly out to satisfy the always-increasing requirement, 
which is more important.)

> I happen to think that a one-second resolution is probably good 
> enough for most practical purposes.

I worry about the "probably" and "most" in that sentence. :-) 
Moreover there's no need to have a time-based granularity at all.

>> 1) LIST -> notice new stamp
>> 2) SELECT mailbox
>> 3) notice the new mail, read it, etc.
>> 4) another mail arrives
>> 5) SELECT another-mailbox
>> The mail is lost unless server notifies client of it before the 
>> SELECT command. I don't think many (any?) servers do that 
>> currently.

I'm not sure the mail is missed, because the timestamp (or sequence 
value) would increase, and if the client never receives a 
notification of (4), it will assume that the mailbox is in the state 
as of the end of (3). The only circumstance you might have problems 
with is if you were foolish enough to use real timestamps, and (4) 
and (1) (or (2), if the timestamp is reported there, too) happened 
within the granularity - but as I say, that would be foolish design 
indeed.

I think there's a huge scope for improvement there. Alexey and I have 
been idly talking about this area in private mail a while ago, so the 
more sensible ideas here are probably his.

If we bring in Alexey's STATUS-in-LIST concept, we can issue 
something like (and I'm stretching LIST-EXTENDED syntax here by quite 
a way):

A01 LIST (SUBSCRIBED (CHANGEDSINCE 20050607202312000)) "" "*" 
(STATUS=UIDVALIDITY,UIDNEXT,MESSAGES)

This could tell us about only the metadata that's changed for a 
mailbox - if any has. This satisfies Simon Tyler's (quite reasonable, 
I thought) request, and should lower the cost for Mark, too - since 
we can assume that, at the very least, a client is likely to issue 
STATUS (if not SELECT) to get the extra information it requires, 
better to allow it to ask for everything in one go. Because the 
"something changed" metadata and the data itself is contained within 
the same message, Timo's concerns about races should vanish.

We can further improve this by adding unsolicited updates, perhaps 
with a syntax like:

A01 LIST (SUBSCRIBED (CHANGEDSINCE 20050607202312000)) "" "*" RETURN 
(STATUS=UIDVALIDITY,UIDNEXT,MESSAGES UPDATES)

This would send updated LIST responses for mailboxes which changed. 
This might merely replace the client's poll with a server-controlled 
poll (thus lowering what Mark describes as abusive behaviour), or 
might be implemented using notifications within the server, avoiding 
polling entirely.

This helps Timo's concerns too - it seems reasonable for a server to 
buffer update LIST responses for a short period to avoid generates 
large numbers of redundant responses.

I appreciate there's a lot more thought that needs to go into this, 
and I'm not wed to any particular detail, so feel free to pull it 
apart.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j57CfoWJ082092; Tue, 7 Jun 2005 05:41:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j57CfofK082091; Tue, 7 Jun 2005 05:41:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j57CfnuN082078 for <ietf-imapext@imc.org>; Tue, 7 Jun 2005 05:41:49 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Tue, 07 Jun 2005 08:41:48 -0400 id 0059811E.42A5960C.00000FCE
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: ietf-imapext@imc.org
Subject: FW: Last Call Expired: <draft-ietf-imapext-2086upd> 
Date: Tue, 7 Jun 2005 08:42:18 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Thread-Index: AcVrEZSAHYbM9ZjhSEGNLtEw7H/cGwATC4ww
Message-ID: <courier.42A5960C.00000FCE@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

The IETF last call for draft-ietf-imapext-2086upd has completed.  I didn't
see any new comments.  Please update the document to correct the remaining
editorial issues (including IDNits) identified before the last call and I'll
schedule it for IESG evaluation once the update is available.  The deadline
to make the next telechat is 16 June.

-Scott-

-----Original Message-----
From: DraftTracker Mail System [mailto:iesg-secretary@ietf.org] 
Sent: Monday, June 06, 2005 11:30 PM
To: iesg@ietf.org
Cc: iesg-secretary@ietf.org
Subject: Last Call Expired: <draft-ietf-imapext-2086upd> 


Please DO NOT reply on this email.
 

ID: draft-ietf-imapext-2086upd

IETF Last Call has ended, and the state has been changed to Waiting for AD
Go-Ahead



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56MJHkG078040; Mon, 6 Jun 2005 15:19:17 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56MJH9Y078039; Mon, 6 Jun 2005 15:19:17 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56MJGrr078032 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 15:19:16 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56MJBKM002392 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 15:19:11 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56MJBv1025659 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 15:19:11 -0700
Date: Mon, 6 Jun 2005 15:17:09 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Message-ID: <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Tue, 7 Jun 2005, Timo Sirainen wrote:
>> We haven't defined what the timestamp interval should be.
> Isn't it going to be seconds if you want it to be supported by legacy stores?

Not necessarily.

A store which only resolves to a second can comply very simply.  It just 
multiplies the second-resolution values by whatever the necessary scale 
factor is, and guarantees that new message(s) delivery always updates the 
value.

I happen to think that a one-second resolution is probably good enough for 
most practical purposes.  I wouldn't be so sanguine if we didn't have the 
atomic multiple-message append mechanism through MULTIAPPEND. 
Fortunately, we do; and it is reasonable to assume that a server that is 
advanced enough to have this new LIST capability can also be advanced 
enough to deliver a flood of messages atomically (and thus require only a 
single interval tick).

> 1) LIST -> notice new stamp
> 2) SELECT mailbox
> 3) notice the new mail, read it, etc.
> 4) another mail arrives
> 5) SELECT another-mailbox
> The mail is lost unless server notifies client of it before the SELECT 
> command. I don't think many (any?) servers do that currently.

You forgot a step: the client obtains the mailbox's current timestamp.

If that happens before (4), then the mailbox will have a different stamp. 
If it happens after (4), an announcement of the new message will as part 
of the operation.

No race.

> Although rare, all that above is possible as a user action.

It's also a terrible abuse of IMAP; it treats IMAP as a stateless protocol 
such as HTTP and completely disregards the entire purpose of IMAP as a 
stateful protocol.

I don't think that we should optimize for abusive use that the protocol 
was never designed to support.

> I sometimes click 
> a mailbox only to notice I mis-clicked and then select another box 
> immediately. I think it's even possible to have time to read the subject, 
> delete the message and switch to another mailbox all within one second.

If you care about monitoring the mailbox, why don't you leave it open?

> 1) select mailbox, do things with it, notice that we need to change the 
> mailbox
> 2) new mail arrives
> 2) issue LIST to get the current timestamp
> 3) new mail arrives
> 4) SELECT another mailbox

In this case, the mailbox's timestamp from (2) will be different (less) 
than the timestamp from (3).

No race.

> But that depends completely on how fast the hardware and software 
> implementation is, and what commands are being used. I think it's still 
> easily possible to handle tens of IMAP commands within one second.

I've looked at lots of IMAP and SMTP/MTA session logs (and not just UW 
imapd).  I think that you're making a stretch.

Even if your case was valid, it is solvable; unlike the alternative which 
is harmful in some server environments *and* breaks a promise made by the 
base specification.

> I don't think that is a "simply". It requires either modifying the existing 
> delivery agent (which I think is not very likely to happen) or running a 
> second delivery agent to fix the timestamp.

Whoa!

If you don't have a whiz-bang format capable of delivering that metadata 
quickly, you aren't going to be able to respond in the times that you 
argue.  And if you have a whiz-bang format then you already have a custom 
mail delivery agent.

If the mail delivery agent isn't part of the IMAP server software package, 
then you haven't a snowball's chance in hell of maintaining a UIDNEXT that 
can be delivered in under a second.

If /bin/mail has appended 500MB of new traditional UNIX format data since 
the last time the IMAP server had the mailbox open, it simply is not going 
to calculate a current UIDNEXT in under a second.

> What probably would happen is that admins will happily enable the timestamp 
> feature in IMAP server without touching the delivery agent. Which makes it 
> unrealiable.

Not at all.

In such circumstances, the IMAP server can also prevent the problem by 
guaranteeing that it never returns a timestamp that equals the current 
time.

> UIDNEXT exists already and can be used for this purpose. If we can't use the 
> timestamps in a portable and useful enough way, I don't think there's a need 
> to invent a new metadata field.

UIDNEXT is potentially expensive and the base specification guarantees 
this fact to a server implementation.  An extension can not repeal a 
guarantee of the base specification.

> I suppose with enough rules (adding arbitrary delays) and requirements (which 
> can be ignored by admins) you could make this reliable enough, but I don't 
> really see them as a good idea.

It's a damned good idea to do something which can be implemented cheaply 
by just about everybody concerned (either at same-cost or much lower cost 
than UIDNEXT), and will not have to do any delays for timing race 
prevention except in extraordinary cases.

> Of course, one possible solution would be to just ignore the races and let 
> IMAP clients deal with it:
> 1) when timestamp changes, check if there is new mail
> 2) one second later check for new mail again
> 3) also check if timestamp changed, if so goto 2

Step (2) accomplishes nothing, because the check for new mail is already 
of newer value than the timestamp.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LepVn074278; Mon, 6 Jun 2005 14:40:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56Lepwx074277; Mon, 6 Jun 2005 14:40:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LeoYG074266 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:40:50 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from [192.168.10.217] (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 1FC6E1C351D2; Tue,  7 Jun 2005 00:40:44 +0300 (EEST)
In-Reply-To: <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary="Apple-Mail-30-200404062"
Message-Id: <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Content-Transfer-Encoding: 7bit
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
From: Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
Date: Tue, 7 Jun 2005 00:40:41 +0300
To: Mark Crispin <MRC@cac.washington.edu>
X-Pgp-Agent: GPGMail 1.0.2
X-Mailer: Apple Mail (2.622)
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--Apple-Mail-30-200404062
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII; format=flowed

On 6.6.2005, at 23:29, Mark Crispin wrote:

>> So, what needs to happen within one
>> second is:
>
> s/second/timestamp interval
>
> We haven't defined what the timestamp interval should be.

Isn't it going to be seconds if you want it to be supported by legacy 
stores?

> You have not demonstrated any benefit that outweighs the substantial 
> cost.

How about these two cases:

1) LIST -> notice new stamp
2) SELECT mailbox
3) notice the new mail, read it, etc.
4) another mail arrives
5) SELECT another-mailbox

The mail is lost unless server notifies client of it before the SELECT 
command. I don't think many (any?) servers do that currently.

Although rare, all that above is possible as a user action. I sometimes 
click a mailbox only to notice I mis-clicked and then select another 
box immediately. I think it's even possible to have time to read the 
subject, delete the message and switch to another mailbox all within 
one second.

And that's only with "normal" client user interface. Some client could 
merge multiple mailboxes within same list and you could easily be 
selecting multiple mailboxes within one second.

Another possibility:

1) select mailbox, do things with it, notice that we need to change the 
mailbox
2) new mail arrives
2) issue LIST to get the current timestamp
3) new mail arrives
4) SELECT another mailbox

Or should it also be required not to issue LIST for mailbox without 
doing a NOOP one timestamp interval later, before closing the mailbox?

>>> But, assuming that we actually have to fix this (perhaps with a 
>>> massive
>>> speedup of RTTs),
>> I think even with your example it could easily happen nowadays. It's
>> maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
>> pings to my ISP's mail server. That's 125 RTTs within one second.
>
> There's a substantial difference between ping RTTs and RTTs in a 
> higher level protocol such as IMAP.

But that depends completely on how fast the hardware and software 
implementation is, and what commands are being used. I think it's still 
easily possible to handle tens of IMAP commands within one second. With 
fast enough servers the network latency is the only measureable one and 
ping reflects that pretty well.

>> I
>> don't see how you could actually get it to work with mboxes (and other
>> legacy stores) as the mails are usually written using tons of 
>> different
>> programs that aren't going to be increasing the timestamp.
>
> Not at all.  The mail delivery agent is simply required to ensure that 
> the timestamp changes, and if necessary delays delivery so that the 
> new message(s) get an updated timestamp.

I don't think that is a "simply". It requires either modifying the 
existing delivery agent (which I think is not very likely to happen) or 
running a second delivery agent to fix the timestamp. And the secondary 
timestamp-agent would also have to lock the mailbox if you want it free 
of race conditions. Yet another software needing to know how mboxes are 
being locked..

What probably would happen is that admins will happily enable the 
timestamp feature in IMAP server without touching the delivery agent. 
Which makes it unrealiable.

>> And I wouldn't want to be polling timestamps either. I think server
>> should be notifying IDLEing clients of changes in them automatically.
>
> Then why are you talking about using UIDNEXT at all?

UIDNEXT exists already and can be used for this purpose. If we can't 
use the timestamps in a portable and useful enough way, I don't think 
there's a need to invent a new metadata field.

I suppose with enough rules (adding arbitrary delays) and requirements 
(which can be ignored by admins) you could make this reliable enough, 
but I don't really see them as a good idea.

Of course, one possible solution would be to just ignore the races and 
let IMAP clients deal with it:

1) when timestamp changes, check if there is new mail
2) one second later check for new mail again
3) also check if timestamp changed, if so goto 2

Actually I'd prefer this if you really must implement non-accurate 
timestamps.

--Apple-Mail-30-200404062
content-type: application/pgp-signature; x-mac-type=70674453;
	name=PGP.sig
content-description: This is a digitally signed message part
content-disposition: inline; filename=PGP.sig
content-transfer-encoding: 7bit

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (Darwin)

iD8DBQFCpMLZyUhSUUBViskRAiGaAJ9KMj0SPPRYIzvxcePe6oQjo5/TewCeMB5z
bgQ/Fw29q9ccCGrbVYYhAKY=
=Nw9o
-----END PGP SIGNATURE-----

--Apple-Mail-30-200404062--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LKXTB072677; Mon, 6 Jun 2005 14:20:33 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56LKX46072676; Mon, 6 Jun 2005 14:20:33 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LKWQi072670 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:20:32 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56LKKAB020378 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 14:20:20 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56LKJJt023853 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 14:20:20 -0700
Date: Mon, 6 Jun 2005 14:18:17 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
cc: ietf-imapext@imc.org, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Message-ID: <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Arnt Gulbrandsen wrote:
>> That's quite a lot to have happen in one interval.
> Really? Consider:

Yes, really.

All of this has to happen in one interval (remember, we have not yet 
defined the magnitude of the interval), *AND* it is necessary to 
accomplish not just two message deliveries but also SELECT/FETCH/close.

On top of that, there is a trivial way to close the race.

This discussion is increasingly sounding like the objections to MIME 
boundary delimiters 15 years ago.  Supposedly, it was impossible to come 
up with a safe algorithm to generate a boundary delimiter that would be 
guaranteed not to occur in the content without having to examine the 
content first.

As it turned out, it wasn't a problem.  Not only is it easy to generate 
boundary delimiters that are impossible in quoted-printable and base64 
content, but it turns out that as long as the delimiter is unlikely enough 
it isn't going to occur in unencoded content for all practical purposes.

An update timestamp works, and lacks the harm caused by breaking a promise 
of the base specification (which, I contend, is out of charter for the 
base specification).

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56L2a2G071181; Mon, 6 Jun 2005 14:02:36 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56L2a3X071180; Mon, 6 Jun 2005 14:02:36 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j56L2VLL071172 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:02:32 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id EC7001A718; Mon,  6 Jun 2005 23:02:29 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 66DC71A70F; Mon,  6 Jun 2005 23:02:19 +0200 (CEST)
Message-Id: <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Date: Mon, 6 Jun 2005 23:01:36 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
In-Reply-To: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin writes:
> In any case, the timing race that you describe is unlikely.  In order 
> to have a false negative, the following all needs to happen within 
> the same interval of time:
>  1) server gets a new message
>  2) client polls and sees that there is a new message
>  3) client select mailbox
>  4) client processes new message
>  5) client closes mailbox
>  6) server gets a second new message
>
> That's quite a lot to have happen in one interval.

Really? Consider:

1) It's just after lunch, and the entire company is sending mail like crazy.
2) An IMAP Client is told that there's mail in mailbox 1.
3) Client selects that mailbox and retrieves the message
4) Message is displayed
5) Client is told there's mail in mailbox 2.
6) Client selects mailbox 2 and fetches the new message
7) Server stores a new message in mailbox 1

If the client uses e.g. SIEVE to store mail in two mailboxes, and two 
threads are active, I don't see this as uncommon. In a thousand-people 
company I could easily see this happening every day. Normally there's a 
step 8

8) Server tells client there's another new message in mailbox 1

which forces the client to discover the missed message, but if the 
message happened to be the last in the thread, that doesn't work.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KVT64067751; Mon, 6 Jun 2005 13:31:29 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56KVTn8067750; Mon, 6 Jun 2005 13:31:29 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KVTD1067743 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:31:29 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56KVIS2012398 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 13:31:18 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56KVIWJ001971 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 13:31:18 -0700
Date: Mon, 6 Jun 2005 13:29:15 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1118089030.15166.502.camel@hurina>
Message-ID: <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Timo Sirainen wrote:
> No, you're assuming that the client is actually interested in reading
> the new message. I think more commonly the timestamp change would just
> cause a STATUS query to request the new number of messages in the
> mailbox (updating a counter in screen - that's why so many clients do
> the STATUS now for all mailboxes).

So, you're saying that, given the very unlikely scenario that you 
describe, it is somehow a problem that, instead of saying "2 new 
messages", it only says "1 new message"?

The fact that there is at least one new message is still shown.

> So, what needs to happen within one
> second is:

s/second/timestamp interval

We haven't defined what the timestamp interval should be.

> 1) new mail is written
> 2) client issues LIST
> 3) client sees new timestamp, issues STATUS command
> 4) server sends reply to STATUS
> 5) another mail is written
> So it really requires just 1,5 RTTs for the problem to occur. The actual
> LIST and STATUS processing and replying is also near instantaneous
> usually.

Even if that was all true (and it isn't -- there is no reason to believe 
that all this will happen in a timestamp interval and you would have to go 
to considerable effect even to contrive it to happen), the only "bad 
effect" is that it says "1 new message" instead of "2 new messages".

It still shows that there are new messages.

Also, there is no way of knowing anything about the message content until 
the mailbox is actually examined.  Perhaps both are spam.

You have not demonstrated any benefit that outweighs the substantial cost.

>> But, assuming that we actually have to fix this (perhaps with a massive
>> speedup of RTTs),
> I think even with your example it could easily happen nowadays. It's
> maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
> pings to my ISP's mail server. That's 125 RTTs within one second.

There's a substantial difference between ping RTTs and RTTs in a higher 
level protocol such as IMAP.

>> we can require that any delivery of new messages write
>> an increased timestamp.
> That would be one implementation-specific way to handle the issue (the
> "timestamp" could be just some 32/64bit increasing number), although I
> don't see how you could actually get it to work with mboxes (and other
> legacy stores) as the mails are usually written using tons of different
> programs that aren't going to be increasing the timestamp.


Not at all.  The mail delivery agent is simply required to ensure that the 
timestamp changes, and if necessary delays delivery so that the new 
message(s) get an updated timestamp.

> Timestamps are also metadata. They just happen to be somewhat more
> easily accessible to you.

s/you/everyone

Even databases have faster access to the operating system's file timestamp 
than they do to a datum contained within the database.

> And I wouldn't want to be polling timestamps either. I think server
> should be notifying IDLEing clients of changes in them automatically.

Then why are you talking about using UIDNEXT at all?

You should be talking about notifications.

The entire discussion at hand has to do with determining which mailboxes 
require synchronization.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KHLfZ066784; Mon, 6 Jun 2005 13:17:21 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56KHLwq066783; Mon, 6 Jun 2005 13:17:21 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KHGvE066770 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:17:17 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 68EFF1C351D2; Mon,  6 Jun 2005 23:17:10 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Mark Crispin <MRC@CAC.Washington.EDU>
Cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-hOSZDezp+WU7UNs3iVAh"
Date: Mon, 06 Jun 2005 23:17:10 +0300
Message-Id: <1118089030.15166.502.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-hOSZDezp+WU7UNs3iVAh
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Mon, 2005-06-06 at 12:38 -0700, Mark Crispin wrote:
> On Mon, 6 Jun 2005, Timo Sirainen wrote:
> > Unless you're going to make the "last updated" timestamp accurate up to
> > microseconds, there's a chance that the mailbox has had new messages
> > since your last check if another message was added just after you
> > checked it. Are you going to require that?
>=20
> RTTs in most of today's networking are measured in milliseconds, not=20
> microseconds.  I doubt very much that we will need millisecond level=20
> resolution.

I was going to say milliseconds first, but then thought maybe it's not
enough in future, and besides I think most OSes/filesystems support
either second or microsecond accuracy, but not millisecond.

> In any case, the timing race that you describe is unlikely.  In order to=20
> have a false negative, the following all needs to happen within the same=20
> interval of time:
>   1) server gets a new message
>   2) client polls and sees that there is a new message
>   3) client select mailbox
>   4) client processes new message
>   5) client closes mailbox
>   6) server gets a second new message
>=20
> That's quite a lot to have happen in one interval.

No, you're assuming that the client is actually interested in reading
the new message. I think more commonly the timestamp change would just
cause a STATUS query to request the new number of messages in the
mailbox (updating a counter in screen - that's why so many clients do
the STATUS now for all mailboxes). So, what needs to happen within one
second is:

1) new mail is written
2) client issues LIST
3) client sees new timestamp, issues STATUS command
4) server sends reply to STATUS
5) another mail is written

So it really requires just 1,5 RTTs for the problem to occur. The actual
LIST and STATUS processing and replying is also near instantaneous
usually.

> But, assuming that we actually have to fix this (perhaps with a massive=20
> speedup of RTTs),=20

I think even with your example it could easily happen nowadays. It's
maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
pings to my ISP's mail server. That's 125 RTTs within one second.

> we can require that any delivery of new messages write=20
> an increased timestamp. =20

That would be one implementation-specific way to handle the issue (the
"timestamp" could be just some 32/64bit increasing number), although I
don't see how you could actually get it to work with mboxes (and other
legacy stores) as the mails are usually written using tons of different
programs that aren't going to be increasing the timestamp.

> > I think the most important thing people want is to see when new mail
> > appears in different mailboxes. They want to see it as fast as possible
> > and without the notifications getting randomly lost. Getting UIDNEXT
> > value is good enough for that.
>=20
> Polling metadata is absolutely the wrong way to accomplish this.  This is=
,=20
> in effect, what getting the UIDNEXT is.

Timestamps are also metadata. They just happen to be somewhat more
easily accessible to you.

And I wouldn't want to be polling timestamps either. I think server
should be notifying IDLEing clients of changes in them automatically.

--=-hOSZDezp+WU7UNs3iVAh
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCpK9FyUhSUUBViskRAjkwAKCPYdifuRm0A1OdBWH9rwkLpUazbQCeOrPV
LTRAf99KQ3RqC3maQUhjwJI=
=TOA/
-----END PGP SIGNATURE-----

--=-hOSZDezp+WU7UNs3iVAh--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56Jcc8F062635; Mon, 6 Jun 2005 12:38:38 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56JccYc062634; Mon, 6 Jun 2005 12:38:38 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56JcbLl062628 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 12:38:37 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56JcTKx027145 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 12:38:29 -0700
Received: from Shimo-Tomobiki.panda.com ([65.122.177.186]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56JcPGp023310 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 12:38:28 -0700
Date: Mon, 6 Jun 2005 12:38:21 -0700
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1118084910.15166.468.camel@hurina>
Message-ID: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Timo Sirainen wrote:
> Unless you're going to make the "last updated" timestamp accurate up to
> microseconds, there's a chance that the mailbox has had new messages
> since your last check if another message was added just after you
> checked it. Are you going to require that?

RTTs in most of today's networking are measured in milliseconds, not 
microseconds.  I doubt very much that we will need millisecond level 
resolution.

In any case, the timing race that you describe is unlikely.  In order to 
have a false negative, the following all needs to happen within the same 
interval of time:
  1) server gets a new message
  2) client polls and sees that there is a new message
  3) client select mailbox
  4) client processes new message
  5) client closes mailbox
  6) server gets a second new message

That's quite a lot to have happen in one interval.

But, assuming that we actually have to fix this (perhaps with a massive 
speedup of RTTs), we can require that any delivery of new messages write 
an increased timestamp.  Although that would seem to limit the rate of new 
message delivery, that is ameliorated by MULTIAPPEND; it isn't necessary 
to require a new timestamp per message, as long as multiple message 
delivery is atomic.

Alternatively, we can require that a SELECT/FETCH/close cycle occupy more 
than one interval.

> I think the most important thing people want is to see when new mail
> appears in different mailboxes. They want to see it as fast as possible
> and without the notifications getting randomly lost. Getting UIDNEXT
> value is good enough for that.

Polling metadata is absolutely the wrong way to accomplish this.  This is, 
in effect, what getting the UIDNEXT is.

This may be my personal problem, but in my opinion people who can't wait a 
minute to be told of newly-arrived spam are badly in need of a life...

> Other indicators would be fine as well,
> as long as they are reliable.

Yes.  Reliable notification.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56J8dpn060388; Mon, 6 Jun 2005 12:08:39 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56J8dLN060387; Mon, 6 Jun 2005 12:08:39 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56J8bcn060377 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 12:08:38 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id F03981C351D2; Mon,  6 Jun 2005 22:08:30 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Mark Crispin <MRC@cac.washington.edu>
Cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-b+b3Crv/3bMDDok3XqXR"
Date: Mon, 06 Jun 2005 22:08:30 +0300
Message-Id: <1118084910.15166.468.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-b+b3Crv/3bMDDok3XqXR
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Fri, 2005-06-03 at 15:11 -0700, Mark Crispin wrote:
> On Fri, 3 Jun 2005, Simon Tyler wrote:
> > It does not detract from the fact that a client needs UIDVALIDITY on al=
l=20
> > synchronised mailboxes which is why I wanted it added to the LIST=20
> > response.
>=20
> Isn't it really the case that you only need UIDVALIDITY/UIDNEXT on=20
> mailboxes which have a changed "last updated" date/time from the last=20
> time?  Do you really need to be told UIDVALIDITY/UIDNEXT on mailboxes=20
> which are unchanged?  If so, why?

Unless you're going to make the "last updated" timestamp accurate up to
microseconds, there's a chance that the mailbox has had new messages
since your last check if another message was added just after you
checked it. Are you going to require that?

I think the most important thing people want is to see when new mail
appears in different mailboxes. They want to see it as fast as possible
and without the notifications getting randomly lost. Getting UIDNEXT
value is good enough for that. Other indicators would be fine as well,
as long as they are reliable.


--=-b+b3Crv/3bMDDok3XqXR
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCpJ8uyUhSUUBViskRAkxDAJ9q4IfgcNlFg2G7VVulJLA06XOLwACdG8eL
6XHp9ao0udFiI8fFzbOkWxU=
=ObUT
-----END PGP SIGNATURE-----

--=-b+b3Crv/3bMDDok3XqXR--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HrGGv054643; Mon, 6 Jun 2005 10:53:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56HrGQt054642; Mon, 6 Jun 2005 10:53:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HrFWo054633 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 10:53:15 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n293en1.watson.ibm.com (sp1n293en1.watson.ibm.com [129.34.20.41]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j56HsSOx012302 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:54:28 -0400
Received: from sp1n293en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n293en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j56Hr9s691436 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:53:09 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n293en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j56Hr8H636476 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:53:08 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118080387.27; Mon, 06 Jun 2005 13:53:07 -0400
Date: Mon, 06 Jun 2005 13:53:07 -0400
From: Barry Leiba <leiba@watson.ibm.com>
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
Message-ID: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <42A4897C.1080609@oceana.com>
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com>
X-Mailer: Mulberry/4.0.0 (Win32)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I've made the editorial changes; thanks.

> - section 6:  should we restrict selection options to be simple atoms
> as we have in this draft, or should we allow for them to be a list as
> well?   I'm thinking that some future selection option might take one
> or more arguments.  For example, the non-standard SCAN extension used
> by Pine and UW-IMAP could be reworked to look something like this:
> 
> x LIST ((SCAN "foo")) "" "bar/%"
> 
> or better yet:
> 
> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"

I have no objection, myself, to extending things further in this way,
but I point out that it *is* a significant additional extension, not
just a simple issue of grammar.  Now there may be many more things to
test in the selection mechanism, for one thing -- it's not just
inspecting whether the mailbox matches some "attribute" (like REMOTE or
SUBSCRIBED), but also whether it meets some other open-ended criteria.
The grammar may also be open-ended (as in your second example; why stop
at two astrings?), which can cause us problems.

Also also, we now have a complication with boolean conditions.  What
might "(SUBSCRIBED (SCAN "foo") (SCAN "bar"))" mean?

We should do this right, while we're doing it.  I'm just not sure, for
this situation, what "right" is.  My sense, from talking with some of
the client authors (mostly with Cyrus), is that this goes beyond what
they need this for.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HbMet053642; Mon, 6 Jun 2005 10:37:22 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56HbMMr053641; Mon, 6 Jun 2005 10:37:22 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from eagle.oceana.com (eagle.oceana.com [208.17.123.12]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HbLHM053632 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 10:37:21 -0700 (PDT) (envelope-from ken@oceana.com)
Received: from [192.168.10.26] (KEN.oceana.com [192.168.10.26]) by eagle.oceana.com (8.13.2/8.13.2) with ESMTP id j56HbCOU013783; Mon, 6 Jun 2005 13:37:12 -0400 (EDT)
Message-ID: <42A4897C.1080609@oceana.com>
Date: Mon, 06 Jun 2005 13:35:56 -0400
From: Ken Murchison <ken@oceana.com>
Organization: Oceana Matrix Ltd.
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7) Gecko/20040514
X-Accept-Language: en,pdf
MIME-Version: 1.0
To: Barry Leiba <leiba@watson.ibm.com>
CC: IMAP extensions Mailing List <ietf-imapext@imc.org>, Alexey Melnikov <Alexey.Melnikov@isode.com>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>
In-Reply-To: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0 (BAYES_00)
X-Scanned-By: MIMEDefang 2.39
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

A couple comments/questions:

- section 3, para 1, sentence 2:  "A mailbox name match a list..." -> "A 
mailbox name matches a list..."

- section 3.3, para 1: "...that the mailbox has st least one..." -> 
"...that the mailbox has at least one..."

- section 5, example 2: response for "Vegetable" mailbox is missing

- section 6:  should we restrict selection options to be simple atoms as 
we have in this draft, or should we allow for them to be a list as well? 
  I'm thinking that some future selection option might take one or more 
arguments.  For example, the non-standard SCAN extension used by Pine 
and UW-IMAP could be reworked to look something like this:

x LIST ((SCAN "foo")) "" "bar/%"

or better yet:

x LIST ((SCAN "utf-8" "foo")) "" "bar/%"


-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j54GrpNb027961; Sat, 4 Jun 2005 09:53:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j54GrpoG027960; Sat, 4 Jun 2005 09:53:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j54GrpTR027945 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 09:53:51 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id BD4FB142275 for <ietf-imapext@imc.org>; Sat,  4 Jun 2005 09:52:37 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 22154-06 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 09:52:37 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 08C0D142274 for <ietf-imapext@imc.org>; Sat,  4 Jun 2005 09:52:36 -0700 (PDT)
Mime-Version: 1.0 (Apple Message framework v622)
To: IMAP Extensions WG <ietf-imapext@imc.org>
Message-Id: <356c352e4f4d75ec6531c13c305ce835@osafoundation.org>
Content-Type: multipart/alternative; boundary=Apple-Mail-6-10380674
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Fwd: Last Call: 'Email Submission Between Independent Networks' to BCP 
Date: Sat, 4 Jun 2005 09:53:37 -0700
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--Apple-Mail-6-10380674
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	format=flowed

FYI

Begin forwarded message:

> From: The IESG <iesg-secretary@ietf.org>
> Date: June 3, 2005 6:47:09 AM PDT
> To: IETF-Announce <ietf-announce@ietf.org>
> Subject: Last Call: 'Email Submission Between Independent Networks' to 
> BCP
> Reply-To: iesg@ietf.org
>
> The IESG has received a request from an individual submitter to 
> consider the
> following document:
>
> - 'Email Submission Between Independent Networks '
>    <draft-hutzler-spamops-04.txt> as a BCP
>
> The IESG plans to make a decision in the next few weeks, and solicits
> final comments on this action.  Please send any comments to the
> iesg@ietf.org or ietf@ietf.org mailing lists by 2005-07-01.
>
> The file can be obtained via
> http://www.ietf.org/internet-drafts/draft-hutzler-spamops-04.txt
>
>
> _______________________________________________
> IETF-Announce mailing list
> IETF-Announce@ietf.org
> https://www1.ietf.org/mailman/listinfo/ietf-announce
>

--Apple-Mail-6-10380674
Content-Transfer-Encoding: 7bit
Content-Type: text/enriched;
	charset=US-ASCII

FYI


Begin forwarded message:


<excerpt><bold><color><param>0000,0000,0000</param>From:
</color></bold>The IESG <<iesg-secretary@ietf.org>

<bold><color><param>0000,0000,0000</param>Date: </color></bold>June 3,
2005 6:47:09 AM PDT

<bold><color><param>0000,0000,0000</param>To:
</color></bold>IETF-Announce <<ietf-announce@ietf.org>

<bold><color><param>0000,0000,0000</param>Subject: </color>Last Call:
'Email Submission Between Independent Networks' to BCP 

<color><param>0000,0000,0000</param>Reply-To:
</color></bold>iesg@ietf.org


The IESG has received a request from an individual submitter to
consider the 

following document:


- 'Email Submission Between Independent Networks '

   <<draft-hutzler-spamops-04.txt> as a BCP


The IESG plans to make a decision in the next few weeks, and solicits

final comments on this action.  Please send any comments to the

iesg@ietf.org or ietf@ietf.org mailing lists by 2005-07-01.


The file can be obtained via

http://www.ietf.org/internet-drafts/draft-hutzler-spamops-04.txt



_______________________________________________

IETF-Announce mailing list

IETF-Announce@ietf.org

https://www1.ietf.org/mailman/listinfo/ietf-announce


</excerpt>
--Apple-Mail-6-10380674--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j549smhk045066; Sat, 4 Jun 2005 02:54:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j549smo9045064; Sat, 4 Jun 2005 02:54:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from foon.sendmail.com (tls.sendmail.com [209.246.26.40]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j549smVS045045 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:48 -0700 (PDT) (envelope-from guenther@sendmail.com)
Received: from lab.smi.sendmail.com ([10.210.100.93]) by foon.sendmail.com (Switch-3.1.7/Switch-3.1.7) with ESMTP id j549sjNs013833 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=OK) for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:45 -0700
X-DomainKeys: Sendmail DomainKeys Filter v0.2.7 foon.sendmail.com j549sjNs013833
DomainKey-Signature: a=rsa-sha1; s=tls; d=sendmail.com; c=nofws; q=dns; b=pw2UXH1vH0IX7hOpsi/HHNXtQPBuQk+i0FLv7u5DbM5nLLaKYIihiOxJ+pEDJ0dtq +oDL786BGXYvWKkO3J6yeKX1x5NvmcKOSQjwNPesJAvGDeXuum+dgzInMOX7JCpspIY X6owaDTcxVaI4q740jDzYAZQRgDytb8feg2JR9M=
Received: from lab.smi.sendmail.com (localhost [127.0.0.1]) by lab.smi.sendmail.com (8.12.11/8.12.11) with ESMTP id j549sjlS015443 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:45 -0700 (PDT) (envelope-from guenther@lab.smi.sendmail.com)
Message-Id: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
To: IMAP Extensions WG <ietf-imapext@imc.org>
From: Philip Guenther <guenther+imap@sendmail.com>
Subject: comments on IMAP-ABNF extensions draft -02
Date: Sat, 04 Jun 2005 02:54:45 -0700
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I'm not convinced that this draft is taking the best course for defining
the syntax for extensions.  I see this document as a place to tweak how
the IMAP grammar is expressed to make extensions easier to express,
*without actually actually extending it*.  That is, extensions should be
able express their syntaxes as amending the grammar in RFC 3501 and this
draft by incremental alternative ("=/") only and without replacing
grammar elements defined in RFC 3501 or this draft, preferably without
having to repeat redundant tokens.

And yes, this goal is of course limited by our ability to predict how
commands will be extended, but the key point is what it excludes:
servers are *never* required to accept syntax for extensions that they
don't support**.  This contrasts against the (few) places where the RFC
3501 grammar *requires* clients to unilaterally accept extensible data
with open-ended semantics: body-extension, flag-extension, and
resp-text-code.  As currently written, several of the rules in this
draft have an open-ended syntax that servers would apparently have to
parse, only so that they can reject them with a BAD.

I also disagree with almost all uses of RFC 2119 keywords.  The only two
that make sense to me are both in section 3, in the lead in prose
("Implementations MUST accept...") and in the comment following the
definition of the 'search' rule ("...to SEARCH MUST be registered...").
Of the other nine uses of "MUST", five of them are directed at the
authors of future RFCs and not at clients or servers and are therefore
neither testable nor enforcable: a future RFC that wanted to violate
such a requirement could simply amend/updated _this_ doc to permit
whatever syntax it wanted.***  The other four require servers to reject
with a BAD response uses of extension syntax that they don't
understand, but that can be dropped if the syntax wasn't defined in the
open-ended manner it currently is.

(If a client sends stuff that doesn't meet the syntax of RFC 3501 plus
the RFCs for the extensions supported the server, then the server can do
d**n well whatever it pleases.  If you don't believe that, please
consider how existing servers *do* react when a client is non-compliant
in such ways as including a CRNL in a quoted string as an argument for
a (possibly unknown) command.  GIGO!)


So, for example, consider the append-opt-* rules.  As currently written,
they are an open-ended syntax that servers have to parse, then reject
with a BAD.  Rather than do that, I think this would be better expressed
as a combination of prose and a rule that is *not* referenced on the
left-hand-side of any other rules.  I.e., something like:

     append          = "APPEND" SP mailbox 1*append-message
                         ;; more than one append-message may appear only
                         ;; if MULTIAPPEND [MULTIAPPEND] capability is
                         ;; present

     append-data     = literal / literal8
                         ;; literal8 may appear only if BINARY [BINARY]
                         ;; capability is present.
                         ;; If append-data is extended, it is strongly
                         ;; recommended that the extension match a
                         ;; subset of the tagged-ext rule syntax

     append-ext      =   <for extension only>
                         ;; This rule exists solely to be augmented by
                         ;; extensions via incremental alternative
                         ;; ("=/") rules.  It is strongly recommended
                         ;; that such extensions match a subset of the
                         ;; tagged-ext rule syntax

     append-message  = append-opts SP append-data

     append-opts     = [SP flag-list] [SP date-time] *(SP append-ext)
...
     tagged-ext      = atom SP tagged-ext-val
                         ;; recommended overarching syntax for
                         ;; extensions

     tagged-ext-val  = astring /
                       "(" tagged-ext-val *(SP tagged-ext-val) ")"
                         ;; extensions that follow this general syntax
                         ;; should use nstring instead of astring when
                         ;; appropriate in the context of the extension
<< or should this just say astring / nstring / "(" ...? >>


The key difference between this and the current draft is that where this
syntax permits something beyond RFC 3501, it's because of (and tied to)
a previously published extension, period.  There's no need of a "MUST be
rejected with a BAD" clause for unknown extensions, because they don't
match the syntax, period.


IMHO, the tagged-ext syntax (or something similarly simple) is Good
Enough to be referenced by FETCH, STORE, SELECT, and SEARCH too.  It
certainly appears to cover the already scoped extensions for CONDSTORE,
ANNOTATE, and several others. Using a single syntax for extensions to
commands would make the syntax part of this draft easier to read.  If it
*isn't* sufficient, then I suspect the Right Thing might be to simply
say that extensions should consist of a atom as a tag, followed by
exactly one argument with balanced parenthesis.


On the issue of what this draft should include to make future extensions
easier to express: it seems unfortunate that RFC 3501's response-data
rule includes the leading "*" SP and the trailing CRLF, such that
extensions like ANNOTATEMORE have to duplicate them, ala:

       annotate-data     = "ANNOTATION" SP mailbox SP entry-list
                           ; empty string for mailbox implies
                           ; server annotation.
...
       response-data     /= "*" SP annotate-data CRLF
                           ; adds to original IMAP data responses

instead of just

       response-foo      /= "ANNOTATION" SP mailbox SP entry-list
                           ; empty string for mailbox implies
                           ; server annotation.

RFC 3501 got it right with resp-text-code; can this draft fix things for
response-data?  Perhaps:

   response-data = "*" SP resp-data CRLF
   resp-data     = resp-cond-state / resp-cond-bye / mailbox-data /
		   message-data / capability-data

Okay, so that's not the greatest pair of names for rules, but you get
the point.

**  It could be argued that x-command breaks this rule, but that syntax
    rule isn't actually defined unambiguously.  IMHO, it's appearance in
    the grammer is a mistake.

*** Whether you can place constraints on your successors is a pretty
    subtle issue in the theory of law and rule-making societies.  I
    haven't been participating in the IETF for very long, but to my eye,
    the _only_ constraint that a substantive RFC (i.e., one that isn't
    actually amending the IETF's own procedures) can place on those that
    seek to amend or replace it is the requirement that they (also) be
    on the standards track, or maybe that they be of the same or higher
    level on the standard track.

    If you enjoy pondering these sorts of issues, I highly recommend
    that you track down and read "The Paradox of Self-Amendment" by
    Peter Suber and consider his discussion of "partially
    self-entrenched" laws.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j544xHMj034886; Fri, 3 Jun 2005 21:59:17 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j544xH6x034885; Fri, 3 Jun 2005 21:59:17 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from foon.sendmail.com (tls.sendmail.com [209.246.26.40]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j544xEbq034879 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 21:59:16 -0700 (PDT) (envelope-from guenther@sendmail.com)
Received: from lab.smi.sendmail.com ([10.210.100.93]) by foon.sendmail.com (Switch-3.1.7/Switch-3.1.7) with ESMTP id j544xCLm011391 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=OK); Fri, 3 Jun 2005 21:59:12 -0700
X-DomainKeys: Sendmail DomainKeys Filter v0.2.7 foon.sendmail.com j544xCLm011391
DomainKey-Signature: a=rsa-sha1; s=tls; d=sendmail.com; c=nofws; q=dns; b=FPrxh9mf9qRelwq7myzPVP1u5Z9OGI97eCwnkpq8C9jNaOZL0kt1PukNHbIheJO5W FLIYiKY1sxlCFjBXlhQhYCvtSvNrKA4UHwhqee8c4PKXyWdJAQBOiujnRuGBpsBxkkY 6Ck1D+N6OCfRSYlAd+1O5oQCS0JhyyBH8yL9yBs=
Received: from lab.smi.sendmail.com (localhost [127.0.0.1]) by lab.smi.sendmail.com (8.12.11/8.12.11) with ESMTP id j544xBhm090612; Fri, 3 Jun 2005 21:59:12 -0700 (PDT) (envelope-from guenther@lab.smi.sendmail.com)
Message-Id: <200506040459.j544xBhm090612@lab.smi.sendmail.com>
To: Cyrus Daboo <daboo@isamet.com>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
From: Philip Guenther <guenther+imap@sendmail.com>
Subject: Re: ANNOTATEMODE uses string rule? 
In-reply-to: <9BCBAA2CA31A73366FE5D0A2@ninevah.local> 
References: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U>  <9BCBAA2CA31A73366FE5D0A2@ninevah.local> 
Date: Fri, 03 Jun 2005 21:59:11 -0700
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Cyrus Daboo <daboo@isamet.com> writes:
>--On June 3, 2005 3:26:14 PM -0700 Mark Crispin <MRC@cac.washington.edu> 
>wrote:
>> I just learned that the ANNOTATEMORE extension uses the string rule,
>> rather than astring or nstring, for several of its rules.
>>
>> This creates an unnecessary new grammar for clients to send and servers
>> to process, both for tokens and the list of tokens.  Please give serious
>> consideration to using astring instead, which will permit the use of
>> existing grammar generators and processors.
...
>Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
>rather than astring. Is there anyone else that feels the same way as Mark?

I basicly agree.  To be precise, I think that 'att-match' and
'entry-match' should be defined as 'list-mailbox', while 'attrib' and
'entry' should be defined as 'astring' (i.e., mirror/reuse the syntax
categories used for mailbox names).


Philip Guenther



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NeNli016880; Fri, 3 Jun 2005 16:40:23 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53NeNl7016879; Fri, 3 Jun 2005 16:40:23 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NeMhw016873 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 16:40:22 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53NeJ8D010118 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 3 Jun 2005 16:40:19 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53NeJic001822 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Fri, 3 Jun 2005 16:40:19 -0700
Date: Fri, 3 Jun 2005 16:38:16 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Cyrus Daboo <daboo@isamet.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: ANNOTATEMODE uses string rule?
In-Reply-To: <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
Message-ID: <Pine.WNT.4.64.0506031636510.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U> <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Fri, 3 Jun 2005, Cyrus Daboo wrote:
> Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
> rather than astring. Is there anyone else that feels the same way as Mark?

I apologize for not noticing this earlier.  My guess is that this use of 
string rather than astring is an influence from ACAP.  However, it is a 
major shift from past IMAP precedent; IMAP uses either astring or nstring.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NG4pp015323; Fri, 3 Jun 2005 16:16:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53NG4Ym015322; Fri, 3 Jun 2005 16:16:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from darius.cyrusoft.com (darius.cyrusoft.com [63.163.82.2]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NG2Kc015312 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 16:16:02 -0700 (PDT) (envelope-from daboo@isamet.com)
Received: from user-152-3-214-115.wireless.duke.edu (pool-151-201-34-147.pitt.east.verizon.net [151.201.34.147]) (authenticated bits=0) by darius.cyrusoft.com (8.12.9/8.12.9) with ESMTP id j53MoNIC008921 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 3 Jun 2005 18:50:24 -0400
Date: Fri, 03 Jun 2005 19:15:57 -0400
From: Cyrus Daboo <daboo@isamet.com>
To: Mark Crispin <MRC@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: ANNOTATEMODE uses string rule?
Message-ID: <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
In-Reply-To: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.EDU>
References:  <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U>
X-Mailer: Mulberry/4.0.0 (Mac OS X)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Spam-Status: No, hits=0.0 tests=none
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Hi Mark,

--On June 3, 2005 3:26:14 PM -0700 Mark Crispin <MRC@cac.washington.edu> 
wrote:

>
> I just learned that the ANNOTATEMORE extension uses the string rule,
> rather than astring or nstring, for several of its rules.
>
> This creates an unnecessary new grammar for clients to send and servers
> to process, both for tokens and the list of tokens.  Please give serious
> consideration to using astring instead, which will permit the use of
> existing grammar generators and processors.
>
> The existing IMAP grammar for clients to send is remarkably small; there
> are only about 10 basic types.  So, adding string (as opposed to astring)
> and list of string is a 20% increase in the client-generated grammar.

Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
rather than astring. Is there anyone else that feels the same way as Mark?

-- 
Cyrus Daboo



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MSIFi011800; Fri, 3 Jun 2005 15:28:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53MSIj3011799; Fri, 3 Jun 2005 15:28:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MSHH2011793 for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:18 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MSHoG029581 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:17 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MSGQg022501 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT) for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:16 -0700
Date: Fri, 3 Jun 2005 15:26:14 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: ANNOTATEMODE uses string rule?
Message-ID: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.EDU>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I just learned that the ANNOTATEMORE extension uses the string rule, 
rather than astring or nstring, for several of its rules.

This creates an unnecessary new grammar for clients to send and servers to 
process, both for tokens and the list of tokens.  Please give serious 
consideration to using astring instead, which will permit the use of 
existing grammar generators and processors.

The existing IMAP grammar for clients to send is remarkably small; there 
are only about 10 basic types.  So, adding string (as opposed to astring) 
and list of string is a 20% increase in the client-generated grammar.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MDxMn011006; Fri, 3 Jun 2005 15:13:59 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53MDxUu011005; Fri, 3 Jun 2005 15:13:59 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MDwcM010998 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 15:13:58 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MDrAg026747 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 3 Jun 2005 15:13:54 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MDrTF019805 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Fri, 3 Jun 2005 15:13:53 -0700
Date: Fri, 3 Jun 2005 15:11:50 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <06480646801216@PD.gordano.com>
Message-ID: <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Fri, 3 Jun 2005, Simon Tyler wrote:
> It does not detract from the fact that a client needs UIDVALIDITY on all 
> synchronised mailboxes which is why I wanted it added to the LIST 
> response.

Isn't it really the case that you only need UIDVALIDITY/UIDNEXT on 
mailboxes which have a changed "last updated" date/time from the last 
time?  Do you really need to be told UIDVALIDITY/UIDNEXT on mailboxes 
which are unchanged?  If so, why?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53A042Z076791; Fri, 3 Jun 2005 03:00:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53A04bF076790; Fri, 3 Jun 2005 03:00:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j53A02mS076765 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 03:00:04 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 501791A745; Fri,  3 Jun 2005 12:00:00 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 863FC1A6B9; Fri,  3 Jun 2005 11:59:53 +0200 (CEST)
Message-Id: <U1Bslde1KJ+IzN/W9M8M+g.md5@prosecco.oryx.com>
Date: Fri, 3 Jun 2005 11:59:26 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>, Mark Crispin <MRC@CAC.Washington.EDU>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin writes:
> Synchronization is difficult.  There is no doubt about that fact.  
> We've spent a decade talking about it with no published 
> specification.  Why don't we roll up our sleeves and talk about how 
> to do it right, rather than consider expensive (to some) 
> half-solutions that are at best "good enough" (to others) but don't 
> quite solve the problem?

The best is the enemy of the good.

A decade has passed, as you say. I think that, after a decade, it's 
about time to lower our aims and be content with a "good enough" 
solution.

My two cents. And I'm not necessarily content with Simon's proposed solution.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536dnba004478; Thu, 2 Jun 2005 23:39:49 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j536dn0Q004477; Thu, 2 Jun 2005 23:39:49 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536dksC004450 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 23:39:46 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id tcuudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:48:06 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id scuudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:48:06 +0100
Date: Fri, 3 Jun 2005 07:48:06 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "IMAP Extensions WG" <ietf-imapext@imc.org>
Message-Id: <06480646801216@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_06480646801217/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_06480646801217/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_06480646801218/PD.gordano.com==_"


--==_06480646801218/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

=20
 =20
(Sorry, resent as I have a horrible habit of not typing the word "not" when=
 I need to!)=20
 =20
CONDSTORE would be a very useful addition for us (although more painful to=
 implement :-( ). Although again, I would like to see some of it's=
 information returned through a LIST response - perhaps the HIGHESTMODSEQ=
 to allow a client to detect changed folders without having to select=
 them.=20
 =20
It does not detract from the fact that a client needs UIDVALIDITY on all=
 synchronised mailboxes which is why I wanted it added to the LIST=
 response.=20
 =20
Simon=20

---- Message from mailto:<mrc@cac.washington.edu Mark Crispin=
 <mrc@cac.washington.edu> at 02-Jun-2005 12:19:44 ------=20


At our WG chair's suggestion (thank you Lisa!), I wish to clarify the=20
following statement:
>  This is the direction of cooperation, rather than trying to boost your=20
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's=20
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into.=20
It isn't just a matter of conscious intent; it is also how it is=
 perceived=20
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST=20
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this=20
does not mean that we have to deploy the first tool is proposed.  We=20
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon=20
last-modified date/time.  We have a start in that direction with=20
CONDSTORE.  Is there a problem with this strawman that I missed?  Of=20
course, it's less painful for my server; but I also believe that it's=20
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.








--==_06480646801218/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><FONT face=3DArial size=3D2>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P>(Sorry, resent as I have a horrible habit of not typing the word "not"=
 when I need to!)</P>
<P>&nbsp;</P>
<P><FONT face=3DArial size=3D2>CONDSTORE would be a very useful addition=
 for us (although more painful to implement :-( ). Although again, I would=
 like to see some of it's information returned through a LIST response -=
 perhaps the HIGHESTMODSEQ to allow a client to detect changed folders=
 without having to select them.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>It does not detract from the fact that a=
 client needs UIDVALIDITY on all synchronised mailboxes which is why I=
 wanted it added to the LIST response.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@cac.washington.edu"=
 target=3D_blank>Mark Crispin &lt;mrc@cac.washington.edu&gt;</A> at=
 02-Jun-2005 12:19:44 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>At our WG chair's suggestion (thank you Lisa!), I=
 wish to clarify the <BR>following statement:<BR>&gt; &nbsp;This is the=
 direction of cooperation, rather than trying to boost your <BR>&gt; own=
 product at the expense of others.<BR><BR>as following:<BR><BR>This is the=
 direction of cooperation, rather than trying to boost one's <BR>own=
 product at the expense of others.<BR><BR>Nobody is doing such boosting=
 now, but it is an easy trap to fall into. <BR>It isn't just a matter of=
 conscious intent; it is also how it is perceived <BR>and the impact upon=
 others.<BR><BR>All of us MUST be sensitive to the problems that others=
 face, and we MUST <BR>NOT be dismissive when someone says "this causes me=
 a problem."<BR><BR>We can all agree that IMAP is lacking in=
 synchronization tools. &nbsp;But this <BR>does not mean that we have to=
 deploy the first tool is proposed. &nbsp;We <BR>should consider=
 alternatives.<BR><BR>I've suggested one such alternative as a strawman: a=
 mechanism based upon <BR>last-modified date/time. &nbsp;We have a start in=
 that direction with <BR>CONDSTORE. &nbsp;Is there a problem with this=
 strawman that I missed? &nbsp;Of <BR>course, it's less painful for my=
 server; but I also believe that it's <BR>technically superior.<BR><BR>--=
 Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science does not emerge=
 from voting, party politics, or public debate.<BR>Si vis pacem, para=
 bellum.<BR><BR><BR><BR></BLOCKQUOTE></FONT><BR><BR></FONT></body></html>

--==_06480646801218/PD.gordano.com==_--
--==_06480646801217/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536Uosf001265; Thu, 2 Jun 2005 23:30:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j536Uoju001264; Thu, 2 Jun 2005 23:30:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536UlcO001228 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 23:30:48 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id dauudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:39:05 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id cauudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:39:05 +0100
Date: Fri, 3 Jun 2005 07:39:05 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "IMAP Extensions WG" <ietf-imapext@imc.org>
Message-Id: <06390556201197@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
In-Reply-To: <Pine.OSX.4.63.0506021204100.266@pangtzu.panda.com>
References: <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com> <1d61bc9cc57fa16d7c700383fe6e7b98@osafoundation.org> <Pine.OSX.4.63.0506021148170.266@pangtzu.panda.com> <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_06390556201198/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_06390556201198/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_06390556201199/PD.gordano.com==_"


--==_06390556201199/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
CONDSTORE would be a very useful addition for us (although more painful to=
 implement :-( ). Although again, I would like to see some of it's=
 information returned through a LIST response - perhaps the HIGHESTMODSEQ=
 to allow a client to detect changed folders without having to select=
 them.=20
 =20
It does detract from the fact that a client needs UIDVALIDITY on all=
 synchronised mailboxes which is why I wanted it added to the LIST=
 response.=20
 =20
Simon=20

---- Message from mailto:<mrc@cac.washington.edu Mark Crispin=
 <mrc@cac.washington.edu> at 02-Jun-2005 12:19:44 ------=20


At our WG chair's suggestion (thank you Lisa!), I wish to clarify the=20
following statement:
>  This is the direction of cooperation, rather than trying to boost your=20
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's=20
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into.=20
It isn't just a matter of conscious intent; it is also how it is=
 perceived=20
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST=20
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this=20
does not mean that we have to deploy the first tool is proposed.  We=20
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon=20
last-modified date/time.  We have a start in that direction with=20
CONDSTORE.  Is there a problem with this strawman that I missed?  Of=20
course, it's less painful for my server; but I also believe that it's=20
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.






--==_06390556201199/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>CONDSTORE would be a very useful addition=
 for us (although more painful to implement :-( ). Although again, I would=
 like to see some of it's information returned through a LIST response -=
 perhaps the HIGHESTMODSEQ to allow a client to detect changed folders=
 without having to select them.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>It does detract from the fact that a client=
 needs UIDVALIDITY on all synchronised mailboxes which is why I wanted it=
 added to the LIST response.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@cac.washington.edu">Mark=
 Crispin &lt;mrc@cac.washington.edu&gt;</A> at 02-Jun-2005 12:19:44=
 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>At our WG chair's suggestion (thank you Lisa!), I=
 wish to clarify the <BR>following statement:<BR>&gt; &nbsp;This is the=
 direction of cooperation, rather than trying to boost your <BR>&gt; own=
 product at the expense of others.<BR><BR>as following:<BR><BR>This is the=
 direction of cooperation, rather than trying to boost one's <BR>own=
 product at the expense of others.<BR><BR>Nobody is doing such boosting=
 now, but it is an easy trap to fall into. <BR>It isn't just a matter of=
 conscious intent; it is also how it is perceived <BR>and the impact upon=
 others.<BR><BR>All of us MUST be sensitive to the problems that others=
 face, and we MUST <BR>NOT be dismissive when someone says "this causes me=
 a problem."<BR><BR>We can all agree that IMAP is lacking in=
 synchronization tools. &nbsp;But this <BR>does not mean that we have to=
 deploy the first tool is proposed. &nbsp;We <BR>should consider=
 alternatives.<BR><BR>I've suggested one such alternative as a strawman: a=
 mechanism based upon <BR>last-modified date/time. &nbsp;We have a start in=
 that direction with <BR>CONDSTORE. &nbsp;Is there a problem with this=
 strawman that I missed? &nbsp;Of <BR>course, it's less painful for my=
 server; but I also believe that it's <BR>technically superior.<BR><BR>--=
 Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science does not emerge=
 from voting, party politics, or public debate.<BR>Si vis pacem, para=
 bellum.<BR><BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_06390556201199/PD.gordano.com==_--
--==_06390556201198/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52LxBpv034445; Thu, 2 Jun 2005 14:59:11 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52LxB9f034444; Thu, 2 Jun 2005 14:59:11 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52Lx9E4034435 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 14:59:10 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 0C0771A75D; Thu,  2 Jun 2005 23:59:06 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id A9FBE1A6E0; Thu,  2 Jun 2005 23:59:00 +0200 (CEST)
Message-Id: <Jqf492O+yMCM5gT0I8WDlg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 23:58:37 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com> <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com> <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com> <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52LxAE4034438
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> So your calculation is a best case calculation where all the STATUS 
> commands are pipelined and mine is the worst case calculation where 
> none of them are.   The answer is likely to be somewhere in between.

IMO, it's likely to be close enough to full pipelining that the 
roundtrip cost becomes insignificant compared to the traffic cost.

...
> So it is your belief that it is easy to do the command batching than 
> to code fallbacks?

Once the cost of testing is added: Yes, I think so, as a rule of thumb. 
Not always. It depends on how many additional test cases the fallback 
creates, etc.

> Perhaps.   I expect there to be widely variant opinions on that as 
> well, but as we are talking about implementation again, it just 
> doesn't seem worth the discussion.

We can't really keep implementation discussion completely away. Unless 
we want to end up like CORBA, of course.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52JJp4i021441; Thu, 2 Jun 2005 12:19:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52JJpRX021440; Thu, 2 Jun 2005 12:19:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52JJoGb021433 for <ietf-imapext@IMC.ORG>; Thu, 2 Jun 2005 12:19:50 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52JJlS9028919 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 12:19:48 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52JJjYt018310 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 12:19:47 -0700
Date: Thu, 2 Jun 2005 12:19:44 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
cc: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
Message-ID: <Pine.OSX.4.63.0506021204100.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com> <1d61bc9cc57fa16d7c700383fe6e7b98@osafoundation.org> <Pine.OSX.4.63.0506021148170.266@pangtzu.panda.com> <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

At our WG chair's suggestion (thank you Lisa!), I wish to clarify the 
following statement:
>  This is the direction of cooperation, rather than trying to boost your 
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's 
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into. 
It isn't just a matter of conscious intent; it is also how it is perceived 
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST 
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this 
does not mean that we have to deploy the first tool is proposed.  We 
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon 
last-modified date/time.  We have a start in that direction with 
CONDSTORE.  Is there a problem with this strawman that I missed?  Of 
course, it's less painful for my server; but I also believe that it's 
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IgeIr018107; Thu, 2 Jun 2005 11:42:40 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Ige46018106; Thu, 2 Jun 2005 11:42:40 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout5.cac.washington.edu (mxout5.cac.washington.edu [140.142.32.135]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IgepF018100 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:42:40 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout5.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Igd0U008082 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:42:39 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Igc9r009519 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:42:39 -0700
Date: Thu, 2 Jun 2005 11:42:37 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021137450.266@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> You're acting like this is the first time that this
> has been asked for and it's not.

With all due respect Steve, you also are sounding like a broken record. 
It isn't as if you've been exactly neutral with respect to your own 
products either.

Can we please leave this baiting aside, and discuss the real issues?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Ia7LF017717; Thu, 2 Jun 2005 11:36:07 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Ia7bL017716; Thu, 2 Jun 2005 11:36:07 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Ia6wm017710 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:36:06 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Ia3ZA016385 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:36:03 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Ia1ee007861 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:36:03 -0700
Date: Thu, 2 Jun 2005 11:36:01 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021126490.266@pangtzu.panda.com>
References: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>    <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>    <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com> <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> The STATUS command was added as a means to acquire folder metadata.
> The reason it was added was because a large number of client implementors
> wanted to markup their displays with this information because their
> customers wanted to see it.

That was not why STATUS was added.  Remember that I'm the person who came 
up with it.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IQERu017058; Thu, 2 Jun 2005 11:26:14 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52IQEYk017057; Thu, 2 Jun 2005 11:26:14 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IQEOS017050 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:26:14 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IQApZ009592 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:26:10 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IQ8uq025304 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:26:09 -0700
Date: Thu, 2 Jun 2005 11:26:07 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> Hogwash.  If you are writing a general purpose client then you have to
> implement fallbacks for the various optional features that are already in
> play and it isn't that hard.

Not so.  Clients exist today that depend upon an optional feature, and 
declare the server to be broken that lack that optional feature.

> This particular requirement is an attempt to make MUCH more difficult
> tasks in the client (synchronization, display markup, user notification)
> easier and perform better.

Except that the proposal in question doesn't really accomplish that 
directly.  It is a possible solution that may be acceptable in some 
circumstances, but it dances around the overall problem (for example, it 
can't detect external deletions).

I have a general dislike of quick and dirty "solutions" that don't solve 
the general problem and impose potentially high costs.

Synchronization is difficult.  There is no doubt about that fact.  We've 
spent a decade talking about it with no published specification.  Why 
don't we roll up our sleeves and talk about how to do it right, rather 
than consider expensive (to some) half-solutions that are at best "good 
enough" (to others) but don't quite solve the problem?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IJiP1016688; Thu, 2 Jun 2005 11:19:44 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52IJiq4016687; Thu, 2 Jun 2005 11:19:44 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IJhV4016680 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:19:43 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IJehP015155 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:19:41 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IJdaO023835 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:19:40 -0700
Date: Thu, 2 Jun 2005 11:19:39 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> If the feature is useful, then it will be implemented by clients and
> servers that wish to implement it.   If this causes feature pressure on
> implementations, then too bad.   The feature will either have merit and be
> useful or it won't.   It is simply one of a set of features that
> contribute to the selection of one implementation over another.   Denying
> features in optional services just because some servers can't implement it
> efficiently, but some clearly can, it ludicrous.   It smacks of software
> socialism :-)

The contrary argument is that certain other protocols are littered with 
pie-in-the-sky features that someone added, but were never widely 
implemented.  The result is that, when reading the protocol specification, 
it is impossible to determine what features need to be implemented and 
what features can be omitted.

This problem has been with us since the beginning; consider TELNET and 
FTP.  TELNET and FTP are full of crud that someone thought was important 
30 years ago, yet are inadequate to do today's needs.  I've heard all 
sorts of horror stories about the FTP meetings 30+ years ago.

This has an additional harmful effect.  There is at least one server 
vendor which believes that SEARCH is optional to implement on the server. 
This vendor realized that nobody implements every aspect of every IMAP 
document, and assumed that this gave him the discretion to apply his own 
razor...to the base specification.

Feature bloat is not a good thing.  It is something to be avoided at all 
cost.  Every new feature proposal must justify its own existance.  It 
can't just be a good idea; it has to be an overwhelmingly good idea above 
and beyond any harm that it will cause.

It has already been pointed out that this proposed LIST extension is 
really a way to determine, for synchronization purposes, what has changed 
since the last synchronization.  Rather than take the first idea that 
comes along for how to do this, we should carefully consider how best to 
do this.

One thing that immediately comes to mind are tools that use a "last 
changed" date/time.  That is basically an extension of the \Marked and 
\UnMarked functionality, only now we're talking about synchronizing a 
disconnected client rather than showing an online client what's new.

Not only does this have less potential cost, it may also lead in new 
desirable directions.  For example, maybe LIST could have a means to 
filter mailboxes based upon "last changed" date/time.

Thinking in the direction of "how to do new things that are good for 
everybody" would be more productive.  This is the direction of 
cooperation, rather than trying to boost your own product at the expense 
of others.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Hte5C014979; Thu, 2 Jun 2005 10:55:40 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HtemF014978; Thu, 2 Jun 2005 10:55:40 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HtdjG014971 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:55:39 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52HtcTn005075 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 10:55:39 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52HtS6g029860 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 10:55:38 -0700
Date: Thu, 2 Jun 2005 10:55:27 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Barry Leiba <leiba@watson.ibm.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
Message-ID: <Pine.OSX.4.63.0506021044240.266@pangtzu.panda.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <429F24BC.6000908@watson.ibm.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Barry Leiba wrote:
> But the protocol MUST do <y>, though THAT is expensive to do with some legacy 
> message stores (that he doesn't care about).  (An example of <y> that comes 
> to mind is "return an exact length for RFC822.SIZE, though clients can easily 
> work around that.)  Again, let's not spend a lot of time talking about 
> RFC822.SIZE now -- I want to make the point that we all bring our 
> implementation biases with us.

For the record, if I had known that there was an issue for any potential 
server with lengths, I would have designed IMAP differently.

When the requirements for RFC822.SIZE were stated, I did not have a clue 
that it could possibly be expensive for Exchange.  Exchange did not exist 
at that time, and Microsoft was a relatively small company that produced 
DOS and hired programmers based on how well they came up with an algorithm 
to do tic-tac-toe with bit arithmetic.

The client can NOT easily work around non-exact lengths.  The purpose of 
exact lengths is for the client to allocate space; the same is the case 
with literals.  If the client does partial fetching, it MUST have exact 
lengths.

I knew about UNIX's use of newline-only newlines, but newline conversion 
had a long history in the network and it was asserted repeatedly that this 
was trivial.  The implementation complaint from the UNIX guys was that 
this IMAP proposal of mine provided for multiple simultaneous access which 
was "impossible" to do on UNIX.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HJ1od010842; Thu, 2 Jun 2005 10:19:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HJ19x010841; Thu, 2 Jun 2005 10:19:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HJ0WP010825 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:19:00 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 11:20:43 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 11:18:54 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com>
References: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com> <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>    <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>    <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>    <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
Message-ID: <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52HJ0WP010828
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 18:49:28 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> No. Of sending all the STATUS commands at the same time. Perfectly legal 
> in IMAP, works with all servers:
> 
> C: a STATUS asfd (UIDNEXT UIDVALIDITY)
> C: b STATUS qwre (UIDNEXT UIDVALIDITY)
> C: c STATUS zxvc (UIDNEXT UIDVALIDITY)
> S: * STATUS asfd (UIDNEXT 124 UIDVALIDITY 43)
> S: a OK
> S: * STATUS qwre (UIDNEXT 1 UIDVALIDITY 1)
> S: b OK
> S: * STATUS zxvc (UIDNEXT 2 UIDVALIDITY 3)
> S: c OK

So your calculation is a best case calculation where all the STATUS 
commands are pipelined and mine is the worst case calculation where none 
of them are.   The answer is likely to be somewhere in between.   Many 
clients will batch to some point, so I'll grant that.   They probably 
should all do that.   I wonder how many actually do.   
 
> More difficult to implement in a client than an RPC-like interface, but 
> IMO less than twice as hard, so quite likely simpler than implementing 
> both with and without an extension.

So it is your belief that it is easy to do the command batching than to 
code fallbacks?   Perhaps.   I expect there to be widely variant opinions 
on that as well, but as we are talking about implementation again, it just
doesn't seem worth the discussion.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HAI8v009984; Thu, 2 Jun 2005 10:10:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HAITn009983; Thu, 2 Jun 2005 10:10:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HAHVc009971 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:10:17 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 11:11:57 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 11:10:08 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com>
References: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>    <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>    <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Message-ID: <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52HAHVc009978
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 18:05:46 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> One excellent way to placate me is to explain how the proposal outlined 
> in this thread relates to this:
> 
> > This particular requirement is an attempt to make MUCH more difficult 
> > tasks in the client (synchronization, display markup, user 
> > notification) easier and perform better.

The STATUS command was added as a means to acquire folder metadata.   
The reason it was added was because a large number of client implementors 
wanted to markup their displays with this information because their 
customers wanted to see it.    At that time there was discussion on 
returning the metadata as part of the LIST response.   It was rejected 
because LIST is mandatory to implement in the base specification and MUST 
be fast.    Adding the requirement for returning metadata to a base 
command puts an unreasonable load on implementations that may not be able 
to bear it.

That doesn't obviate the requirement for the metadata ... thus the STATUS 
command.

Roundtrips are painful from a client standpoint.    This is especially 
true when your client is in fact a service -- ie. a J2EE or Web Service.  
If you need the information for your display then you currently have to 
get it by issuing multiple STATUS commands.   This is mandated by the 
protocol.    It doesn't matter how fast your server is at answering the 
question, because the roundtrips and associated IO overwhelm all other 
performance issues.

If you optionally add the metadata to the LIST response you eliminate the 
roundtrips.   Then, to get the expected performance, your server must be 
able perform the requested operation efficiently.   Many can.   If you are
building service class infrastructure, then you will select one that does.

But, most importantly, you have eliminated the protocol inefficiency that 
causes services delays regardless of how good or bad your store 
implementation is.

To summarize.   The requirement has some merit.   It was a clear 
performance and feature win in a proprietary form for us both with service 
clients and desktop clients.   It is just one of several useful features 
that have been discussed wrt LIST extensions.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Gpd75008560; Thu, 2 Jun 2005 09:51:39 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Gpd05008559; Thu, 2 Jun 2005 09:51:39 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52GpUcI008545 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:51:39 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 10:53:12 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 10:51:24 -0700
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
Message-ID: <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52GpdcI008554
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 06:14:03 -0700 (PDT) Mark Crispin 
<mrc@cac.washington.edu> wrote:

> 
> On Thu, 2 Jun 2005, Simon Tyler wrote:
> > Whether or not a server stores UIDNEXT or calculates it is a server 
> > implementation decision. It is irrelevant when designing a protocol.
> 
> This is completely wrong.

No he's not.   His request is just difficult for you to implement.

> For better or worse, Internet protocols are molded to accomodate existing 
> implementations.  You will receive strong pushback on any proposed 
> protocol or protocol enhancement whose impact is perceived as harmful to 
> one or more implementations.  Complaints that "it's a poor specification" 
> if your demands are not fufilled will fall on deaf ears.

This is an argument that has been seriously overused and has prevented 
some very useful features from being added.   Make it optional and 
don't implement it.   If people *really* want it, then you will have to 
figure out a way to make it work.   If they don't, then continue to not 
implement it.
 
> In fact, what you want is to have IMAP molded to accomodate *your* client 
> implementation.  Rather than fix your client so it doesn't do something as 
> ill-conceived as a STATUS on every LISTed mailbox, you want to change the 
> protocol so the server does the stupid thing for you automatically, all to 
> save your client some RTTs.  It doesn't address the underlying problem.

Also not true Mark.   You're acting like this is the first time that this 
has been asked for and it's not.   In fact it has come up repeatedly over 
the years.   It is a legitmate client requirement.   Yes, it requires more
work from the server.    That's why it is called a "server".

> The entire point of the Internet protocol architecture is a set of 
> interoperable protocols which can be implemented across a wide variety of 
> systems, including systems with legacy infrastructure.

Which definitely applies to base protocols but does not necessarily apply 
to extensions.  Extensions are optional.

I believe that the problem you've had with the LIST extensions all along 
is that you believe client implementors will implement them -- in droves.
Because they are useful.  And that will put pressure on server 
implementations that have difficulty in providing these services.

> The design of protocols for a perfect world, in which you never need to 
> concern yourself with the other guy's problems, is not the realm of 
> Internet.  It is the realm of proprietary protocols and systems.

> In the Internet world, interoperability is king, and everything else must 
> submit to the demands of that king.

And I submit that you are currently hurting interoperability wrt to LIST 
extensions.   The metadata features are so useful that many people are 
implementing proprietary extensions none of which have a chance of 
interoperating.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Go0eW008430; Thu, 2 Jun 2005 09:50:00 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Go0wB008429; Thu, 2 Jun 2005 09:50:00 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52GnwoW008423 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:49:59 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 30C341A74F; Thu,  2 Jun 2005 18:49:58 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id BF55C1A707; Thu,  2 Jun 2005 18:49:51 +0200 (CEST)
Message-Id: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 18:49:28 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com> <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> On Thu, 2 Jun 2005 10:12:42 +0200 Arnt Gulbrandsen
> <arnt@gulbrandsen.priv.no> wrote:
>> Alternative 1 has a cost of two roundtrips, plus traffic costs for 
>> the LIST responses and for the responses to STATUS, plus CPU costs 
>> for LIST, STATUS and quite a bit of parsing.
>
> Almost right.   Alternative 1 has a cost of 1 roundtrip for the LIST 
> and 1 roundtrip for each STATUS per interesting mailbox.  Perhaps you 
> are thinking of a STATUS extension that accepts multiple mailboxes as 
> arguments?

No. Of sending all the STATUS commands at the same time. Perfectly legal 
in IMAP, works with all servers:

C: a STATUS asfd (UIDNEXT UIDVALIDITY)
C: b STATUS qwre (UIDNEXT UIDVALIDITY)
C: c STATUS zxvc (UIDNEXT UIDVALIDITY)
S: * STATUS asfd (UIDNEXT 124 UIDVALIDITY 43)
S: a OK
S: * STATUS qwre (UIDNEXT 1 UIDVALIDITY 1)
S: b OK
S: * STATUS zxvc (UIDNEXT 2 UIDVALIDITY 3)
S: c OK

More difficult to implement in a client than an RPC-like interface, but 
IMO less than twice as hard, so quite likely simpler than implementing 
both with and without an extension.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52GbiP0007212; Thu, 2 Jun 2005 09:37:44 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52GbiEk007211; Thu, 2 Jun 2005 09:37:44 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Gbhmu007203 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:37:43 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 10:39:20 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 10:37:32 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>
References: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>    <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Message-ID: <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52Gbhmu007205
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 10:12:42 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> 
> Let me step back. AFAICT, there are two approaches to this:
> 
> 1. Client sends LIST. When an interesting LIST response arrives, it 
> sends a STATUS command. As the STATUS commands' responses arrive, the 
> client has the information it wants.
> 
> 2. Client sends LIST with some extension. For each mailbox matching the 
> supplied pattern, the server does the same work that STATUS would do, 
> and sends the result in a LIST response.
> 
> Do I understand it correctly?
> 
> Alternative 1 has a cost of two roundtrips, plus traffic costs for the 
> LIST responses and for the responses to STATUS, plus CPU costs for 
> LIST, STATUS and quite a bit of parsing.

Almost right.   Alternative 1 has a cost of 1 roundtrip for the LIST and
1 roundtrip for each STATUS per interesting mailbox.  Perhaps you are 
thinking of a STATUS extension that accepts multiple mailboxes as 
arguments?

> Alternative 2 has a cost of one roundtrip, plus traffic for the LIST 
> responses, plus CPU cost for LIST including STATUS.
>
> If we assume that there's 100 LIST responses and 10 interesting 
> mailboxes, then 1 costs one roundtrip more and 2 has 90 unnecessary 
> STATUS processings and 90 unnecessary transmitted UIDNEXT/UIDVALIDITY 
> pairs. (And of course, a client that implements 2 has to implement 1 as 
> well, or else it won't work with plain servers.)

Based on the roundtrip calculation correction:

1 costs 1 LIST + 10 STATUS roundtrips.
2 costs 1 LIST roundtrip.

I believe the response processing to be roughly equivalent in both cases.

>From a client point of view, avoiding the 10 additional roundtrips is very
useful -- especially in a highly latent situation.

> Hm. I suppose I must misunderstand terribly, because I cannot see any 
> appeal in 2 at all.

You've missed the roundtrip cost of multiple STATUS commands.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G6Fb7004555; Thu, 2 Jun 2005 09:06:15 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52G6F4K004554; Thu, 2 Jun 2005 09:06:15 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52G6ESx004545 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:06:14 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id F34BD1A753; Thu,  2 Jun 2005 18:06:12 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id A83BA1A6F5; Thu,  2 Jun 2005 18:06:06 +0200 (CEST)
Message-Id: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 18:05:46 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> Sorry Arnt, it appears that I'm picking on your messages in 
> particular,  and I'm not really.

I am much offended. You must now placate me. ;)

One excellent way to placate me is to explain how the proposal outlined 
in this thread relates to this:

> This particular requirement is an attempt to make MUCH more difficult 
> tasks in the client (synchronization, display markup, user 
> notification) easier and perform better.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G1cHn003828; Thu, 2 Jun 2005 09:01:38 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52G1cth003827; Thu, 2 Jun 2005 09:01:38 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G1YbD003818 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:01:35 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id insudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 17:09:45 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id hnsudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 17:09:45 +0100
Date: Thu, 2 Jun 2005 17:09:45 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: <ietf-imapext@imc.org>
Message-Id: <16094559300656@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
References: <08173804605816@PD.gordano.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_16094560900657/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_16094560900657/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_16094560900658/PD.gordano.com==_"


--==_16094560900658/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I do seem to have opened a can of worms here. Apologies to all concerned -=
 I'm rapidly learning how this process works.=20
 =20
I agree with pretty much all of what you have said. =20
 =20
The only points I would like to make are:=20
 =20
1: In my experience IMAP clients appear to do a LIST followed by a STATUS=
 for every folder. It may be nasty, it may be expensive but they already do=
 it. =20
 =20
2. The idea that users wish to only synchronise a subset of folders is not=
 correct. Again, in my experience most people use IMAP to obtain an offline=
 store whilst retaining the files server side for other uses (WebMail, home=
 access, whatever). Obviously there are exceptions to this - more=
 sophistaticated users\clients or mobile device access but these are the=
 exception rather than the rule. If you are in the business of=
 synchronising a folder then you need UIDVALIDITY so you have to do a=
 STATUS. If you are synchronising all folders then you have to do a STATUS=
 on all folders.=20
 =20
If we come up with extensions that are useful to client developers then=
 they will be pressured to use them. I work for a company that primarily=
 writes servers but we need a client now. The real driver here is to make=
 the end user experience as quick and as easy as possible. If one client=
 works better than another then the worse client will have to improve or=
 die.=20
 =20
Simon=20

---- Message from mailto:<leiba@watson.ibm.com Barry Leiba=
 <leiba@watson.ibm.com> at 02-Jun-2005 11:24:44 ------=20


I don't agree with Mark on every aspect of the IMAP protocol, nor even=20
on every aspect of protocol design in general.  I agree with him on this=20
point.

First, one thing (and I don't mean to get into an extended discussion=20
here; I just want to make a point and move on): there is some=20
inconsistency in Mark's stand.  We can't have the protocol do <x>=20
because it is expensive to do with some legacy message stores (at least=20
one of which he needs to implement for).  But the protocol MUST do <y>,=20
though THAT is expensive to do with some legacy message stores (that he=20
doesn't care about).  (An example of <y> that comes to mind is "return=20
an exact length for RFC822.SIZE, though clients can easily work around=20
that.)  Again, let's not spend a lot of time talking about RFC822.SIZE=20
now -- I want to make the point that we all bring our implementation=20
biases with us.

I also think Simon is right that it's easy to add an extension that's=20
cheap to implement in some (most?) servers, and that it's nice to give=20
that optimization to the client.  We did talk, in fact, about adding=20
STATUS items to the LISTEXT response selections.  I would side with=20
Simon here IF (and this is the key point) we could rely on clients to=20
ask for the information when they need it, and to interpret the server's=20
lack of support for it as an indication that this is an expensive thing=20
to do on this server... and to use a different approach.

What I'm saying is that the client would say "list all the mailboxes,=20
and while you're at it give me <this other information>".  If the server=20
didn't not return the other information, the client would understand=20
that it's because the server can't do that efficiently, and the client=20
would proceed *without* the information, using a different method of=20
processing.

We can't rely on that.  What will, instead, happen, is that clients=20
(perhaps not yours, Simon, but others) will fall back to extracting the=20
information the old, slow way, incurring all the overhead.  Which they=20
can do anyway, of course, but...

My usual approach is to say that if we can provide clients with=20
information they want in a way that the server might be able to=20
optimize, the server might as well support it because the clients will=20
get it anyway.  In this case, though, I agree with Mark that the client=20
does NOT need the information always, and should instead be more=20
selective about obtaining it.  So the remainder of the "but..." sentence=20
above is, "...but it's best not to encourage behaviour that is sometimes=20
bad, and generally unnecessary."

I actually LIKE the idea of the server's having an option of responding=20
"this is a very expensive operation", and wish we could rely on clients=20
to pay attention to that.  Alas, we can't.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam






--==_16094560900658/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I do seem to have opened a can of worms=
 here. Apologies to all concerned - I'm rapidly learning how this process=
 works.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I agree with pretty much all of what you=
 have said.&nbsp;</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>The only points I would like to make=
 are:</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>1:&nbsp;In my experience IMAP clients appear=
 to do a LIST followed by a STATUS for every folder. It may be nasty, it=
 may be expensive but they already do it. </FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>2. The idea that users wish to only=
 synchronise a subset of folders is not correct. Again, in my experience=
 most people use IMAP to obtain an offline store whilst retaining the files=
 server side for other uses (WebMail, home access, whatever). Obviously=
 there are exceptions to this - more sophistaticated users\clients or=
 mobile device access but these are the exception rather than the rule. If=
 you are in the business of synchronising a folder then you need=
 UIDVALIDITY so you have to do a STATUS. If you are synchronising all=
 folders then you have to do a STATUS on all folders.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>If we come up with extensions that are=
 useful to client developers then they will be pressured to use them. I=
 work for a company that primarily writes servers but we need a client now.=
 The real driver here is to make the end user experience as quick and as=
 easy as possible. If one client works better than another then the worse=
 client will have to improve or die.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<leiba@watson.ibm.com">Barry=
 Leiba &lt;leiba@watson.ibm.com&gt;</A> at 02-Jun-2005 11:24:44 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>I don't agree with Mark on every aspect of the IMAP=
 protocol, nor even <BR>on every aspect of protocol design in general.=
 &nbsp;I agree with him on this <BR>point.<BR><BR>First, one thing (and I=
 don't mean to get into an extended discussion <BR>here; I just want to=
 make a point and move on): there is some <BR>inconsistency in Mark's=
 stand. &nbsp;We can't have the protocol do &lt;x&gt; <BR>because it is=
 expensive to do with some legacy message stores (at least <BR>one of which=
 he needs to implement for). &nbsp;But the protocol MUST do &lt;y&gt;,=
 <BR>though THAT is expensive to do with some legacy message stores (that=
 he <BR>doesn't care about). &nbsp;(An example of &lt;y&gt; that comes to=
 mind is "return <BR>an exact length for RFC822.SIZE, though clients can=
 easily work around <BR>that.) &nbsp;Again, let's not spend a lot of time=
 talking about RFC822.SIZE <BR>now -- I want to make the point that we all=
 bring our implementation <BR>biases with us.<BR><BR>I also think Simon is=
 right that it's easy to add an extension that's <BR>cheap to implement in=
 some (most?) servers, and that it's nice to give <BR>that optimization to=
 the client. &nbsp;We did talk, in fact, about adding <BR>STATUS items to=
 the LISTEXT response selections. &nbsp;I would side with <BR>Simon here IF=
 (and this is the key point) we could rely on clients to <BR>ask for the=
 information when they need it, and to interpret the server's <BR>lack of=
 support for it as an indication that this is an expensive thing <BR>to do=
 on this server... and to use a different approach.<BR><BR>What I'm saying=
 is that the client would say "list all the mailboxes, <BR>and while you're=
 at it give me &lt;this other information&gt;". &nbsp;If the server=
 <BR>didn't not return the other information, the client would understand=
 <BR>that it's because the server can't do that efficiently, and the client=
 <BR>would proceed *without* the information, using a different method of=
 <BR>processing.<BR><BR>We can't rely on that. &nbsp;What will, instead,=
 happen, is that clients <BR>(perhaps not yours, Simon, but others) will=
 fall back to extracting the <BR>information the old, slow way, incurring=
 all the overhead. &nbsp;Which they <BR>can do anyway, of course,=
 but...<BR><BR>My usual approach is to say that if we can provide clients=
 with <BR>information they want in a way that the server might be able to=
 <BR>optimize, the server might as well support it because the clients will=
 <BR>get it anyway. &nbsp;In this case, though, I agree with Mark that the=
 client <BR>does NOT need the information always, and should instead be=
 more <BR>selective about obtaining it. &nbsp;So the remainder of the=
 "but..." sentence <BR>above is, "...but it's best not to encourage=
 behaviour that is sometimes <BR>bad, and generally unnecessary."<BR><BR>I=
 actually LIKE the idea of the server's having an option of responding=
 <BR>"this is a very expensive operation", and wish we could rely on=
 clients <BR>to pay attention to that. &nbsp;Alas, we=
 can't.<BR><BR>Barry<BR><BR>--<BR>Barry Leiba, Pervasive Computing=
 Technology=
 &nbsp;(leiba@watson.ibm.com)<BR>http://www.research.ibm.com/people/l/leiba=
<BR>http://www.research.ibm.com/spam<BR><BR><BR><BR></BLOCKQUOTE></FONT></b=
ody></html>

--==_16094560900658/PD.gordano.com==_--
--==_16094560900657/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FqDNZ002900; Thu, 2 Jun 2005 08:52:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FqDHg002899; Thu, 2 Jun 2005 08:52:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52FqBxE002892 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:52:12 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 90CED1A747; Thu,  2 Jun 2005 17:52:10 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 8E18F1A6EF; Thu,  2 Jun 2005 17:52:04 +0200 (CEST)
Message-Id: <G54Annq/kZ1hROmZzY1qKQ.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 17:51:44 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> We are in the IMAP *Extensions* working group.   Presumably we are 
> developing extensions that meet the requirements of the community.
>
> A requirement has been expressed and we have devolved, almost 
> instantly, into a discussion on the relative abilities of some 
> servers to implement the feature that meets the requirement.    This 
> is just wrong.

So it is. Stupid of me to use that argument; it's not central to my position.

I'm not opposed to extensions as such. I do, however, think that 
extensions need to provide some minimum level of utility: Each 
extension brings with it an increased interoperability and testing 
burden, and its utility should be worth that.

In the case of this discussion, I don't see that proposed extension 
does. Perhaps that's because it doesn't, or perhaps that's because the 
people who like the idea aren't pointing out its utility.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Fpms9002871; Thu, 2 Jun 2005 08:51:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Fpmvu002870; Thu, 2 Jun 2005 08:51:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FpmaS002852 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:51:48 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id BFC0A142275; Thu,  2 Jun 2005 08:50:53 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 29352-09; Thu, 2 Jun 2005 08:50:53 -0700 (PDT)
Received: from [152.3.214.109] (user-152-3-214-109.wireless.duke.edu [152.3.214.109]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 26F05142274; Thu,  2 Jun 2005 08:50:53 -0700 (PDT)
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <429F24BC.6000908@watson.ibm.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <2562b68792998dcd3baff37762c3480c@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: ietf-imapext@imc.org
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IMAP4 List Command Extensions
Date: Thu, 2 Jun 2005 08:51:36 -0700
To: Barry Leiba <leiba@watson.ibm.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Thanks Steve and Barry -- I'm glad this discussion naturally got to the 
point of exploring solutions and costs.

  - Client implementors, is this important enough to really use an 
optional extension in the area? Would client implementors actually 
implement both halves of the "IF (supports optional feature)... ELSE 
..."?

  - Are there any other solutions, workarounds or approaches that would 
address the problem described, and be implementable by most (not 
necessarily all) servers?

thx,
Lisa

On Jun 2, 2005, at 8:24 AM, Barry Leiba wrote:

>
> I don't agree with Mark on every aspect of the IMAP protocol, nor even 
> on every aspect of protocol design in general.  I agree with him on 
> this point.
>
> First, one thing (and I don't mean to get into an extended discussion 
> here; I just want to make a point and move on): there is some 
> inconsistency in Mark's stand.  We can't have the protocol do <x> 
> because it is expensive to do with some legacy message stores (at 
> least one of which he needs to implement for).  But the protocol MUST 
> do <y>, though THAT is expensive to do with some legacy message stores 
> (that he doesn't care about).  (An example of <y> that comes to mind 
> is "return an exact length for RFC822.SIZE, though clients can easily 
> work around that.)  Again, let's not spend a lot of time talking about 
> RFC822.SIZE now -- I want to make the point that we all bring our 
> implementation biases with us.
>
> I also think Simon is right that it's easy to add an extension that's 
> cheap to implement in some (most?) servers, and that it's nice to give 
> that optimization to the client.  We did talk, in fact, about adding 
> STATUS items to the LISTEXT response selections.  I would side with 
> Simon here IF (and this is the key point) we could rely on clients to 
> ask for the information when they need it, and to interpret the 
> server's lack of support for it as an indication that this is an 
> expensive thing to do on this server... and to use a different 
> approach.
>
> What I'm saying is that the client would say "list all the mailboxes, 
> and while you're at it give me <this other information>".  If the 
> server didn't not return the other information, the client would 
> understand that it's because the server can't do that efficiently, and 
> the client would proceed *without* the information, using a different 
> method of processing.
>
> We can't rely on that.  What will, instead, happen, is that clients 
> (perhaps not yours, Simon, but others) will fall back to extracting 
> the information the old, slow way, incurring all the overhead.  Which 
> they can do anyway, of course, but...
>
> My usual approach is to say that if we can provide clients with 
> information they want in a way that the server might be able to 
> optimize, the server might as well support it because the clients will 
> get it anyway.  In this case, though, I agree with Mark that the 
> client does NOT need the information always, and should instead be 
> more selective about obtaining it.  So the remainder of the "but..." 
> sentence above is, "...but it's best not to encourage behaviour that 
> is sometimes bad, and generally unnecessary."
>
> I actually LIKE the idea of the server's having an option of 
> responding "this is a very expensive operation", and wish we could 
> rely on clients to pay attention to that.  Alas, we can't.
>
> Barry
>
> --
> Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
> http://www.research.ibm.com/people/l/leiba
> http://www.research.ibm.com/spam
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FZqJE000929; Thu, 2 Jun 2005 08:35:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FZq9B000928; Thu, 2 Jun 2005 08:35:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FZq9N000921 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:35:52 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 09:37:29 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 09:35:41 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
Message-ID: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52FZq9N000922
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Sorry Arnt, it appears that I'm picking on your messages in particular, 
and I'm not really.   You simply authored some conveniently short messages 
with summary positions that it was easier to do reply editing on.    
Please note that I'm addressing the whole thread rather than a single 
message or viewpoint. 

On Thu, 2 Jun 2005 10:16:20 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> It's difficult to implement efficiently for 
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.) 

So?   Make it an optional feature, and then implement it if you can.

> and not too easy for clients, either, because they have to implement a
> fallback in addition, and test both.

Hogwash.  If you are writing a general purpose client then you have to 
implement fallbacks for the various optional features that are already in 
play and it isn't that hard.   If you are writing a single purpose client 
(say a dedicated webmail service) then you just implement the features you
want and make sure that your backend complies.

I've always found the arguments for "difficult to implement in the client"
to be interesting when the requirements come from client implementors.    
This particular requirement is an attempt to make MUCH more difficult 
tasks in the client (synchronization, display markup, user notification) 
easier and perform better.  The cost of fallback code in the low level 
protocol drivers is laughably trivial compared to these other issues.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FOsEx099938; Thu, 2 Jun 2005 08:24:54 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FOsd3099937; Thu, 2 Jun 2005 08:24:54 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FOrtF099922 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:24:54 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j52FQ5lr022843 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:26:05 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j52FOlk94748 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:24:47 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j52FOld94746 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:24:47 -0400
Received: from SATURN ([9.12.241.7]) by mars.watson.ibm.com ID IMFd1117725886.26; Thu, 02 Jun 2005 11:24:46 -0400
Message-ID: <429F24BC.6000908@watson.ibm.com>
Date: Thu, 02 Jun 2005 11:24:44 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
In-Reply-To: <08173804605816@PD.gordano.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I don't agree with Mark on every aspect of the IMAP protocol, nor even 
on every aspect of protocol design in general.  I agree with him on this 
point.

First, one thing (and I don't mean to get into an extended discussion 
here; I just want to make a point and move on): there is some 
inconsistency in Mark's stand.  We can't have the protocol do <x> 
because it is expensive to do with some legacy message stores (at least 
one of which he needs to implement for).  But the protocol MUST do <y>, 
though THAT is expensive to do with some legacy message stores (that he 
doesn't care about).  (An example of <y> that comes to mind is "return 
an exact length for RFC822.SIZE, though clients can easily work around 
that.)  Again, let's not spend a lot of time talking about RFC822.SIZE 
now -- I want to make the point that we all bring our implementation 
biases with us.

I also think Simon is right that it's easy to add an extension that's 
cheap to implement in some (most?) servers, and that it's nice to give 
that optimization to the client.  We did talk, in fact, about adding 
STATUS items to the LISTEXT response selections.  I would side with 
Simon here IF (and this is the key point) we could rely on clients to 
ask for the information when they need it, and to interpret the server's 
lack of support for it as an indication that this is an expensive thing 
to do on this server... and to use a different approach.

What I'm saying is that the client would say "list all the mailboxes, 
and while you're at it give me <this other information>".  If the server 
didn't not return the other information, the client would understand 
that it's because the server can't do that efficiently, and the client 
would proceed *without* the information, using a different method of 
processing.

We can't rely on that.  What will, instead, happen, is that clients 
(perhaps not yours, Simon, but others) will fall back to extracting the 
information the old, slow way, incurring all the overhead.  Which they 
can do anyway, of course, but...

My usual approach is to say that if we can provide clients with 
information they want in a way that the server might be able to 
optimize, the server might as well support it because the clients will 
get it anyway.  In this case, though, I agree with Mark that the client 
does NOT need the information always, and should instead be more 
selective about obtaining it.  So the remainder of the "but..." sentence 
above is, "...but it's best not to encourage behaviour that is sometimes 
bad, and generally unnecessary."

I actually LIKE the idea of the server's having an option of responding 
"this is a very expensive operation", and wish we could rely on clients 
to pay attention to that.  Alas, we can't.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FNKR2099814; Thu, 2 Jun 2005 08:23:20 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FNK3P099813; Thu, 2 Jun 2005 08:23:20 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FNJtw099807 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:23:20 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA  for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:25:04 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 09:23:14 -0700
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
Message-ID: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52FNKtw099808
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 10:16:20 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> It's easy to implement efficiently for an IMAP server that has exclusive 
> control over its storage. It's difficult to implement efficiently for 
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.) and not 
> too easy for clients, either, because they have to implement a fallback 
> in addition, and test both.

We are in the IMAP *Extensions* working group.   Presumably we are 
developing extensions that meet the requirements of the community.

A requirement has been expressed and we have devolved, almost instantly, 
into a discussion on the relative abilities of some servers to implement 
the feature that meets the requirement.    This is just wrong.

The trick, as always, is to design protocol so that useful features may be
provided by servers and consumed by clients.   There has been a *long* 
history of denying useful protocol features in IMAP, particularly in the 
folder management set, based on the premise that:

 "If we provide an optional feature, then certain clients will depend
  on it and not work properly with servers that don't implement it."

And, this is exactly what happens.  If it's useful, then it's useful and 
people will use it.   The reason that there is resistance to this, 
primarily by server vendors, is that this causes pressure to implement 
something that, possibly, their server just can't do very well.

Oh well!

If the feature is useful, then it will be implemented by clients and 
servers that wish to implement it.   If this causes feature pressure on 
implementations, then too bad.   The feature will either have merit and be
useful or it won't.   It is simply one of a set of features that 
contribute to the selection of one implementation over another.   Denying
features in optional services just because some servers can't implement it
efficiently, but some clearly can, it ludicrous.   It smacks of software 
socialism :-)

The craft of good protocol design is to identify useful but potentially 
inefficient features and make sure that:

1.  Either the feature can be removed from the larger optional set (an 
optional feature within a larger optional feature set).

or

2.  Make the optional features disjoint (no dependencies).

If it was easy, everyone would do protocol design.

As to the specifics of this last proposal, this is perhaps the 5th or 6th 
time that I have see this requirement with more or less the same rhetoric
following the expression of the requirement.

The fact that it has been raised this many times should tell you something 
about the *requirement* at least.   An efficient means of acquiring the 
status/metadata (contents to be defined) of a collection of folders would 
be an *extremely* useful feature.   So useful in fact, that I know of 
several instances of private extensions in this area (ours being one of 
them).  No interoperbility of course, because they are non-standard.

So, I encourage those who are interested in proposing a standard for such a
feature to do so.   Don't be shouted down on the basis of a subset of 
implementations that don't support the underlying infrastructure.  
Discussions on the merit of the requirement/feature should be based in
terms of its general usefulness and applicability to existing feature 
context.   This is software engineering process 101. 

I encourage those who have lots of protocol design experience to make the 
feature conform with the rest of IMAP and be all the other things that we 
want in a protocol: simple, elegant, easy to use.  If it is a feature that
cannot be implemented in your server, then clearly to need to lobby for 
the feature being optional in some appropriate form.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52DECm5086026; Thu, 2 Jun 2005 06:14:12 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52DECfU086025; Thu, 2 Jun 2005 06:14:12 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52DE8uw085997 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 06:14:09 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52DE5xl016312 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 06:14:05 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52DE3vj007934 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 06:14:04 -0700
Date: Thu, 2 Jun 2005 06:14:03 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: Alexey Melnikov <alexey.melnikov@isode.com>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <07455621805752@PD.gordano.com>
Message-ID: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Simon Tyler wrote:
> Whether or not a server stores UIDNEXT or calculates it is a server 
> implementation decision. It is irrelevant when designing a protocol.

This is completely wrong.

For better or worse, Internet protocols are molded to accomodate existing 
implementations.  You will receive strong pushback on any proposed 
protocol or protocol enhancement whose impact is perceived as harmful to 
one or more implementations.  Complaints that "it's a poor specification" 
if your demands are not fufilled will fall on deaf ears.

In fact, what you want is to have IMAP molded to accomodate *your* client 
implementation.  Rather than fix your client so it doesn't do something as 
ill-conceived as a STATUS on every LISTed mailbox, you want to change the 
protocol so the server does the stupid thing for you automatically, all to 
save your client some RTTs.  It doesn't address the underlying problem.

The entire point of the Internet protocol architecture is a set of 
interoperable protocols which can be implemented across a wide variety of 
systems, including systems with legacy infrastructure.

The design of protocols for a perfect world, in which you never need to 
concern yourself with the other guy's problems, is not the realm of 
Internet.  It is the realm of proprietary protocols and systems.

In the Internet world, implementors MUST consider the impact of what their 
application does; they can not pretend that they are in a vacuum where 
only they count.  There will be strong pushback against any proposal that 
is perceived as allowing one vendor to count coup against another.

In the proprietary world, you don't have to worry about such things.  You 
control both the client and the server, and any other implementor with the 
temerity to implement your protocol must follow your rules.  At any time, 
you can change the protocol to hurt the other implementor, and you're 
motivated to do so because they are your competitor.

In the Internet world, you have to limit the protocol to what all 
implementations can do well, so that everybody can play ball on a level 
playing field.

In the proprietary world, you can fine-tune the protocol to optimize your 
implementation.  It's your ball and your field.

In the Internet world, interoperability is king, and everything else must 
submit to the demands of that king.

In the proprietary world, the owner is king, and that king receives 
similar submission.

You seem to want the best of both worlds, while paying the price of 
neither.  It doesn't work that way.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52CfDcx073693; Thu, 2 Jun 2005 05:41:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52CfD17073691; Thu, 2 Jun 2005 05:41:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52CfCC2073670 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 05:41:12 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Cf6SA023084 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 05:41:06 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Cf4m9015257 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 05:41:05 -0700
Date: Thu, 2 Jun 2005 05:41:04 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1117701978.15166.414.camel@hurina>
Message-ID: <Pine.OSX.4.63.0506020528510.266@pangtzu.panda.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>  <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <1117701978.15166.414.camel@hurina>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Timo Sirainen wrote:
> I wouldn't mind implementing combined LIST+STATUS reply. The client I
> use in fact already does STATUS for all mailboxes and I like it that
> way, it's all over in less than a second (~300MB in 30 mboxes, with
> non-Dovecot aware procmail appending mails).

30 is not very many mailboxes.

Consider what happens when there are 300 mailboxes, or 3000 mailboxes, or 
30,000 mailboxes.

There are users who have that many!  And we're not talking about tiny 10MB 
mailboxes either.

Plenty of seemingly-good ideas fail when they are scaled upwards a few 
orders of magnitude more than originally considered.

I can guarantee that those orders of magnitude higher scaling *will* 
occur.  It's a miracle that IMAP, a protocol designed at a time when total 
mailbox size was limited to 1MB, has managed to survive nearly 20 years. 
IMHO, a large reason for IMAP's durability was because IMAP was designed 
to be scalable.

> Anyway, I agree that doing a LIST and sending STATUS to non-\Unmarked
> mailboxes is better.

Indeed.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528kQtU092453; Thu, 2 Jun 2005 01:46:26 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528kQZM092452; Thu, 2 Jun 2005 01:46:26 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528kP8m092410 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:46:25 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 110C41C134C6 for <ietf-imapext@imc.org>; Thu,  2 Jun 2005 11:46:19 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: ietf-imapext@imc.org
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-dbyS/srAGiLHWT2++WS8"
Date: Thu, 02 Jun 2005 11:46:18 +0300
Message-Id: <1117701978.15166.414.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-dbyS/srAGiLHWT2++WS8
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Thu, 2005-06-02 at 10:16 +0200, Arnt Gulbrandsen wrote:
> Simon Tyler writes:
> > I would like to see quality in both realms. I don't see that my=20
> > UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds=20
> > to it. I think it is sufficiently easy to implement efficiently that=20
> > it can be included. It is also optional, so a server that cannot=20
> > implement efficienty it can ignore it.
>=20
> It's easy to implement efficiently for an IMAP server that has exclusive=20
> control over its storage. It's difficult to implement efficiently for=20
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.)

I'd just like to point out that while it may be difficult, it's not
impossible. Dovecot has taken ages to get into v1.0 and it's still not
there, partially because it has attempted to implement support for
legacy stores in very efficient way.

I wouldn't mind implementing combined LIST+STATUS reply. The client I
use in fact already does STATUS for all mailboxes and I like it that
way, it's all over in less than a second (~300MB in 30 mboxes, with
non-Dovecot aware procmail appending mails).

Anyway, I agree that doing a LIST and sending STATUS to non-\Unmarked
mailboxes is better. That could however be implemented to the LIST
+STATUS extension if needed. Wouldn't save more than one roundtrip and a
few bytes of traffic though.

Better yet would be instant notification of mailboxes which have
changed, and client could then do STATUS for them if it wanted. Mark
said a year ago he was working on something for this, but I don't
remember anything coming out of it?


--=-dbyS/srAGiLHWT2++WS8
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCnsdayUhSUUBViskRArHdAKCIqG8r8Snh55hhM5qMxi03xN29gQCfd96E
Ey5MFs/qQti/te/l1SGLfws=
=stuj
-----END PGP SIGNATURE-----

--=-dbyS/srAGiLHWT2++WS8--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528Goo1082401; Thu, 2 Jun 2005 01:16:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528GoCv082400; Thu, 2 Jun 2005 01:16:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j528Gn0w082385 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:16:49 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 675001A6E0; Thu,  2 Jun 2005 10:16:48 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 102881A6EC; Thu,  2 Jun 2005 10:16:38 +0200 (CEST)
Message-Id: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:16:20 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
In-Reply-To: <08173804605816@PD.gordano.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Simon Tyler writes:
> I would like to see quality in both realms. I don't see that my 
> UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds 
> to it. I think it is sufficiently easy to implement efficiently that 
> it can be included. It is also optional, so a server that cannot 
> implement efficienty it can ignore it.

It's easy to implement efficiently for an IMAP server that has exclusive 
control over its storage. It's difficult to implement efficiently for 
other IMAP servers (uw-imapd, dovecot, binc, courier et al.) and not 
too easy for clients, either, because they have to implement a fallback 
in addition, and test both.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528DDgV081268; Thu, 2 Jun 2005 01:13:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528DDUT081267; Thu, 2 Jun 2005 01:13:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j528DCjg081252 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:13:13 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id B3D411A757; Thu,  2 Jun 2005 10:13:11 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id D65CD1A6E0; Thu,  2 Jun 2005 10:13:00 +0200 (CEST)
Message-Id: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:12:42 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
In-Reply-To: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Let me step back. AFAICT, there are two approaches to this:

1. Client sends LIST. When an interesting LIST response arrives, it 
sends a STATUS command. As the STATUS commands' responses arrive, the 
client has the information it wants.

2. Client sends LIST with some extension. For each mailbox matching the 
supplied pattern, the server does the same work that STATUS would do, 
and sends the result in a LIST response.

Do I understand it correctly?

Alternative 1 has a cost of two roundtrips, plus traffic costs for the 
LIST responses and for the responses to STATUS, plus CPU costs for 
LIST, STATUS and quite a bit of parsing.

Alternative 2 has a cost of one roundtrip, plus traffic for the LIST 
responses, plus CPU cost for LIST including STATUS.

If we assume that there's 100 LIST responses and 10 interesting 
mailboxes, then 1 costs one roundtrip more and 2 has 90 unnecessary 
STATUS processings and 90 unnecessary transmitted UIDNEXT/UIDVALIDITY 
pairs. (And of course, a client that implements 2 has to implement 1 as 
well, or else it won't work with plain servers.)

Hm. I suppose I must misunderstand terribly, because I cannot see any 
appeal in 2 at all.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5289O6u080036; Thu, 2 Jun 2005 01:09:24 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5289OP4080035; Thu, 2 Jun 2005 01:09:24 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5289LCZ080004 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:09:22 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id gbqudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 09:17:38 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id fbqudaaa; Thu, 2 Jun 2005 09:17:38 +0100
Date: Thu, 2 Jun 2005 09:17:38 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Arnt Gulbrandsen" <arnt@gulbrandsen.priv.no>, <ietf-imapext@imc.org>
Cc: "Alexey Melnikov" <alexey.melnikov@isode.com>, "Mark Crispin" <mrc@cac.washington.edu>
Message-Id: <08173804605816@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_08173804605817/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_08173804605817/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_08173804605818/PD.gordano.com==_"


--==_08173804605818/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I would like to see quality in both realms. I don't see that my=
 UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds to=
 it. I think it is sufficiently easy to implement efficiently that it can=
 be included. It is also optional, so a server that cannot implement=
 efficienty it can ignore it.=20

Simon=20

---- Message from mailto:<arnt@gulbrandsen.priv.no Arnt Gulbrandsen=
 <arnt@gulbrandsen.priv.no> at 02-Jun-2005 10:01:52 ------=20

You seem to be pushing "quality" into the server implementation realm=20
and pretending that a protocol itself has no quality.

But it has. A protocol can good, in which case it's easy to implement=20
efficiently, and it can be bad, in which case it usually is difficult=20
to implement efficiently.

Arnt





--==_08173804605818/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P>&nbsp;</P>
<P>I would like to&nbsp;see quality in both realms. I don't see that my=
 UIDNEXT\UIDVALIDITY detracts from protocol quality but&nbsp; actually adds=
 to it. I think it is sufficiently easy to implement efficiently that it=
 can be included. It is also optional, so a server that cannot implement=
 efficienty&nbsp;it can ignore it.</P>
<P><BR>Simon</P>
<P><FONT face=3DArial size=3D2><BR>---- Message from <A=
 href=3D"mailto:<arnt@gulbrandsen.priv.no">Arnt Gulbrandsen=
 &lt;arnt@gulbrandsen.priv.no&gt;</A> at 02-Jun-2005 10:01:52 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">You seem to be pushing "quality" into the server=
 implementation realm <BR>and pretending that a protocol itself has no=
 quality.<BR><BR>But it has. A protocol can good, in which case it's easy=
 to implement <BR>efficiently, and it can be bad, in which case it usually=
 is difficult <BR>to implement=
 efficiently.<BR><BR>Arnt<BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_08173804605818/PD.gordano.com==_--
--==_08173804605817/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5282Q4t077811; Thu, 2 Jun 2005 01:02:26 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5282Qsr077810; Thu, 2 Jun 2005 01:02:26 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j5282Lw8077775 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:02:23 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 657251A752; Thu,  2 Jun 2005 10:02:20 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 44DFA1A6E7; Thu,  2 Jun 2005 10:02:10 +0200 (CEST)
Message-Id: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:01:52 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
In-Reply-To: <07455621805752@PD.gordano.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

You seem to be pushing "quality" into the server implementation realm 
and pretending that a protocol itself has no quality.

But it has. A protocol can good, in which case it's easy to implement 
efficiently, and it can be bad, in which case it usually is difficult 
to implement efficiently.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j527bfhM069164; Thu, 2 Jun 2005 00:37:41 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j527bfMw069163; Thu, 2 Jun 2005 00:37:41 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j527baua069131 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 00:37:37 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id xxpudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 08:45:56 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id wxpudaaa; Thu, 2 Jun 2005 08:45:56 +0100
Date: Thu, 2 Jun 2005 08:45:56 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Mark Crispin" <mrc@CAC.Washington.EDU>
Cc: "Alexey Melnikov" <alexey.melnikov@isode.com>, <ietf-imapext@imc.org>
Message-Id: <07455621805752@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <08394718706473@PD.gordano.com>
References: <08261932806426@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_07455623405753/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_07455623405753/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_07455623405754/PD.gordano.com==_"


--==_07455623405754/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I'm afraid I still disagree on this.=20
 =20
Whether or not a server stores UIDNEXT or calculates it is a server=
 implementation decision. It is irrelevant when designing a protocol. A=
 well designed server will implement the specification in an efficient way,=
 if calculation is expensive then it will store it as metadata, if=
 accessing the metadata is expensive it will cache the metadata. Moulding a=
 specification to a perceived implementation issue (such as performance of=
 obtaining UIDNEXT) will simply result in a poor specification. If we=
 follow your logic then we can equally argue that the SUBSCRIBED option in=
 the draft command extension can also be expense - how expensive is it to=
 decide if a folder is subscribed? I would argue it is implementation=
 dependent.=20
 =20
Simon=20

---- Message from mailto:<mrc@CAC.Washington.EDU Mark Crispin=
 <mrc@CAC.Washington.EDU> at 01-Jun-2005 11:44:10 ------=20

On Wed, 1 Jun 2005, Simon Tyler wrote:
> My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I=20
> fully understand Mark's issue with performance - getting the number of=20
> recent messages back could be inefficient. Perhaps if we just limited=
 to=20
> the two UID values then these performance issue would go away?

It won't.

UIDNEXT is not necessarily in a fixed metadata value; if new mail has=
 been=20
appended then it may be necessary to go through a UID assignment step in=20
order to calculate UIDNEXT.  [This same problem afflicts the UIDPLUS=20
extension; but in a COPY or APPEND command it may be alright if it takes=20
a little bit longer.]

UIDVALIDITY is normally constant, but it could be affected by the UIDNEXT=20
step if it's determined that something is fouled up in the mailbox.

The bottom line is that unless the mail store has an index of all=
 messages=20
for the mailbox (and legacy mail stores do not), these are expensive=20
calculations.

It gets worse.

Even opening a file to sniff at metadata can be costly.  The very first=20
versions of UW imapd did that 15 years ago, to discriminate between=20
mailbox files and non-mailbox files.  This was removed because it=
 resulted=20
in a huge performance penalty.  It may be only a fraction of a second,=
 but=20
when doing a LIST command that fraction of a section per mailbox adds up=20
to several minutes.  Or worse.

This isn't as much of a problem on modern filesystems as in the past, but=20
the problem has not been exterminated.

Even inode metadata from stat() can be expensive.  For example, a timing=20
test on my home directory shows that "ls -RF" takes about twice as long=
 as=20
"ls -R", even when all the directories are in the buffer cache.  That's=20
much better than 10 years ago, but still is not good.

Small systems with few users have less of a problem that large systems=20
with tens or hundreds of thousands of users.

Dedicated mail stores designed for IMAP only use, with its own internal=20
hierarchy that does not attempt to mirror a filesystem hierarchy, also=20
have less of a problem.  [The problem with mirroring a filesystem=20
hierarchy, especially on systems used for more than just IMAP, is that=20
users make symlinks that create filename loops.]

But saying that "the problem can be reduced in some circumstances" does=20
not equate to "the problem does not exist."

A well-designed client works well in a less than ideal world.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.





--==_07455623405754/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I'm afraid I still disagree on=
 this.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Whether or not a server stores UIDNEXT or=
 calculates it&nbsp;is a server implementation decision. It is irrelevant=
 when designing a protocol. A well designed server will implement the=
 specification in an efficient way, if calculation is expensive then it=
 will store it as metadata, if accessing the metadata is expensive&nbsp;it=
 will cache the metadata.&nbsp;Moulding a specification to a perceived=
 implementation issue (such as performance of obtaining UIDNEXT) will=
 simply result in a poor specification. If we follow your logic then we can=
 equally argue that the SUBSCRIBED option in the draft command extension=
 can also be expense - how expensive is it to decide if a folder is=
 subscribed? I would argue it is implementation dependent.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@CAC.Washington.EDU">Mark=
 Crispin &lt;mrc@CAC.Washington.EDU&gt;</A> at 01-Jun-2005 11:44:10=
 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">On Wed, 1 Jun 2005, Simon Tyler wrote:<BR>&gt; My main=
 driver behind this is to get UIDVALIDITY and UIDNEXT back. I <BR>&gt;=
 fully understand Mark's issue with performance - getting the number of=
 <BR>&gt; recent messages back could be inefficient. Perhaps if we just=
 limited to <BR>&gt; the two UID values then these performance issue would=
 go away?<BR><BR>It won't.<BR><BR>UIDNEXT is not necessarily in a fixed=
 metadata value; if new mail has been <BR>appended then it may be necessary=
 to go through a UID assignment step in <BR>order to calculate UIDNEXT.=
 &nbsp;[This same problem afflicts the UIDPLUS <BR>extension; but in a COPY=
 or APPEND command it may be alright if it takes <BR>a little bit=
 longer.]<BR><BR>UIDVALIDITY is normally constant, but it could be affected=
 by the UIDNEXT <BR>step if it's determined that something is fouled up in=
 the mailbox.<BR><BR>The bottom line is that unless the mail store has an=
 index of all messages <BR>for the mailbox (and legacy mail stores do not),=
 these are expensive <BR>calculations.<BR><BR>It gets worse.<BR><BR>Even=
 opening a file to sniff at metadata can be costly. &nbsp;The very first=
 <BR>versions of UW imapd did that 15 years ago, to discriminate between=
 <BR>mailbox files and non-mailbox files. &nbsp;This was removed because it=
 resulted <BR>in a huge performance penalty. &nbsp;It may be only a=
 fraction of a second, but <BR>when doing a LIST command that fraction of a=
 section per mailbox adds up <BR>to several minutes. &nbsp;Or=
 worse.<BR><BR>This isn't as much of a problem on modern filesystems as in=
 the past, but <BR>the problem has not been exterminated.<BR><BR>Even inode=
 metadata from stat() can be expensive. &nbsp;For example, a timing=
 <BR>test on my home directory shows that "ls -RF" takes about twice as=
 long as <BR>"ls -R", even when all the directories are in the buffer=
 cache. &nbsp;That's <BR>much better than 10 years ago, but still is not=
 good.<BR><BR>Small systems with few users have less of a problem that=
 large systems <BR>with tens or hundreds of thousands of=
 users.<BR><BR>Dedicated mail stores designed for IMAP only use, with its=
 own internal <BR>hierarchy that does not attempt to mirror a filesystem=
 hierarchy, also <BR>have less of a problem. &nbsp;[The problem with=
 mirroring a filesystem <BR>hierarchy, especially on systems used for more=
 than just IMAP, is that <BR>users make symlinks that create filename=
 loops.]<BR><BR>But saying that "the problem can be reduced in some=
 circumstances" does <BR>not equate to "the problem does not=
 exist."<BR><BR>A well-designed client works well in a less than ideal=
 world.<BR><BR>-- Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science=
 does not emerge from voting, party politics, or public debate.<BR>Si vis=
 pacem, para bellum.<BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_07455623405754/PD.gordano.com==_--
--==_07455623405753/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j51IiJUV059686; Wed, 1 Jun 2005 11:44:19 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j51IiJS2059685; Wed, 1 Jun 2005 11:44:19 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j51IiG9V059675 for <ietf-imapext@imc.org>; Wed, 1 Jun 2005 11:44:16 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j51IiCtx014921 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 1 Jun 2005 11:44:12 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j51IiAXj002892 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Wed, 1 Jun 2005 11:44:11 -0700
Date: Wed, 1 Jun 2005 11:44:10 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: Alexey Melnikov <alexey.melnikov@isode.com>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <07144132803948@PD.gordano.com>
Message-ID: <Pine.OSX.4.63.0506011125420.266@pangtzu.panda.com>
References: <07363107802703@PD.gordano.com> <07062778102640@PD.gordano.com> <07062778102640@PD.gordano.com> <07363107802703@PD.gordano.com> <07144132803948@PD.gordano.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 1 Jun 2005, Simon Tyler wrote:
> My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I 
> fully understand Mark's issue with performance - getting the number of 
> recent messages back could be inefficient. Perhaps if we just limited to 
> the two UID values then these performance issue would go away?

It won't.

UIDNEXT is not necessarily in a fixed metadata value; if new mail has been 
appended then it may be necessary to go through a UID assignment step in 
order to calculate UIDNEXT.  [This same problem afflicts the UIDPLUS 
extension; but in a COPY or APPEND command it may be alright if it takes 
a little bit longer.]

UIDVALIDITY is normally constant, but it could be affected by the UIDNEXT 
step if it's determined that something is fouled up in the mailbox.

The bottom line is that unless the mail store has an index of all messages 
for the mailbox (and legacy mail stores do not), these are expensive 
calculations.

It gets worse.

Even opening a file to sniff at metadata can be costly.  The very first 
versions of UW imapd did that 15 years ago, to discriminate between 
mailbox files and non-mailbox files.  This was removed because it resulted 
in a huge performance penalty.  It may be only a fraction of a second, but 
when doing a LIST command that fraction of a section per mailbox adds up 
to several minutes.  Or worse.

This isn't as much of a problem on modern filesystems as in the past, but 
the problem has not been exterminated.

Even inode metadata from stat() can be expensive.  For example, a timing 
test on my home directory shows that "ls -RF" takes about twice as long as 
"ls -R", even when all the directories are in the buffer cache.  That's 
much better than 10 years ago, but still is not good.

Small systems with few users have less of a problem that large systems 
with tens or hundreds of thousands of users.

Dedicated mail stores designed for IMAP only use, with its own internal 
hierarchy that does not attempt to mirror a filesystem hierarchy, also 
have less of a problem.  [The problem with mirroring a filesystem 
hierarchy, especially on systems used for more than just IMAP, is that 
users make symlinks that create filename loops.]

But saying that "the problem can be reduced in some circumstances" does 
not equate to "the problem does not exist."

A well-designed client works well in a less than ideal world.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5176WN3097803; Wed, 1 Jun 2005 00:06:32 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5176WBB097802; Wed, 1 Jun 2005 00:06:32 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5176RmH097767 for <ietf-imapext@imc.org>; Wed, 1 Jun 2005 00:06:28 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id fekudaaa for ietf-imapext@imc.org; Wed, 1 Jun 2005 08:14:41 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id eekudaaa; Wed, 1 Jun 2005 08:14:41 +0100
Date: Wed, 1 Jun 2005 08:14:41 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Alexey Melnikov" <alexey.melnikov@isode.com>
Cc: <ietf-imapext@imc.org>
Message-Id: <07144132803948@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <Pine.OSX.4.63.0505310031220.305@pangtzu.panda.com>
References: <07363107802703@PD.gordano.com> <07062778102640@PD.gordano.com> <07062778102640@PD.gordano.com> <07363107802703@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_07144132803949/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_07144132803949/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_07144132803950/PD.gordano.com==_"


--==_07144132803950/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
Looks like almost the same idea I had just with slightly different=
 syntax.=20
 =20
My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I fully=
 understand Mark's issue with performance - getting the number of recent=
 messages back could be inefficient. Perhaps if we just limited to the two=
 UID values then these performance issue would go away?=20
 =20
Simon=20

---- Message from mailto:<alexey.melnikov@isode.com Alexey Melnikov=
 <alexey.melnikov@isode.com> at 31-May-2005 18:02:00 ------=20

Simon Tyler wrote:

> Indeed the protocol should consider a wide variety of implementations.=20
> Such as one which has no large overhead in obtaining the extra=20
> information I was proposing. Our client will clearly backdown to using=20
> a STATUS command should the extended LIST response not be found.
>
> =20
>
> I fully understand all the issues, that is why I felt it was a=20
> reasonable extension - servers which have issues returning the UID=20
> information in short timescales would obviously not choose to add in=20
> such an extension. Those that can obtain the information could do so.
>
Exactly.

I've posted a similar proposal some times ago:

<http://www.imc.org/ietf-imapext/mail-archive/msg02245.html>

which I intend to submit as a draft, once the LIST-EXTENDED is done.






--==_07144132803950/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Looks like almost the same idea I had just=
 with slightly different syntax.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>My main driver behind this is to get=
 UIDVALIDITY and UIDNEXT back. I fully understand Mark's issue with=
 performance - getting the number of recent messages back could be=
 inefficient. Perhaps if we just limited to the two UID values then these=
 performance issue would go away?</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A=
 href=3D"mailto:<alexey.melnikov@isode.com">Alexey Melnikov=
 &lt;alexey.melnikov@isode.com&gt;</A> at 31-May-2005 18:02:00 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">Simon Tyler wrote:<BR><BR>&gt; Indeed the protocol=
 should consider a wide variety of implementations. <BR>&gt; Such as one=
 which has no large overhead in obtaining the extra <BR>&gt; information I=
 was proposing. Our client will clearly backdown to using <BR>&gt; a STATUS=
 command should the extended LIST response not be found.<BR>&gt;<BR>&gt;=
 &nbsp;<BR>&gt;<BR>&gt; I fully understand all the issues, that is why I=
 felt it was a <BR>&gt; reasonable extension - servers which have issues=
 returning the UID <BR>&gt; information in short timescales would obviously=
 not choose to add in <BR>&gt; such an extension. Those that can obtain the=
 information could do so.<BR>&gt;<BR>Exactly.<BR><BR>I've posted a similar=
 proposal some times=
 ago:<BR><BR>&lt;http://www.imc.org/ietf-imapext/mail-archive/msg02245.html=
&gt;<BR><BR>which I intend to submit as a draft, once the LIST-EXTENDED is=
 done.<BR><BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_07144132803950/PD.gordano.com==_--
--==_07144132803949/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5SJo3E5046603; Tue, 28 Jun 2005 12:50:03 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5SJo3GZ046602; Tue, 28 Jun 2005 12:50:03 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from newodin.ietf.org ([132.151.6.50]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5SJo2Ym046594 for <ietf-imapext@imc.org>; Tue, 28 Jun 2005 12:50:03 -0700 (PDT) (envelope-from mlee@newodin.ietf.org)
Received: from mlee by newodin.ietf.org with local (Exim 4.43) id 1DnM5q-000887-MN; Tue, 28 Jun 2005 15:50:02 -0400
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
To: i-d-announce@ietf.org
Cc: ietf-imapext@imc.org
From: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-imapext-2086upd-08.txt 
Message-Id: <E1DnM5q-000887-MN@newodin.ietf.org>
Date: Tue, 28 Jun 2005 15:50:02 -0400
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Internet Message Access Protocol Extension Working Group of the IETF.

	Title		: IMAP4 ACL extension
	Author(s)	: A. Melnikov
	Filename	: draft-ietf-imapext-2086upd-08.txt
	Pages		: 29
	Date		: 2005-6-28
	
The ACL (Access Control List) extension (RFC 2086) of the Internet
   Message Access Protocol (IMAP) permits mailbox access control lists
   to be retrieved and manipulated through the IMAP protocol.

   This document is a revision of RFC 2086.  It defines several new
   access control rights and clarifies which rights are required for
   different IMAP commands.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-imapext-2086upd-08.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-imapext-2086upd-08.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-imapext-2086upd-08.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<2005-6-28150727.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-imapext-2086upd-08.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-imapext-2086upd-08.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<2005-6-28150727.I-D@ietf.org>

--OtherAccess--

--NextPart--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFxGCa007874; Fri, 24 Jun 2005 08:59:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFxGJu007871; Fri, 24 Jun 2005 08:59:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFxF5E007794 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:59:15 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so201zzb for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Received: by 172.25.12.97 with SMTP id 17mr57zzb; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Fri, 24 Jun 2005 08:59:09 -0700 (PDT)
Message-ID: <5b93232d05062408593a7c0f83@mail.google.com>
Date: Fri, 24 Jun 2005 08:59:09 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Alexey Melnikov <alexey.melnikov@isode.com>
Subject: Re: user names vs. identifiers
Cc: Lisa Dusseault <lisa@osafoundation.org>, Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <42BC2D45.3080407@isode.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com> <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org> <42BC2D45.3080407@isode.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5OFxF5E007857
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/24/05, Alexey Melnikov <alexey.melnikov@isode.com> wrote:
> Lisa Dusseault wrote:
> 
> >
> > What about some wording up front (before the word is first used) that
> > explains once that "identifier" is used for access identifiers which
> > may be userids or other authorization tokens?
> 
> I would rather do what you suggest, as this minimizes the amount of
> changes to the document and thus it is less prone to introducing errors.

I agree, this seems much less invasive and still correct.

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFv0Cx005789; Fri, 24 Jun 2005 08:57:00 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFv0fH005788; Fri, 24 Jun 2005 08:57:00 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFuxRF005764 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:56:59 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.176] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Fri, 24 Jun 2005 16:56:52 +0100
Message-ID: <42BC2D45.3080407@isode.com>
Date: Fri, 24 Jun 2005 16:56:53 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Lisa Dusseault <lisa@osafoundation.org>
CC: Rob Siemborski <robsiemb@google.com>, Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: user names vs. identifiers
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com> <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
In-Reply-To: <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Lisa Dusseault wrote:

>
> What about some wording up front (before the word is first used) that 
> explains once that "identifier" is used for access identifiers which 
> may be userids or other authorization tokens?

I would rather do what you suggest, as this minimizes the amount of 
changes to the document and thus it is less prone to introducing errors.

> lisa
>
> On Jun 24, 2005, at 8:12 AM, Rob Siemborski wrote:
>
>>
>> On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
>>
>>> On Thu, 23 Jun 2005, Rob Siemborski wrote:
>>>
>>>> In the case of the ACL document, this isn't accurate, as the
>>>> identifiers can refer to groups as well as usernames/userids.
>>>
>>>
>>> Suggestion: the term "access identifier" should be used instead.  This
>>> term is used by the URLAUTH document as well, but the two documents 
>>> don't
>>> overlap in this area so the term can be reused.
>>
>>
>> Sure, this works for me.
>>
>>> "Access identifier" should certainly be defined so that one of its
>>> possible values is "userid".  AFAICT, all userids are access 
>>> identifiers
>>> for ACL purposes, but not all access identifiers are userids, right?
>>
>>
>> Yes.
>


-- 
Alexey Melnikov
__________________________________________
Isode Mbox Message Store developer

Isode Limited, http://www.isode.com

IETF standard related pages:
http://www.melnikov.ca/mel/devel/Links.html

Personal Home Page: http://www.melnikov.ca
__________________________________________



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFhU5X090987; Fri, 24 Jun 2005 08:43:30 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFhUd6090984; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFhU4L090971 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:43:30 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 367D2142264; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 28569-04; Fri, 24 Jun 2005 08:43:30 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id BBFA314225E; Fri, 24 Jun 2005 08:43:29 -0700 (PDT)
In-Reply-To: <42BBDAAA.50508@isode.com>
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org> <42BBDAAA.50508@isode.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <865da06559bc551eb0e875fa72c135c5@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Fri, 24 Jun 2005 08:43:27 -0700
To: Alexey Melnikov <alexey.melnikov@isode.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Jun 24, 2005, at 3:04 AM, Alexey Melnikov wrote:

>>
>> So, another possible sentence to add is something along the lines of 
>> "When an IMAP server is responsible for creating user names for new 
>> accounts, the server SHOULD NOT create such ambiguous user names."
>
> Regarding this sentence in particular: user management is outside the 
> scope of the 2086upd, so I don't think this sentence belongs to the 
> document.

That's just fine by me, I see your point.
Lisa



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFghBJ090156; Fri, 24 Jun 2005 08:42:43 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFghZ8090155; Fri, 24 Jun 2005 08:42:43 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFgfhQ090033 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:42:41 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 2698F142264; Fri, 24 Jun 2005 08:42:41 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 24565-08; Fri, 24 Jun 2005 08:42:40 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 6CF6914225E; Fri, 24 Jun 2005 08:42:40 -0700 (PDT)
In-Reply-To: <5b93232d050624081278553f7e@mail.google.com>
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com> <5b93232d050624081278553f7e@mail.google.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <0a1406ab108432c1fceafff1c5bcd397@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: Mark Crispin <mrc@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: user names vs. identifiers
Date: Fri, 24 Jun 2005 08:42:38 -0700
To: Rob Siemborski <robsiemb@google.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

What about some wording up front (before the word is first used) that 
explains once that "identifier" is used for access identifiers which 
may be userids or other authorization tokens?

lisa

On Jun 24, 2005, at 8:12 AM, Rob Siemborski wrote:

>
> On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
>> On Thu, 23 Jun 2005, Rob Siemborski wrote:
>>> In the case of the ACL document, this isn't accurate, as the
>>> identifiers can refer to groups as well as usernames/userids.
>>
>> Suggestion: the term "access identifier" should be used instead.  This
>> term is used by the URLAUTH document as well, but the two documents 
>> don't
>> overlap in this area so the term can be reused.
>
> Sure, this works for me.
>
>> "Access identifier" should certainly be defined so that one of its
>> possible values is "userid".  AFAICT, all userids are access 
>> identifiers
>> for ACL purposes, but not all access identifiers are userids, right?
>
> Yes.
>
> -Rob
>
> -- 
> Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
> Software Engineer     |                       |            650-623-6925
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFCqR3063038; Fri, 24 Jun 2005 08:12:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OFCprJ063029; Fri, 24 Jun 2005 08:12:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OFCoiH062949 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:12:50 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so165zzb for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Received: by 172.25.12.188 with SMTP id 28mr49zzc; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Fri, 24 Jun 2005 08:12:45 -0700 (PDT)
Message-ID: <5b93232d050624081278553f7e@mail.google.com>
Date: Fri, 24 Jun 2005 08:12:45 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: user names vs. identifiers
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com> <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5OFCoiH063017
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
> On Thu, 23 Jun 2005, Rob Siemborski wrote:
> > In the case of the ACL document, this isn't accurate, as the
> > identifiers can refer to groups as well as usernames/userids.
> 
> Suggestion: the term "access identifier" should be used instead.  This
> term is used by the URLAUTH document as well, but the two documents don't
> overlap in this area so the term can be reused.

Sure, this works for me.

> "Access identifier" should certainly be defined so that one of its
> possible values is "userid".  AFAICT, all userids are access identifiers
> for ACL purposes, but not all access identifiers are userids, right?

Yes.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OA6YUQ008917; Fri, 24 Jun 2005 03:06:34 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5OA4kWB006820; Fri, 24 Jun 2005 03:04:46 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5OA4TRR006350 for <ietf-imapext@imc.org>; Fri, 24 Jun 2005 03:04:30 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Fri, 24 Jun 2005 11:04:26 +0100
Message-ID: <42BBDAAA.50508@isode.com>
Date: Fri, 24 Jun 2005 11:04:26 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Lisa Dusseault <lisa@osafoundation.org>
CC: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
In-Reply-To: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Lisa Dusseault wrote:

> On Jun 23, 2005, at 9:34 AM, Alexey Melnikov wrote:
>
>> ...
>>
>> So, I think this is a SASLPrep issue and should be fixed there. Two 
>> other alternatives for 2086upd I see:
>> 1). override SASLPrep rules to say that PR29 should be used (the 
>> proper procedure is described there).
>> 2). say that an 2086upd server MAY reject such sequences.
>> Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
>> sure that our document is the proper place for the fix.
>
> I would prefer to add something along the lines of #2.  Some of my 
> reasoning:
>  - Although we don't have to say that the server MAY reject such user 
> names, because there's nothing elsewhere in the document that forbids 
> it, it does make it clear what one possible, immediate, practical 
> solution to the problem is.
>  - I'd rather not have the suggested solution be merely  to wait on 
> future speculative work.
>  - I don't see any drawbacks to adding such a MAY clause.

Ok, you've convinced me that this solution is fine. My proposed text 
should still work. I just need to add a sentence about "MAY reject such 
identifiers".

> The problem is of course not entirely solvable by 2086upd, although we 
> can do our best by drawing attention to it.  If I understand 
> correctly, the software that typically creates the problem by creating 
> ambiguous user names is not IMAP clients, and not always IMAP servers, 
> but often directory and authentication services -- and the 
> implementors of these services probably have no need to read 2086bis.
>
> So, another possible sentence to add is something along the lines of 
> "When an IMAP server is responsible for creating user names for new 
> accounts, the server SHOULD NOT create such ambiguous user names."

Regarding this sentence in particular: user management is outside the 
scope of the 2086upd, so I don't think this sentence belongs to the 
document.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6oU3i044161; Thu, 23 Jun 2005 23:50:30 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O6oUFX044160; Thu, 23 Jun 2005 23:50:30 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6oT1K044148 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:50:29 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O6oRDj017601 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 23 Jun 2005 23:50:27 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O6oPwK024604 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 23 Jun 2005 23:50:27 -0700
Date: Thu, 23 Jun 2005 23:50:25 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Rob Siemborski <robsiemb@google.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: user names vs. identifiers
In-Reply-To: <5b93232d05062323352dbc2c68@mail.google.com>
Message-ID: <Pine.OSX.4.63.0506232348430.21436@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com> <5b93232d05062323352dbc2c68@mail.google.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 23 Jun 2005, Rob Siemborski wrote:
> In the case of the ACL document, this isn't accurate, as the
> identifiers can refer to groups as well as usernames/userids.

Suggestion: the term "access identifier" should be used instead.  This 
term is used by the URLAUTH document as well, but the two documents don't 
overlap in this area so the term can be reused.

"Access identifier" should certainly be defined so that one of its 
possible values is "userid".  AFAICT, all userids are access identifiers 
for ACL purposes, but not all access identifiers are userids, right?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6ZEUb031130; Thu, 23 Jun 2005 23:35:14 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O6ZEMs031129; Thu, 23 Jun 2005 23:35:14 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O6ZDTn031065 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:35:14 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so101zzb for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Received: by 172.25.12.188 with SMTP id 28mr26zzc; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Thu, 23 Jun 2005 23:35:08 -0700 (PDT)
Message-ID: <5b93232d05062323352dbc2c68@mail.google.com>
Date: Thu, 23 Jun 2005 23:35:08 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: user names vs. identifiers
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5O6ZETn031122
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Mark Crispin <mrc@cac.washington.edu> wrote:
> I agree that "user name" should be deprecated, although I would advocate
> "userid" or "user identifier" instead of plain "identifier".

In the case of the ACL document, this isn't accurate, as the
identifiers can refer to groups as well as usernames/userids.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O38N2D079107; Thu, 23 Jun 2005 20:08:23 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O38Not079106; Thu, 23 Jun 2005 20:08:23 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O38NOn079091 for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:23 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O38MRc004374 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:22 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j5O38K5x004723 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@IMC.ORG>; Thu, 23 Jun 2005 20:08:21 -0700
Date: Thu, 23 Jun 2005 20:08:19 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: user names vs. identifiers
Message-ID: <Pine.OSX.4.63.0506231957420.21436@pangtzu.panda.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

For what it's worth, RFC 3501 uses both "user name" and "userid"; 
"identifier" is used primarily in the context of "unique identifier" 
although once is used to refer to a charset identifier.

The URLAUTH document in Lemonade uses userid throughout, except in one 
error mesage where it uses "user name".

I agree that "user name" should be deprecated, although I would advocate 
"userid" or "user identifier" instead of plain "identifier".

Another problem is such terms as "current userid", etc. when what is 
really intended is this session's authorization identity.  I think that 
it's probably alright to use "current userid" as long as we define 
someplace that this refers to the authorization identity.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O2XPu6056963; Thu, 23 Jun 2005 19:33:25 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5O2XPUJ056962; Thu, 23 Jun 2005 19:33:25 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5O2XOCV056917 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 19:33:24 -0700 (PDT) (envelope-from robsiemb@google.com)
Received: by 216-239-45-4.google.com with SMTP id 14so9zzb for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 19:33:19 -0700 (PDT)
Received: by 172.25.12.97 with SMTP id 17mr3zzb; Thu, 23 Jun 2005 19:33:19 -0700 (PDT)
Received: by 172.25.12.91 with HTTP; Thu, 23 Jun 2005 19:33:18 -0700 (PDT)
Message-ID: <5b93232d0506231933df1c637@mail.google.com>
Date: Thu, 23 Jun 2005 19:33:18 -0700
From: Rob Siemborski <robsiemb@google.com>
Reply-To: Rob Siemborski <robsiemb@google.com>
To: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Cc: Alexey Melnikov <alexey.melnikov@isode.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com> <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j5O2XPCV056956
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On 6/23/05, Lisa Dusseault <lisa@osafoundation.org> wrote:
> So, another possible sentence to add is something along the lines of
> "When an IMAP server is responsible for creating user names for new
> accounts, the server SHOULD NOT create such ambiguous user names."

We should be sure that however we fix this we refer to 'identifiers'
and not 'user names'.

-Rob

-- 
Rob Siemborski        |     PGP:0x5CE32FCC    |     robsiemb@google.com
Software Engineer     |                       |            650-623-6925



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NHWGbc045923; Thu, 23 Jun 2005 10:32:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NHWGdG045922; Thu, 23 Jun 2005 10:32:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NHWG6d045904 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:32:16 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id 1429114225D; Thu, 23 Jun 2005 10:32:39 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 15041-03; Thu, 23 Jun 2005 10:32:38 -0700 (PDT)
Received: from [192.168.101.178] (w002.z065106067.sjc-ca.dsl.cnc.net [65.106.67.2]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id CC99C14225B; Thu, 23 Jun 2005 10:32:38 -0700 (PDT)
In-Reply-To: <42BAE4A7.9080004@isode.com>
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com> <42BAE4A7.9080004@isode.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <6e29a9095c7fad1251bd769b6ff84fe4@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Thu, 23 Jun 2005 10:32:12 -0700
To: Alexey Melnikov <alexey.melnikov@isode.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Jun 23, 2005, at 9:34 AM, Alexey Melnikov wrote:

> ...
>
> So, I think this is a SASLPrep issue and should be fixed there. Two 
> other alternatives for 2086upd I see:
> 1). override SASLPrep rules to say that PR29 should be used (the 
> proper procedure is described there).
> 2). say that an 2086upd server MAY reject such sequences.
> Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
> sure that our document is the proper place for the fix.

I would prefer to add something along the lines of #2.  Some of my 
reasoning:
  - Although we don't have to say that the server MAY reject such user 
names, because there's nothing elsewhere in the document that forbids 
it, it does make it clear what one possible, immediate, practical 
solution to the problem is.
  - I'd rather not have the suggested solution be merely  to wait on 
future speculative work.
  - I don't see any drawbacks to adding such a MAY clause.

The problem is of course not entirely solvable by 2086upd, although we 
can do our best by drawing attention to it.  If I understand correctly, 
the software that typically creates the problem by creating ambiguous 
user names is not IMAP clients, and not always IMAP servers, but often 
directory and authentication services -- and the implementors of these 
services probably have no need to read 2086bis.

So, another possible sentence to add is something along the lines of 
"When an IMAP server is responsible for creating user names for new 
accounts, the server SHOULD NOT create such ambiguous user names."

Lisa


>
> Assuming we go with "punt the issue to SASLPrep", I suggest to add the 
> following text to the Security Considerations section:
>
>     This document relies on [SASLPrep] to describe steps required to 
> perform user name
>     canonicalization (preparation). The preparation algorithm in 
> SASLPrep was specifically
>     designed such that its output is canonical, and it is well-formed.
>     However, due to an anomaly [PR29] in the specification of Unicode
>     normalization, canonical equivalence is not guaranteed for a select
>     few character sequences.
>
>     User names prepared with SASLPrep can be stored and returned
>     by an ACL server. The anomaly affects ACL manipulation and 
> evaluation of
>     user names containing the selected character sequences. These 
> sequences, however,
>     do not appear in well-formed text. This specification was 
> published despite this
>     known technical problem. It is expected that the problem will be
>     addressed in a future revision of [Unicode] and/or [StringPrep]
>     specifications.
>
> Comments?
>
> Alexey
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NH1lFN012646; Thu, 23 Jun 2005 10:01:47 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NH1lXv012645; Thu, 23 Jun 2005 10:01:47 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from numenor.qualcomm.com (numenor.qualcomm.com [129.46.51.58]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NH1leA012638 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:47 -0700 (PDT) (envelope-from ameylan@qualcomm.com)
Received: from crowley.qualcomm.com (crowley.qualcomm.com [129.46.61.151]) by numenor.qualcomm.com (8.12.10/8.12.5/1.0) with ESMTP id j5NH1egj027659 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL) for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:40 -0700 (PDT)
Received: from AMEYLAN.qualcomm.com (ameylan.na.qualcomm.com [129.46.159.213]) by crowley.qualcomm.com (8.12.10/8.12.5/1.0) with ESMTP id j5NH1cv6003414 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 10:01:38 -0700 (PDT)
Message-Id: <6.2.2.1.2.20050623100115.03d0b0f0@unixmail.qualcomm.com>
X-Mailer: QUALCOMM Windows Eudora Version 6.2.2.1
Date: Thu, 23 Jun 2005 10:01:34 -0700
To: ietf-imapext@imc.org
From: Arnaud Meylan <ameylan@qualcomm.com>
Subject: 
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

unsubscribe ameylan@qualcomm.com



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NGZIrr086764; Thu, 23 Jun 2005 09:35:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NGZIDx086763; Thu, 23 Jun 2005 09:35:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NGZHSL086757 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 09:35:18 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.177] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 23 Jun 2005 17:34:46 +0100
Message-ID: <42BAE4A7.9080004@isode.com>
Date: Thu, 23 Jun 2005 17:34:47 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: Re: IESG Evaluation of draft-ietf-imapext-2086upd-07
References: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
In-Reply-To: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Scott Hollenbeck wrote:

>The IESG evaluation of draft-ietf-imapext-2086upd-07 has completed.  There's
>only one discuss comment that we started to talk about.  Revised ID needed.
>
>Lisa: as document shepherd, please work with Russ and Alexey to come up with
>appropriate text.
>
>  
>
After reading <http://www.unicode.org/review/pr-29.html> I think I 
understand the issue now (please let me know if this is not the case). 
Here is a scenario when this issue might be relevant to the document:

An 2086upd client adds permissions for user 
"<U+1100><U+0300><U+1161><U+0323>" (the value is taken from PR29 and is 
contrived) using SETACL command. The server performs normalization as 
described in SASLprep and will store permissions for the username 
"<U+AC00><U+0300><U+0323>".

The client then lists permissions (for display purposes) on the mailbox 
using GETACL. The server returns canonicalized name 
"<U+AC00><U+0300><U+0323>".

The user then decides that the permissions are not right, so the client 
will update the permissions for "<U+AC00><U+0300><U+0323>". The server 
performs canonicalization which results in username 
"<U+AC00><U+0323><U+0300>", which is not the same as 
"<U+AC00><U+0300><U+0323>". The server will fail to treat them as the 
same, so instead of replacing permissions on "<U+AC00><U+0300><U+0323>", 
it will add permissions for "<U+AC00><U+0323><U+0300>".

The problem for the user is that all 3 variants should be displayed 
exactly the same.
The problem for access calculation is that the result will depend on the 
form of identifier used. This can be unexpected at best.

The good news is that such sequences are very-very-unlikely to happen in 
real life.

So, I think this is a SASLPrep issue and should be fixed there. Two 
other alternatives for 2086upd I see:
1). override SASLPrep rules to say that PR29 should be used (the proper 
procedure is described there).
2). say that an 2086upd server MAY reject such sequences.
Both of the alternatives try to fix SASLPrep/StringPrep, but I am not 
sure that our document is the proper place for the fix.

Assuming we go with "punt the issue to SASLPrep", I suggest to add the 
following text to the Security Considerations section:

     This document relies on [SASLPrep] to describe steps required to 
perform user name
     canonicalization (preparation). The preparation algorithm in 
SASLPrep was specifically
     designed such that its output is canonical, and it is well-formed.
     However, due to an anomaly [PR29] in the specification of Unicode
     normalization, canonical equivalence is not guaranteed for a select
     few character sequences.

     User names prepared with SASLPrep can be stored and returned
     by an ACL server. The anomaly affects ACL manipulation and 
evaluation of
     user names containing the selected character sequences. These 
sequences, however,
     do not appear in well-formed text. This specification was published 
despite this
     known technical problem. It is expected that the problem will be
     addressed in a future revision of [Unicode] and/or [StringPrep]
     specifications.

Comments?

Alexey



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NG9KrD055882; Thu, 23 Jun 2005 09:09:20 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5NG9KjJ055876; Thu, 23 Jun 2005 09:09:20 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5NG9Heo055763 for <ietf-imapext@imc.org>; Thu, 23 Jun 2005 09:09:19 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Thu, 23 Jun 2005 12:09:12 -0400 id 0059C1E8.42BADEA8.000069DF
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: "'Alexey Melnikov'" <alexey.melnikov@isode.com>, lisa@osafoundation.org
Cc: "'IMAP Extensions WG'" <ietf-imapext@imc.org>
Subject: IESG Evaluation of draft-ietf-imapext-2086upd-07
Date: Thu, 23 Jun 2005 12:08:52 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
Thread-Index: AcV4DdkWELCFMBPyS2qQvyUyoUfwrQ==
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Message-ID: <courier.42BADEA8.000069DF@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

The IESG evaluation of draft-ietf-imapext-2086upd-07 has completed.  There's
only one discuss comment that we started to talk about.  Revised ID needed.

Lisa: as document shepherd, please work with Russ and Alexey to come up with
appropriate text.

-Scott-



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5LBBMgB028665; Tue, 21 Jun 2005 04:11:22 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5LBBM3E028664; Tue, 21 Jun 2005 04:11:22 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5LBBLUC028648 for <ietf-imapext@imc.org>; Tue, 21 Jun 2005 04:11:21 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Tue, 21 Jun 2005 07:11:20 -0400 id 0059C319.42B7F5D8.00003168
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: ietf-imapext@imc.org
Subject: FW: DISCUSS: draft-ietf-imapext-2086upd-07
Date: Tue, 21 Jun 2005 07:11:07 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Thread-Index: AcV13Do5bZLxzRHfSj2hlkIUvWFDJAAdWYSw
Message-ID: <courier.42B7F5D8.00003168@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

FYI -- here's an IESG evaluation comment that will need to be resolved.
More may come before Thursday.  Do NOT change the document now.  We can talk
about the needed text and talk to Russ to find something he's happy with.

-Scott-

-----Original Message-----
From: Russ Housley [mailto:housley@vigilsec.com] 
Sent: Monday, June 20, 2005 4:56 PM
To: iesg@ietf.org
Subject: DISCUSS: draft-ietf-imapext-2086upd-07

IMAP4 ACL extension (Proposed Standard)

DISCUSS

   This specification depends on SASLprep, which is defined in RFC 4013.
   The security considerations of RFC 4013 say:
   >
   > This profile is intended to prepare simple user name and password
   > strings for comparison or use in cryptographic functions (e.g.,
   > message digests).  The preparation algorithm was specifically
   > designed such that its output is canonical, and it is well-formed.
   > However, due to an anomaly [PR29] in the specification of Unicode
   > normalization, canonical equivalence is not guaranteed for a select
   > few character sequences.  These sequences, however, do not appear in
   > well-formed text.  This specification was published despite this
   > known technical problem.  It is expected that this specification will
   > be revised before further progression on the Standards Track (after
   > [Unicode] and/or [StringPrep] specifications have been updated to
   > address this problem).
   >
   The security considerations of this document need to address this point.
   How does this situation impact ACL processing?





Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5DJe89l078654; Mon, 13 Jun 2005 12:40:08 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5DJe8QK078653; Mon, 13 Jun 2005 12:40:08 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from ietf.org (odin.ietf.org [132.151.1.176]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5DJe7Em078646 for <ietf-imapext@imc.org>; Mon, 13 Jun 2005 12:40:08 -0700 (PDT) (envelope-from dinaras@cnri.reston.va.us)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15881; Mon, 13 Jun 2005 15:40:05 -0400 (EDT)
Message-Id: <200506131940.PAA15881@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: i-d-announce@ietf.org
Cc: ietf-imapext@imc.org
From: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-imapext-2086upd-07.txt
Date: Mon, 13 Jun 2005 15:40:05 -0400
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Internet Message Access Protocol Extension Working Group of the IETF.

	Title		: IMAP4 ACL extension
	Author(s)	: A. Melnikov
	Filename	: draft-ietf-imapext-2086upd-07.txt
	Pages		: 28
	Date		: 2005-6-13
	
The ACL (Access Control List) extension (RFC 2086) of the Internet
   Message Access Protocol (IMAP) permits mailbox access control lists
   to be retrieved and manipulated through the IMAP protocol.

   This document is a revision of RFC 2086.  It defines several new
   access control rights and clarifies which rights are required for
   different IMAP commands.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-imapext-2086upd-07.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-imapext-2086upd-07.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-imapext-2086upd-07.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<2005-6-13160813.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-imapext-2086upd-07.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-imapext-2086upd-07.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<2005-6-13160813.I-D@ietf.org>

--OtherAccess--

--NextPart--




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AvWhN084164; Thu, 9 Jun 2005 03:57:32 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59AvWQH084163; Thu, 9 Jun 2005 03:57:32 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AvVXZ084153 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:57:31 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.110] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 11:57:28 +0100
Message-ID: <42A82098.4010306@isode.com>
Date: Thu, 09 Jun 2005 11:57:28 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Dave Cridland <dave@cridland.net>
CC: Barry Leiba <leiba@watson.ibm.com>, IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com> <28904.1118310684.785236@localhost.localdomain>
In-Reply-To: <28904.1118310684.785236@localhost.localdomain>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Dave Cridland wrote:

> On Wed Jun  8 19:27:01 2005, Barry Leiba wrote:
>
>>>> - section 6:  should we restrict selection options to be simple atoms
>>>> as we have in this draft, or should we allow for them to be a list as
>>>> well?   I'm thinking that some future selection option might take one
>>>> or more arguments.  For example, the non-standard SCAN extension used
>>>> by Pine and UW-IMAP could be reworked to look something like this:
>>>>
>>>> x LIST ((SCAN "foo")) "" "bar/%"
>>>>
>>>> or better yet:
>>>>
>>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
>>>
>>
>> No responses.......
>> Since I think it adds too much complexity, unless there's some 
>> support of Ken's suggestion Very Soon, I will toss the idea, and buy 
>> Ken a glass of Bordeaux in Paris as compensation.
>
> I can see some uses for this - as well as SCAN, there's the discussion 
> of a CHANGEDSINCE possibility, and I think Cyrus mentioned something 
> about searching mailbox-level annotations, which could use this kind 
> of syntax.

A clarification to my previous post: as any unrecognized criterion 
should cause BAD response anyway, it is Ok not to define selection 
criteria parameters in the LISTEXT document, even though it would be 
nice to have either generic syntax or some recommendations on what 
should be used by extensions.

> However, I'm assuming a later specification could update the syntax to 
> include whatever we find we need - is there anything that would 
> prevent that?

Nothing I can think of at the moment.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AhmFN079427; Thu, 9 Jun 2005 03:43:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59AhmtW079426; Thu, 9 Jun 2005 03:43:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59AhlIv079415 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:43:47 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [172.16.2.110] (shiny.isode.com [62.3.217.250])  by rufus.isode.com via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 11:43:43 +0100
Message-ID: <42A81D60.3030503@isode.com>
Date: Thu, 09 Jun 2005 11:43:44 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Philip Guenther <guenther+imap@sendmail.com>
CC: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: comments on IMAP-ABNF extensions draft -02
References: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
In-Reply-To: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Philip Guenther wrote:

>I'm not convinced that this draft is taking the best course for defining
>the syntax for extensions.  I see this document as a place to tweak how
>the IMAP grammar is expressed to make extensions easier to express,
>*without actually actually extending it*.  That is, extensions should be
>able express their syntaxes as amending the grammar in RFC 3501 and this
>draft by incremental alternative ("=/") only and without replacing
>grammar elements defined in RFC 3501 or this draft, preferably without
>having to repeat redundant tokens.
>
>And yes, this goal is of course limited by our ability to predict how
>commands will be extended, but the key point is what it excludes:
>servers are *never* required to accept syntax for extensions that they
>don't support**.  This contrasts against the (few) places where the RFC
>3501 grammar *requires* clients to unilaterally accept extensible data
>with open-ended semantics: body-extension, flag-extension, and
>resp-text-code.  As currently written, several of the rules in this
>draft have an open-ended syntax that servers would apparently have to
>parse, only so that they can reject them with a BAD.
>  
>
I don't think that this statement is true, at least it wasn't the intent.

[...]

>IMHO, the tagged-ext syntax (or something similarly simple) is Good
>Enough to be referenced by FETCH, STORE, SELECT, and SEARCH too.  It
>certainly appears to cover the already scoped extensions for CONDSTORE,
>ANNOTATE, and several others. Using a single syntax for extensions to
>commands would make the syntax part of this draft easier to read.
>
Yes, I know. Apart from the SEARCH command extensions the grammer in the 
document is taken from other documents like CONDSTORE, ANNOTATE, etc.
What I've discovered was that all commands *were* extended differently, 
which is a pitty.

>If it *isn't* sufficient, then I suspect the Right Thing might be to simply
>say that extensions should consist of a atom as a tag, followed by
>exactly one argument with balanced parenthesis.
>
Yes, this can be done for the newly defined things.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59ALpiX072082; Thu, 9 Jun 2005 03:21:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59ALp2Q072081; Thu, 9 Jun 2005 03:21:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout5.cac.washington.edu (mxout5.cac.washington.edu [140.142.32.135]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59ALo7G072073 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:21:50 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout5.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j59ALfrM019510 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 9 Jun 2005 03:21:42 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j59ALc6u027146 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 9 Jun 2005 03:21:41 -0700
Date: Thu, 9 Jun 2005 03:21:37 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118309793.629223@localhost.localdomain>
Message-ID: <Pine.OSX.4.63.0506090317440.12669@pangtzu.panda.com>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>   <28904.1118220817.700674@localhost.localdomain>   <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>  <28904.1118258976.392523@localhost.localdomain>  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118309793.629223@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 9 Jun 2005, Dave Cridland wrote:
> Of course, you're right in that we require the ability to base the value on 
> stat() time_t values for some implementations, and I agree we should provide 
> recommendations for implementors choosing this method, but it is and should 
> remain an implementation issue, and confined to an appendix.

I have no problem with this, as long as the word "require" above remains 
operative.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j59A01L1063913; Thu, 9 Jun 2005 03:00:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j59A01qp063912; Thu, 9 Jun 2005 03:00:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xxKs063889 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 03:00:00 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:59:52 +0100
Message-ID: <42A8130E.2060301@isode.com>
Date: Thu, 09 Jun 2005 10:59:42 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Barry Leiba <leiba@watson.ibm.com>
CC: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba wrote:

>I've made the editorial changes; thanks.
>  
>
>>- section 6:  should we restrict selection options to be simple atoms
>>as we have in this draft, or should we allow for them to be a list as
>>well?   I'm thinking that some future selection option might take one
>>or more arguments.  For example, the non-standard SCAN extension used
>>by Pine and UW-IMAP could be reworked to look something like this:
>>
>>x LIST ((SCAN "foo")) "" "bar/%"
>>
>>or better yet:
>>
>>x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
>>    
>>
>
>I have no objection, myself, to extending things further in this way,
>but I point out that it *is* a significant additional extension, not
>just a simple issue of grammar.  Now there may be many more things to
>test in the selection mechanism, for one thing -- it's not just
>inspecting whether the mailbox matches some "attribute" (like REMOTE or
>SUBSCRIBED), but also whether it meets some other open-ended criteria.
>
Yes, and I believe several people have proposed this already. So 
ignoring the open ended syntax issue: this is something that should be 
allowed by the spec.

>The grammar may also be open-ended (as in your second example; why stop
>at two astrings?), which can cause us problems.
>
>Also also, we now have a complication with boolean conditions.  What
>might "(SUBSCRIBED (SCAN "foo") (SCAN "bar"))" mean?
>
>We should do this right, while we're doing it.  I'm just not sure, for
>this situation, what "right" is.  My sense, from talking with some of
>the client authors (mostly with Cyrus), is that this goes beyond what
>they need this for.
>  
>
Actually Dave Cridland wanted to have parameters to selection criteria, 
for example to list all mailboxes changed since MODSEQ. So, if we can 
nail the basic syntax now, that would be good.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xuDw063874; Thu, 9 Jun 2005 02:59:56 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599xtWo063872; Thu, 9 Jun 2005 02:59:55 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599xtpG063824 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:59:55 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 5881B154009; Thu,  9 Jun 2005 09:59:47 +0000 (UTC)
Date: Thu, 09 Jun 2005 10:51:24 +0100
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
MIME-Version: 1.0
Message-Id: <28904.1118310684.785236@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Barry Leiba <leiba@watson.ibm.com>
Cc: IMAP extensions Mailing List <ietf-imapext@imc.org>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 19:27:01 2005, Barry Leiba wrote:
> 
>>> - section 6:  should we restrict selection options to be simple 
>>> atoms
>>> as we have in this draft, or should we allow for them to be a 
>>> list as
>>> well?   I'm thinking that some future selection option might take 
>>> one
>>> or more arguments.  For example, the non-standard SCAN extension 
>>> used
>>> by Pine and UW-IMAP could be reworked to look something like this:
>>> 
>>> x LIST ((SCAN "foo")) "" "bar/%"
>>> 
>>> or better yet:
>>> 
>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
> 
> No responses.......
> Since I think it adds too much complexity, unless there's some 
> support of Ken's suggestion Very Soon, I will toss the idea, and 
> buy Ken a glass of Bordeaux in Paris as compensation.

I can see some uses for this - as well as SCAN, there's the 
discussion of a CHANGEDSINCE possibility, and I think Cyrus mentioned 
something about searching mailbox-level annotations, which could use 
this kind of syntax.

However, I'm assuming a later specification could update the syntax 
to include whatever we find we need - is there anything that would 
prevent that?

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599oqTW060829; Thu, 9 Jun 2005 02:50:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599oqka060828; Thu, 9 Jun 2005 02:50:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599oqFw060816 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:50:52 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:50:50 +0100
Message-ID: <42A810FA.2060304@isode.com>
Date: Thu, 09 Jun 2005 10:50:50 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
CC: ietf-imapext@imc.org, Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, Dave Cridland <dave@cridland.net>
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <1118233852.29860.17.camel@localhost.localdomain> <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
In-Reply-To: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Arnt Gulbrandsen wrote:

> Timo Sirainen writes:
>
>> If we assume that timestamp can be increased after each modification, 
>> can't we just assume that timestamp is one implementation of 
>> CONDSTORE's MODSEQ? In that case we don't need a whole new timestamp 
>> thing, just a possibility to get latest MODSEQ from LIST reply.
>
> Now there's a good idea. How much baggage does it drag along - how 
> much of CONDSTORE has to be provided to provide MODSEQ?

Server code needs to update MODSEQs on each STORE operation. That should 
be sufficient.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599j387058874; Thu, 9 Jun 2005 02:45:03 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599j3Pm058873; Thu, 9 Jun 2005 02:45:03 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599j2Y0058831 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:45:02 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id D0F63154009; Thu,  9 Jun 2005 09:44:55 +0000 (UTC)
Date: Thu, 09 Jun 2005 10:36:33 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>   <28904.1118220817.700674@localhost.localdomain>   <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>  <28904.1118258976.392523@localhost.localdomain>  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To:  <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
MIME-Version: 1.0
Message-Id: <28904.1118309793.629223@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <MRC@CAC.Washington.EDU>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu Jun  9 00:20:57 2005, Mark Crispin wrote:
> On Wed, 8 Jun 2005, Dave Cridland wrote:
>> b) Nothing can show the actual time anyway, at least with the 
>> technology we had back in 2005.
> 
> Is this really true?  Suppose the client wants to show a "human 
> readable" last date/time?
> 
> 
I believe it is true. I also believe it's an orthogonal problem. Any 
direct, protocol-specified use of timestamps as modification 
indicators (as opposed to their use as timestamps) would inevitably 
cause problems. Not least, didn't those idiots back in 2005 realise 
that we'd be updating mailboxes faster than once per millisecond?


>> If ever we need to know the system time of a particular update, 
>> we'll need to do that in a distinct way anyway - system time can, 
>> under some circumstances, go backwards. (Or it could in 2005, 
>> anyway.)
> 
> Leaving aside the question of drifting clocks, there's still the 
> question of a "human readable" last date/time for user display.
> 
> 
. STATUS INBOX (TIMESTAMP)
* STATUS INBOX (TIMESTAMP "09-Jun-2005 09:32:00+0000"
. OK STATUS completed.
. FETCH 1 TIMESTAMP
* 1 FETCH (TIMESTAMP "09-Jun-2005 09:32:00+0000")
. OK FETCH completed.

This does assume that humans aren't interested in millisecond 
accuracy, of course.

But that's no use to test for modifications, purely for use to 
display to 2005-era humans.


>>> I like the idea of using a CONDSTORE type mechanism, as long as 
>>> it can be implemented using a file date from the operating system 
>>> (e.g. from the stat() system call on UNIX).  Maybe not a "best" 
>>> implementation (in that there may be requirements for 
>>> synchronization delays if the value's resolution is too low) but 
>>> nonetheless a viable and compliant implementation.
>> Your earlier description of taking a timestamp, multiplying it to 
>> emulate an increased resolution, and "making up" the extra 
>> resolution to fit the specification, ought to work fine.
> 
> We also need some sort of stipulation that this value is not 
> allowed to go backwards, and recommendations for preventing that 
> from happening, such as detecting an old time-based value that is 
> in the future, and monotonically increasing it on update rather 
> than setting it to time.

Yes, we need this, but it damages the relationship between the 64-bit 
modseq and the actual time, which in turn means that clients cannot 
rely on it to provide a human-readable, accurate time.

I really believe that timestamps are irrelevant to the actual 
protocol here - all we need is some kind of data type which is 
comparable, either less than, greater than, or equal to. A 64-bit 
integer is fast, and has precendence in CONDSTORE.

Of course, you're right in that we require the ability to base the 
value on stat() time_t values for some implementations, and I agree 
we should provide recommendations for implementors choosing this 
method, but it is and should remain an implementation issue, and 
confined to an appendix.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599WtXe054888; Thu, 9 Jun 2005 02:32:55 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j599WtMH054887; Thu, 9 Jun 2005 02:32:55 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j599WrkC054868 for <ietf-imapext@imc.org>; Thu, 9 Jun 2005 02:32:54 -0700 (PDT) (envelope-from alexey.melnikov@isode.com)
Received: from [192.168.0.3] ([62.3.217.253]) by rufus.isode.com  via TCP (internal) with ESMTPA; Thu, 9 Jun 2005 10:32:49 +0100
Message-ID: <42A80CC1.8080001@isode.com>
Date: Thu, 09 Jun 2005 10:32:49 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
To: Mark Crispin <mrc@cac.washington.edu>
CC: Dave Cridland <dave@cridland.net>, IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain> <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
MIME-version: 1.0
Content-type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-transfer-encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin wrote:

> On Wed, 8 Jun 2005, Dave Cridland wrote:
>
>> The client only needs to know that a change has happened, never the 
>> actual times involved.
>
> The question is if a client ever will care about the time.  I don't 
> want to have to revisit this issue some years from now due to a "what 
> you did in 2005 sucks since it doesn't show the actual time, and 
> that's what we need now."
>
>>> I happen to think that a one-second resolution is probably good 
>>> enough for most practical purposes.
>>
>> I worry about the "probably" and "most" in that sentence. :-) 
>> Moreover there's no need to have a time-based granularity at all.
>
> Words such as "probably" and "most" are the earmarks of handwaving; 
> and that statement was intended as a handwave (and an opinion at 
> that), not a statement of fact.  :-)
>
> I like the idea of using a CONDSTORE type mechanism, as long as it can 
> be implemented using a file date from the operating system (e.g. from 
> the stat() system call on UNIX).  Maybe not a "best" implementation 
> (in that there may be requirements for synchronization delays if the 
> value's resolution is too low) but nonetheless a viable and compliant 
> implementation.

As long as OS time doesn't go back, that should be possible.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j591CuaG020098; Wed, 8 Jun 2005 18:12:56 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j591CuIu020097; Wed, 8 Jun 2005 18:12:56 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from eagle.oceana.com (eagle.oceana.com [208.17.123.12]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j591CtRU020087 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 18:12:55 -0700 (PDT) (envelope-from ken@oceana.com)
Received: from [192.168.137.140] (69-161-65-27.bflony.adelphia.net [69.161.65.27]) (authenticated bits=0) by eagle.oceana.com (8.13.2/8.13.2) with ESMTP id j591ClNP005617 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 21:12:48 -0400 (EDT)
Message-ID: <42A7978A.9070605@oceana.com>
Date: Wed, 08 Jun 2005 21:12:42 -0400
From: Ken Murchison <ken@oceana.com>
Organization: Oceana Matrix Ltd.
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040514
X-Accept-Language: en,pdf
MIME-Version: 1.0
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com> <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0 (BAYES_00)
X-Scanned-By: MIMEDefang 2.39
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba wrote:

> 
>>> - section 6:  should we restrict selection options to be simple atoms
>>> as we have in this draft, or should we allow for them to be a list as
>>> well?   I'm thinking that some future selection option might take one
>>> or more arguments.  For example, the non-standard SCAN extension used
>>> by Pine and UW-IMAP could be reworked to look something like this:
>>>
>>> x LIST ((SCAN "foo")) "" "bar/%"
>>>
>>> or better yet:
>>>
>>> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"
> 
> 
> No responses.......
> Since I think it adds too much complexity, unless there's some support 
> of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a glass 
> of Bordeaux in Paris as compensation.


It was just a thought, since I was thinking about Mark's SCAN extension 
at the time.  I won't be devastated if we leave things alone.

Unfortunately, I won't be in Paris, and the compensation wouldn't be 
necessary anyways.  :)

-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58NNGrw012465; Wed, 8 Jun 2005 16:23:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58NNGiD012464; Wed, 8 Jun 2005 16:23:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58NNCJZ012455 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 16:23:12 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58NMxtL004968 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 8 Jun 2005 16:22:59 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58NMxTA022133 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Wed, 8 Jun 2005 16:22:59 -0700
Date: Wed, 8 Jun 2005 16:20:57 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118258976.392523@localhost.localdomain>
Message-ID: <Pine.WNT.4.64.0506081613190.2288@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com> <28904.1118258976.392523@localhost.localdomain>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 8 Jun 2005, Dave Cridland wrote:
> b) Nothing can show the actual time anyway, at least with the technology we 
> had back in 2005.

Is this really true?  Suppose the client wants to show a "human readable" 
last date/time?

A 32-bit ulong (and all IMAP 32-bit values *are* ulongs) since 2005-Jan-01 
00:00:00 UTC will work until February 2106.  The much talked-about 
doomsday at 2038-Jan-19 03:14:08 is because UNIX uses longs, not ulongs.

> If ever we need to know the system time of a particular update, we'll need to 
> do that in a distinct way anyway - system time can, under some circumstances, 
> go backwards. (Or it could in 2005, anyway.)

Leaving aside the question of drifting clocks, there's still the question 
of a "human readable" last date/time for user display.

>> I like the idea of using a CONDSTORE type mechanism, as long as it can be 
>> implemented using a file date from the operating system (e.g. from the 
>> stat() system call on UNIX).  Maybe not a "best" implementation (in that 
>> there may be requirements for synchronization delays if the value's 
>> resolution is too low) but nonetheless a viable and compliant 
>> implementation.
> Your earlier description of taking a timestamp, multiplying it to emulate an 
> increased resolution, and "making up" the extra resolution to fit the 
> specification, ought to work fine.

We also need some sort of stipulation that this value is not allowed to go 
backwards, and recommendations for preventing that from happening, such as 
detecting an old time-based value that is in the future, and monotonically 
increasing it on update rather than setting it to time.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jlq5C096261; Wed, 8 Jun 2005 12:47:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58JlqAg096260; Wed, 8 Jun 2005 12:47:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58JlpCf096251 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 12:47:51 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id C2ED9142271; Wed,  8 Jun 2005 12:46:00 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 06673-08; Wed, 8 Jun 2005 12:46:00 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 264E514226E; Wed,  8 Jun 2005 12:46:00 -0700 (PDT)
Message-ID: <42A74B60.9040102@osafoundation.org>
Date: Wed, 08 Jun 2005 12:47:44 -0700
From: Lisa Dusseault <lisa@osafoundation.org>
User-Agent: Mozilla Thunderbird 1.0 (Macintosh/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Cc: ietf-imapext@imc.org, Barry Leiba <leiba@watson.ibm.com>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>  <42A4897C.1080609@oceana.com>  <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>  <42A73875.3040501@watson.ibm.com> <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
In-Reply-To: <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Yes, let's have a Friends of IMAP supper in Paris!  I've asked for a 
meeting slot but I don't know what day it will be on yet.

Lisa

Arnt Gulbrandsen wrote:

>
> Barry Leiba writes:
>
>> No responses.......
>> Since I think it adds too much complexity, unless there's some 
>> support of Ken's suggestion Very Soon, I will toss the idea, and buy 
>> Ken a glass of Bordeaux in Paris as compensation.
>
>
> I didn't like it, sounded like a bug magnet to me, but I decided to 
> say nothing in case someone came up with justification.
>
> How about not having an IMAP lunch, btw? How about a proper dinner at 
> a marginally expensive place instead? (You can buy Ken a glass of 
> vintage armagnac afterwards, that's sure to make him forgive you.)
>
> Arnt
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jbk8O095712; Wed, 8 Jun 2005 12:37:46 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58Jbkn2095711; Wed, 8 Jun 2005 12:37:46 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Jbi1x095703 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 12:37:45 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id C5C58154009; Wed,  8 Jun 2005 19:37:38 +0000 (UTC)
Date: Wed, 08 Jun 2005 20:29:36 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>   <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
MIME-Version: 1.0
Message-Id: <28904.1118258976.392523@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <mrc@CAC.Washington.EDU>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 16:18:27 2005, Mark Crispin wrote:
> On Wed, 8 Jun 2005, Dave Cridland wrote:
>> The client only needs to know that a change has happened, never 
>> the actual times involved.
> 
> The question is if a client ever will care about the time.  I don't 
> want to have to revisit this issue some years from now due to a 
> "what you did in 2005 sucks since it doesn't show the actual time, 
> and that's what we need now."
> 
> 
I'd argue then that:
a) CONDSTORE had already prepackaged that can of worms for us, by 
using an essentially opaque 64-bit int.
b) Nothing can show the actual time anyway, at least with the 
technology we had back in 2005.

If ever we need to know the system time of a particular update, we'll 
need to do that in a distinct way anyway - system time can, under 
some circumstances, go backwards. (Or it could in 2005, anyway.)

> I like the idea of using a CONDSTORE type mechanism, as long as it 
> can be implemented using a file date from the operating system 
> (e.g. from the stat() system call on UNIX).  Maybe not a "best" 
> implementation (in that there may be requirements for 
> synchronization delays if the value's resolution is too low) but 
> nonetheless a viable and compliant implementation.

Your earlier description of taking a timestamp, multiplying it to 
emulate an increased resolution, and "making up" the extra resolution 
to fit the specification, ought to work fine.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58Ig4CC091741; Wed, 8 Jun 2005 11:42:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58Ig4UK091740; Wed, 8 Jun 2005 11:42:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j58Ig3dY091734 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:42:04 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id DCEC81A754; Wed,  8 Jun 2005 20:42:01 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 5694F1A6E1; Wed,  8 Jun 2005 20:41:55 +0200 (CEST)
Message-Id: <DW483QjsNo1oyRmmkRHQTQ.md5@prosecco.oryx.com>
Date: Wed, 8 Jun 2005 20:40:55 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: draft-ietf-imapext-list-extensions-13
Cc: Barry Leiba <leiba@watson.ibm.com>
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com> <42A73875.3040501@watson.ibm.com>
In-Reply-To: <42A73875.3040501@watson.ibm.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Barry Leiba writes:
> No responses.......
> Since I think it adds too much complexity, unless there's some support 
> of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a 
> glass of Bordeaux in Paris as compensation.

I didn't like it, sounded like a bug magnet to me, but I decided to say 
nothing in case someone came up with justification.

How about not having an IMAP lunch, btw? How about a proper dinner at a 
marginally expensive place instead? (You can buy Ken a glass of vintage 
armagnac afterwards, that's sure to make him forgive you.)

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58IR9QZ090492; Wed, 8 Jun 2005 11:27:09 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58IR9wZ090491; Wed, 8 Jun 2005 11:27:09 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58IR8ud090483 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:27:08 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j58ISMLV005282 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:28:22 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j58IR2k123308 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:27:02 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j58IR1d109726 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:27:01 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118255221.284; Wed, 08 Jun 2005 14:27:01 -0400
Message-ID: <42A73875.3040501@watson.ibm.com>
Date: Wed, 08 Jun 2005 14:27:01 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com> <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

>>- section 6:  should we restrict selection options to be simple atoms
>>as we have in this draft, or should we allow for them to be a list as
>>well?   I'm thinking that some future selection option might take one
>>or more arguments.  For example, the non-standard SCAN extension used
>>by Pine and UW-IMAP could be reworked to look something like this:
>>
>>x LIST ((SCAN "foo")) "" "bar/%"
>>
>>or better yet:
>>
>>x LIST ((SCAN "utf-8" "foo")) "" "bar/%"

No responses.......
Since I think it adds too much complexity, unless there's some support 
of Ken's suggestion Very Soon, I will toss the idea, and buy Ken a glass 
of Bordeaux in Paris as compensation.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58INZR6090317; Wed, 8 Jun 2005 11:23:35 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58INZI1090316; Wed, 8 Jun 2005 11:23:35 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58INYnh090305 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 11:23:34 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j58IOlmT004737 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:24:48 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j58INSk101542 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:23:28 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j58INRd102098 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 14:23:27 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118255007.283; Wed, 08 Jun 2005 14:23:27 -0400
Message-ID: <42A7379F.2080007@watson.ibm.com>
Date: Wed, 08 Jun 2005 14:23:27 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com> <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To: <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Folks, really... this has strayed widely from what I think Simon's 
original issue was.  If I have this right, Simon was talking about 
synchronizing client cached state when coming back online.  What Timo 
has been arguing about is, as Mark noted, notification.  These are 
vastly different issues, and people have already proposed models for 
multi-mailbox notification schemes.

If we're concerned about notification, we should work on a proposal that 
fits that.  Then a client can have one socket open to the server to be 
notified on, and can provide the server with a list of mailboxes to 
monitor.  Sorta like a multi-mailbox IDLE.  This is waaaay better than 
polling, regardless of how you poll.

If that doesn't work because you can't keep a socket open, then look at 
the Lemonade work, and at Stephane Maes's "push IMAP" spec -- which has 
a lot of complexity, but the widely useful parts of it are being plucked 
out separately.

Now, back to the original question:  Can Simon (and others) resynch his 
client cache effectively without doing STATUS on every mailbox?  Or 
should we have an extension that enables STATUS options on an extended 
LIST command, to make his work more efficient on the servers that 
support it?

Note that however that last question is answered, I do NOT think this 
should affect the LIST extensions spec we're working on now.  It would, 
if anything, be another extension, to be supported by servers or not, 
independently of whether they support the base LIST extensions.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58FIv5N070862; Wed, 8 Jun 2005 08:18:57 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58FIvoE070861; Wed, 8 Jun 2005 08:18:57 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58FIrj3070853 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 08:18:53 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58FIhea015050 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 8 Jun 2005 08:18:43 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j58FISWI018977 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Wed, 8 Jun 2005 08:18:42 -0700
Date: Wed, 8 Jun 2005 08:18:27 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Dave Cridland <dave@cridland.net>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <28904.1118220817.700674@localhost.localdomain>
Message-ID: <Pine.OSX.4.63.0506080807400.266@pangtzu.panda.com>
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 8 Jun 2005, Dave Cridland wrote:
> The client only needs to know that a change has happened, never the actual 
> times involved.

The question is if a client ever will care about the time.  I don't want 
to have to revisit this issue some years from now due to a "what you did 
in 2005 sucks since it doesn't show the actual time, and that's what we 
need now."

>> I happen to think that a one-second resolution is probably good enough for 
>> most practical purposes.
> I worry about the "probably" and "most" in that sentence. :-) Moreover 
> there's no need to have a time-based granularity at all.

Words such as "probably" and "most" are the earmarks of handwaving; and 
that statement was intended as a handwave (and an opinion at that), not a 
statement of fact.  :-)

I like the idea of using a CONDSTORE type mechanism, as long as it can be 
implemented using a file date from the operating system (e.g. from the 
stat() system call on UNIX).  Maybe not a "best" implementation (in that 
there may be requirements for synchronization delays if the value's 
resolution is too low) but nonetheless a viable and compliant 
implementation.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58EdJTu066910; Wed, 8 Jun 2005 07:39:19 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58EdJej066909; Wed, 8 Jun 2005 07:39:19 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58EdIAA066900 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 07:39:19 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 72D30154009; Wed,  8 Jun 2005 14:39:12 +0000 (UTC)
Date: Wed, 08 Jun 2005 15:31:17 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>  <28904.1118220817.700674@localhost.localdomain>  <1118233852.29860.17.camel@localhost.localdomain>  <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
In-Reply-To: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
MIME-Version: 1.0
Message-Id: <28904.1118241077.035520@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, <ietf-imapext@imc.org>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed Jun  8 13:38:58 2005, Arnt Gulbrandsen wrote:
> Timo Sirainen writes:
>> If we assume that timestamp can be increased after each 
>> modification, can't we just assume that timestamp is one 
>> implementation of CONDSTORE's MODSEQ? In that case we don't need a 
>> whole new timestamp thing, just a possibility to get latest MODSEQ 
>> from LIST reply.
> 
> Now there's a good idea. How much baggage does it drag along - how 
> much of CONDSTORE has to be provided to provide MODSEQ?

Well, there's two things here.

Firstly, even if CONDSTORE is implemented, not every mailbox has a 
HIGHESTMODSEQ.

Secondly, although any increase in HIGHESTMODSEQ must - assuming 
we're using Alexey's STATUS-inna-LIST concept - also increase the 
LIST's magic sequence, HIGHESTMODSEQ is per-mailbox, while the 
LISTS's sequence is per-logical-server (ie, per server or per 
cluster, whichever applies). This means that when we wrap the LIST 
sequence, we have to wrap all the mailboxes' sequences as well.

The second issue is pretty minor - I'd fully expect any server 
implementing both CONDSTORE and CHANGEDSINCE LIST to use the same 
sequence number for both purposes, however, it's not essential. The 
first is more crucial - it seems fairly logical, to me, that certain 
types of mailbox can be made to support CONDSTORE easier than others, 
so an IMAP implementation with multiple backends (which, IIRC, 
includes both Mark and Timo's implementations) may well only support 
CONDSTORE on some.

So while I'd expect that in many cases, servers implementing 
CONDSTORE would havea  single sequence, I wouldn't want to mandate 
that.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CeCQO045166; Wed, 8 Jun 2005 05:40:12 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58CeCdh045165; Wed, 8 Jun 2005 05:40:12 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j58CeApY045150 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 05:40:11 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id B93F01A750; Wed,  8 Jun 2005 14:40:09 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 987271A710; Wed,  8 Jun 2005 14:39:56 +0200 (CEST)
Message-Id: <954MjjA2BDBrdsuP066SFg.md5@prosecco.oryx.com>
Date: Wed, 8 Jun 2005 14:38:58 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>, Dave Cridland <dave@cridland.net>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi> <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain> <1118233852.29860.17.camel@localhost.localdomain>
In-Reply-To: <1118233852.29860.17.camel@localhost.localdomain>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Timo Sirainen writes:
> If we assume that timestamp can be increased after each modification, 
> can't we just assume that timestamp is one implementation of 
> CONDSTORE's MODSEQ? In that case we don't need a whole new timestamp 
> thing, just a possibility to get latest MODSEQ from LIST reply.

Now there's a good idea. How much baggage does it drag along - how much 
of CONDSTORE has to be provided to provide MODSEQ?

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CV1aM041998; Wed, 8 Jun 2005 05:31:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j58CV1Bs041997; Wed, 8 Jun 2005 05:31:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from dmgw.movial.fi (dmgw.movial.fi [62.236.91.5]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j58CV0Mq041952 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 05:31:00 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from movial.fi (hellmail.movial.fi [62.236.91.6]) by dmgw.movial.fi (Postfix) with ESMTP id D1A53EC12B; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Received: from localhost (hellmail.movial.fi [127.0.0.1]) by movial.fi (Postfix) with ESMTP id AB2538005D; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Received: from movial.fi ([127.0.0.1]) by localhost (hellmail [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 06628-05; Wed, 8 Jun 2005 15:30:53 +0300 (EEST)
Received: from sd022.hel.movial.fi (sd022.hel.movial.fi [172.17.49.22]) (using SSLv3 with cipher RC4-MD5 (128/128 bits)) (Client did not present a certificate) by movial.fi (Postfix) with ESMTP id 79F2780052; Wed,  8 Jun 2005 15:30:53 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Dave Cridland <dave@cridland.net>
Cc: Mark Crispin <MRC@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
In-Reply-To: <28904.1118220817.700674@localhost.localdomain>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi> <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU> <28904.1118220817.700674@localhost.localdomain>
Content-Type: text/plain
Date: Wed, 08 Jun 2005 12:30:52 +0000
Message-Id: <1118233852.29860.17.camel@localhost.localdomain>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: by amavisd-new-20030616-p10 (Debian) at movial.fi
X-Amavis-Alert: BAD HEADER Improper use of control character (char 0D hex) in message header 'References' References: ...46801216@PD.gordano.com>\n\t \r <Pine.WNT.4.6... ^
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 2005-06-08 at 09:53 +0100, Dave Cridland wrote:
> > On Tue, 7 Jun 2005, Timo Sirainen wrote:
> >>> We haven't defined what the timestamp interval should be.
> >> Isn't it going to be seconds if you want it to be supported by 
> >> legacy stores?
> > 
> > Not necessarily.
> > 
> > A store which only resolves to a second can comply very simply.  It 
> > just multiplies the second-resolution values by whatever the 
> > necessary scale factor is, and guarantees that new message(s) 
> > delivery always updates the value.
> > 
> Exactly - that's essentially what an ACAP modtime is. Or you could 
> use a system similar to CONDSTORE's MODSEQ - an always-increasing 
> 64-bit integer. The client only needs to know that a change has 
> happened, never the actual times involved.

If we assume that timestamp can be increased after each modification,
can't we just assume that timestamp is one implementation of CONDSTORE's
MODSEQ? In that case we don't need a whole new timestamp thing, just a
possibility to get latest MODSEQ from LIST reply.




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5891gqT069179; Wed, 8 Jun 2005 02:01:42 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5891gTP069178; Wed, 8 Jun 2005 02:01:42 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from gw2.gestalt.entity.net (dsl-217-155-137-62.zen.co.uk [217.155.137.62]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5891Zrb069103 for <ietf-imapext@imc.org>; Wed, 8 Jun 2005 02:01:36 -0700 (PDT) (envelope-from dave@cridland.net)
Received: from peirce (dsl-217-155-137-61.zen.co.uk [217.155.137.61]) by gw2.gestalt.entity.net (Postfix) with ESMTP id 43E9A154009; Wed,  8 Jun 2005 09:01:25 +0000 (UTC)
Date: Wed, 08 Jun 2005 09:53:37 +0100
Subject: Re: IMAP4 List Command Extensions
References: <06480646801216@PD.gordano.com>   <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>   <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>  <1118089030.15166.502.camel@hurina>  <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>  <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
In-Reply-To:  <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
MIME-Version: 1.0
Message-Id: <28904.1118220817.700674@localhost.localdomain>
From: Dave Cridland <dave@cridland.net>
To: Mark Crispin <MRC@cac.washington.edu>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon Jun  6 23:17:09 2005, Mark Crispin wrote:
> 
> On Tue, 7 Jun 2005, Timo Sirainen wrote:
>>> We haven't defined what the timestamp interval should be.
>> Isn't it going to be seconds if you want it to be supported by 
>> legacy stores?
> 
> Not necessarily.
> 
> A store which only resolves to a second can comply very simply.  It 
> just multiplies the second-resolution values by whatever the 
> necessary scale factor is, and guarantees that new message(s) 
> delivery always updates the value.
> 
> 
Exactly - that's essentially what an ACAP modtime is. Or you could 
use a system similar to CONDSTORE's MODSEQ - an always-increasing 
64-bit integer. The client only needs to know that a change has 
happened, never the actual times involved.

I do like ACAP's modtimes, because they lend a human-readability to 
it, but the human semantics aren't required for the client at all. 
(Indeed, it makes no sense to try - the modtimes my ACAP 
implementation uses are only "usually" accurate to within a second, 
and may be wildly out to satisfy the always-increasing requirement, 
which is more important.)

> I happen to think that a one-second resolution is probably good 
> enough for most practical purposes.

I worry about the "probably" and "most" in that sentence. :-) 
Moreover there's no need to have a time-based granularity at all.

>> 1) LIST -> notice new stamp
>> 2) SELECT mailbox
>> 3) notice the new mail, read it, etc.
>> 4) another mail arrives
>> 5) SELECT another-mailbox
>> The mail is lost unless server notifies client of it before the 
>> SELECT command. I don't think many (any?) servers do that 
>> currently.

I'm not sure the mail is missed, because the timestamp (or sequence 
value) would increase, and if the client never receives a 
notification of (4), it will assume that the mailbox is in the state 
as of the end of (3). The only circumstance you might have problems 
with is if you were foolish enough to use real timestamps, and (4) 
and (1) (or (2), if the timestamp is reported there, too) happened 
within the granularity - but as I say, that would be foolish design 
indeed.

I think there's a huge scope for improvement there. Alexey and I have 
been idly talking about this area in private mail a while ago, so the 
more sensible ideas here are probably his.

If we bring in Alexey's STATUS-in-LIST concept, we can issue 
something like (and I'm stretching LIST-EXTENDED syntax here by quite 
a way):

A01 LIST (SUBSCRIBED (CHANGEDSINCE 20050607202312000)) "" "*" 
(STATUS=UIDVALIDITY,UIDNEXT,MESSAGES)

This could tell us about only the metadata that's changed for a 
mailbox - if any has. This satisfies Simon Tyler's (quite reasonable, 
I thought) request, and should lower the cost for Mark, too - since 
we can assume that, at the very least, a client is likely to issue 
STATUS (if not SELECT) to get the extra information it requires, 
better to allow it to ask for everything in one go. Because the 
"something changed" metadata and the data itself is contained within 
the same message, Timo's concerns about races should vanish.

We can further improve this by adding unsolicited updates, perhaps 
with a syntax like:

A01 LIST (SUBSCRIBED (CHANGEDSINCE 20050607202312000)) "" "*" RETURN 
(STATUS=UIDVALIDITY,UIDNEXT,MESSAGES UPDATES)

This would send updated LIST responses for mailboxes which changed. 
This might merely replace the client's poll with a server-controlled 
poll (thus lowering what Mark describes as abusive behaviour), or 
might be implemented using notifications within the server, avoiding 
polling entirely.

This helps Timo's concerns too - it seems reasonable for a server to 
buffer update LIST responses for a short period to avoid generates 
large numbers of redundant responses.

I appreciate there's a lot more thought that needs to go into this, 
and I'm not wed to any particular detail, so feel free to pull it 
apart.

Dave.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j57CfoWJ082092; Tue, 7 Jun 2005 05:41:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j57CfofK082091; Tue, 7 Jun 2005 05:41:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mail.verisignlabs.com (cliffie.verisignlabs.com [65.201.175.9]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j57CfnuN082078 for <ietf-imapext@imc.org>; Tue, 7 Jun 2005 05:41:49 -0700 (PDT) (envelope-from sah@428cobrajet.net)
Received: from dul1shollenbl1 ([::ffff:216.168.239.87]) (AUTH: LOGIN shollenb, SSL: TLSv1/SSLv3,128bits,RC4-MD5) by mail.verisignlabs.com with esmtp; Tue, 07 Jun 2005 08:41:48 -0400 id 0059811E.42A5960C.00000FCE
From: "Scott Hollenbeck" <sah@428cobrajet.net>
To: ietf-imapext@imc.org
Subject: FW: Last Call Expired: <draft-ietf-imapext-2086upd> 
Date: Tue, 7 Jun 2005 08:42:18 -0400
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook, Build 11.0.6353
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
Thread-Index: AcVrEZSAHYbM9ZjhSEGNLtEw7H/cGwATC4ww
Message-ID: <courier.42A5960C.00000FCE@mail.verisignlabs.com>
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

The IETF last call for draft-ietf-imapext-2086upd has completed.  I didn't
see any new comments.  Please update the document to correct the remaining
editorial issues (including IDNits) identified before the last call and I'll
schedule it for IESG evaluation once the update is available.  The deadline
to make the next telechat is 16 June.

-Scott-

-----Original Message-----
From: DraftTracker Mail System [mailto:iesg-secretary@ietf.org] 
Sent: Monday, June 06, 2005 11:30 PM
To: iesg@ietf.org
Cc: iesg-secretary@ietf.org
Subject: Last Call Expired: <draft-ietf-imapext-2086upd> 


Please DO NOT reply on this email.
 

ID: draft-ietf-imapext-2086upd

IETF Last Call has ended, and the state has been changed to Waiting for AD
Go-Ahead



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56MJHkG078040; Mon, 6 Jun 2005 15:19:17 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56MJH9Y078039; Mon, 6 Jun 2005 15:19:17 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56MJGrr078032 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 15:19:16 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56MJBKM002392 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 15:19:11 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56MJBv1025659 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 15:19:11 -0700
Date: Mon, 6 Jun 2005 15:17:09 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Message-ID: <Pine.WNT.4.64.0506061447000.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU> <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Tue, 7 Jun 2005, Timo Sirainen wrote:
>> We haven't defined what the timestamp interval should be.
> Isn't it going to be seconds if you want it to be supported by legacy stores?

Not necessarily.

A store which only resolves to a second can comply very simply.  It just 
multiplies the second-resolution values by whatever the necessary scale 
factor is, and guarantees that new message(s) delivery always updates the 
value.

I happen to think that a one-second resolution is probably good enough for 
most practical purposes.  I wouldn't be so sanguine if we didn't have the 
atomic multiple-message append mechanism through MULTIAPPEND. 
Fortunately, we do; and it is reasonable to assume that a server that is 
advanced enough to have this new LIST capability can also be advanced 
enough to deliver a flood of messages atomically (and thus require only a 
single interval tick).

> 1) LIST -> notice new stamp
> 2) SELECT mailbox
> 3) notice the new mail, read it, etc.
> 4) another mail arrives
> 5) SELECT another-mailbox
> The mail is lost unless server notifies client of it before the SELECT 
> command. I don't think many (any?) servers do that currently.

You forgot a step: the client obtains the mailbox's current timestamp.

If that happens before (4), then the mailbox will have a different stamp. 
If it happens after (4), an announcement of the new message will as part 
of the operation.

No race.

> Although rare, all that above is possible as a user action.

It's also a terrible abuse of IMAP; it treats IMAP as a stateless protocol 
such as HTTP and completely disregards the entire purpose of IMAP as a 
stateful protocol.

I don't think that we should optimize for abusive use that the protocol 
was never designed to support.

> I sometimes click 
> a mailbox only to notice I mis-clicked and then select another box 
> immediately. I think it's even possible to have time to read the subject, 
> delete the message and switch to another mailbox all within one second.

If you care about monitoring the mailbox, why don't you leave it open?

> 1) select mailbox, do things with it, notice that we need to change the 
> mailbox
> 2) new mail arrives
> 2) issue LIST to get the current timestamp
> 3) new mail arrives
> 4) SELECT another mailbox

In this case, the mailbox's timestamp from (2) will be different (less) 
than the timestamp from (3).

No race.

> But that depends completely on how fast the hardware and software 
> implementation is, and what commands are being used. I think it's still 
> easily possible to handle tens of IMAP commands within one second.

I've looked at lots of IMAP and SMTP/MTA session logs (and not just UW 
imapd).  I think that you're making a stretch.

Even if your case was valid, it is solvable; unlike the alternative which 
is harmful in some server environments *and* breaks a promise made by the 
base specification.

> I don't think that is a "simply". It requires either modifying the existing 
> delivery agent (which I think is not very likely to happen) or running a 
> second delivery agent to fix the timestamp.

Whoa!

If you don't have a whiz-bang format capable of delivering that metadata 
quickly, you aren't going to be able to respond in the times that you 
argue.  And if you have a whiz-bang format then you already have a custom 
mail delivery agent.

If the mail delivery agent isn't part of the IMAP server software package, 
then you haven't a snowball's chance in hell of maintaining a UIDNEXT that 
can be delivered in under a second.

If /bin/mail has appended 500MB of new traditional UNIX format data since 
the last time the IMAP server had the mailbox open, it simply is not going 
to calculate a current UIDNEXT in under a second.

> What probably would happen is that admins will happily enable the timestamp 
> feature in IMAP server without touching the delivery agent. Which makes it 
> unrealiable.

Not at all.

In such circumstances, the IMAP server can also prevent the problem by 
guaranteeing that it never returns a timestamp that equals the current 
time.

> UIDNEXT exists already and can be used for this purpose. If we can't use the 
> timestamps in a portable and useful enough way, I don't think there's a need 
> to invent a new metadata field.

UIDNEXT is potentially expensive and the base specification guarantees 
this fact to a server implementation.  An extension can not repeal a 
guarantee of the base specification.

> I suppose with enough rules (adding arbitrary delays) and requirements (which 
> can be ignored by admins) you could make this reliable enough, but I don't 
> really see them as a good idea.

It's a damned good idea to do something which can be implemented cheaply 
by just about everybody concerned (either at same-cost or much lower cost 
than UIDNEXT), and will not have to do any delays for timing race 
prevention except in extraordinary cases.

> Of course, one possible solution would be to just ignore the races and let 
> IMAP clients deal with it:
> 1) when timestamp changes, check if there is new mail
> 2) one second later check for new mail again
> 3) also check if timestamp changed, if so goto 2

Step (2) accomplishes nothing, because the check for new mail is already 
of newer value than the timestamp.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LepVn074278; Mon, 6 Jun 2005 14:40:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56Lepwx074277; Mon, 6 Jun 2005 14:40:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LeoYG074266 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:40:50 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from [192.168.10.217] (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 1FC6E1C351D2; Tue,  7 Jun 2005 00:40:44 +0300 (EEST)
In-Reply-To: <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina> <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary="Apple-Mail-30-200404062"
Message-Id: <751e0a0e2aa225ab02abc1ada2e19e22@iki.fi>
Content-Transfer-Encoding: 7bit
Cc: IMAP Extensions WG <ietf-imapext@imc.org>, Simon Tyler <simont@gordano.com>
From: Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
Date: Tue, 7 Jun 2005 00:40:41 +0300
To: Mark Crispin <MRC@cac.washington.edu>
X-Pgp-Agent: GPGMail 1.0.2
X-Mailer: Apple Mail (2.622)
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--Apple-Mail-30-200404062
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII; format=flowed

On 6.6.2005, at 23:29, Mark Crispin wrote:

>> So, what needs to happen within one
>> second is:
>
> s/second/timestamp interval
>
> We haven't defined what the timestamp interval should be.

Isn't it going to be seconds if you want it to be supported by legacy 
stores?

> You have not demonstrated any benefit that outweighs the substantial 
> cost.

How about these two cases:

1) LIST -> notice new stamp
2) SELECT mailbox
3) notice the new mail, read it, etc.
4) another mail arrives
5) SELECT another-mailbox

The mail is lost unless server notifies client of it before the SELECT 
command. I don't think many (any?) servers do that currently.

Although rare, all that above is possible as a user action. I sometimes 
click a mailbox only to notice I mis-clicked and then select another 
box immediately. I think it's even possible to have time to read the 
subject, delete the message and switch to another mailbox all within 
one second.

And that's only with "normal" client user interface. Some client could 
merge multiple mailboxes within same list and you could easily be 
selecting multiple mailboxes within one second.

Another possibility:

1) select mailbox, do things with it, notice that we need to change the 
mailbox
2) new mail arrives
2) issue LIST to get the current timestamp
3) new mail arrives
4) SELECT another mailbox

Or should it also be required not to issue LIST for mailbox without 
doing a NOOP one timestamp interval later, before closing the mailbox?

>>> But, assuming that we actually have to fix this (perhaps with a 
>>> massive
>>> speedup of RTTs),
>> I think even with your example it could easily happen nowadays. It's
>> maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
>> pings to my ISP's mail server. That's 125 RTTs within one second.
>
> There's a substantial difference between ping RTTs and RTTs in a 
> higher level protocol such as IMAP.

But that depends completely on how fast the hardware and software 
implementation is, and what commands are being used. I think it's still 
easily possible to handle tens of IMAP commands within one second. With 
fast enough servers the network latency is the only measureable one and 
ping reflects that pretty well.

>> I
>> don't see how you could actually get it to work with mboxes (and other
>> legacy stores) as the mails are usually written using tons of 
>> different
>> programs that aren't going to be increasing the timestamp.
>
> Not at all.  The mail delivery agent is simply required to ensure that 
> the timestamp changes, and if necessary delays delivery so that the 
> new message(s) get an updated timestamp.

I don't think that is a "simply". It requires either modifying the 
existing delivery agent (which I think is not very likely to happen) or 
running a second delivery agent to fix the timestamp. And the secondary 
timestamp-agent would also have to lock the mailbox if you want it free 
of race conditions. Yet another software needing to know how mboxes are 
being locked..

What probably would happen is that admins will happily enable the 
timestamp feature in IMAP server without touching the delivery agent. 
Which makes it unrealiable.

>> And I wouldn't want to be polling timestamps either. I think server
>> should be notifying IDLEing clients of changes in them automatically.
>
> Then why are you talking about using UIDNEXT at all?

UIDNEXT exists already and can be used for this purpose. If we can't 
use the timestamps in a portable and useful enough way, I don't think 
there's a need to invent a new metadata field.

I suppose with enough rules (adding arbitrary delays) and requirements 
(which can be ignored by admins) you could make this reliable enough, 
but I don't really see them as a good idea.

Of course, one possible solution would be to just ignore the races and 
let IMAP clients deal with it:

1) when timestamp changes, check if there is new mail
2) one second later check for new mail again
3) also check if timestamp changed, if so goto 2

Actually I'd prefer this if you really must implement non-accurate 
timestamps.

--Apple-Mail-30-200404062
content-type: application/pgp-signature; x-mac-type=70674453;
	name=PGP.sig
content-description: This is a digitally signed message part
content-disposition: inline; filename=PGP.sig
content-transfer-encoding: 7bit

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (Darwin)

iD8DBQFCpMLZyUhSUUBViskRAiGaAJ9KMj0SPPRYIzvxcePe6oQjo5/TewCeMB5z
bgQ/Fw29q9ccCGrbVYYhAKY=
=Nw9o
-----END PGP SIGNATURE-----

--Apple-Mail-30-200404062--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LKXTB072677; Mon, 6 Jun 2005 14:20:33 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56LKX46072676; Mon, 6 Jun 2005 14:20:33 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56LKWQi072670 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:20:32 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56LKKAB020378 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 14:20:20 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56LKJJt023853 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 14:20:20 -0700
Date: Mon, 6 Jun 2005 14:18:17 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
cc: ietf-imapext@imc.org, Simon Tyler <simont@gordano.com>, Timo Sirainen <tss@iki.fi>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Message-ID: <Pine.WNT.4.64.0506061404040.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Arnt Gulbrandsen wrote:
>> That's quite a lot to have happen in one interval.
> Really? Consider:

Yes, really.

All of this has to happen in one interval (remember, we have not yet 
defined the magnitude of the interval), *AND* it is necessary to 
accomplish not just two message deliveries but also SELECT/FETCH/close.

On top of that, there is a trivial way to close the race.

This discussion is increasingly sounding like the objections to MIME 
boundary delimiters 15 years ago.  Supposedly, it was impossible to come 
up with a safe algorithm to generate a boundary delimiter that would be 
guaranteed not to occur in the content without having to examine the 
content first.

As it turned out, it wasn't a problem.  Not only is it easy to generate 
boundary delimiters that are impossible in quoted-printable and base64 
content, but it turns out that as long as the delimiter is unlikely enough 
it isn't going to occur in unencoded content for all practical purposes.

An update timestamp works, and lacks the harm caused by breaking a promise 
of the base specification (which, I contend, is out of charter for the 
base specification).

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56L2a2G071181; Mon, 6 Jun 2005 14:02:36 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56L2a3X071180; Mon, 6 Jun 2005 14:02:36 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j56L2VLL071172 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 14:02:32 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id EC7001A718; Mon,  6 Jun 2005 23:02:29 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 66DC71A70F; Mon,  6 Jun 2005 23:02:19 +0200 (CEST)
Message-Id: <o4FMas4k/sHM05LZt7LZ4A.md5@prosecco.oryx.com>
Date: Mon, 6 Jun 2005 23:01:36 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Mark Crispin <MRC@cac.washington.edu>, Timo Sirainen <tss@iki.fi>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
In-Reply-To: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin writes:
> In any case, the timing race that you describe is unlikely.  In order 
> to have a false negative, the following all needs to happen within 
> the same interval of time:
>  1) server gets a new message
>  2) client polls and sees that there is a new message
>  3) client select mailbox
>  4) client processes new message
>  5) client closes mailbox
>  6) server gets a second new message
>
> That's quite a lot to have happen in one interval.

Really? Consider:

1) It's just after lunch, and the entire company is sending mail like crazy.
2) An IMAP Client is told that there's mail in mailbox 1.
3) Client selects that mailbox and retrieves the message
4) Message is displayed
5) Client is told there's mail in mailbox 2.
6) Client selects mailbox 2 and fetches the new message
7) Server stores a new message in mailbox 1

If the client uses e.g. SIEVE to store mail in two mailboxes, and two 
threads are active, I don't see this as uncommon. In a thousand-people 
company I could easily see this happening every day. Normally there's a 
step 8

8) Server tells client there's another new message in mailbox 1

which forces the client to discover the missed message, but if the 
message happened to be the last in the thread, that doesn't work.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KVT64067751; Mon, 6 Jun 2005 13:31:29 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56KVTn8067750; Mon, 6 Jun 2005 13:31:29 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KVTD1067743 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:31:29 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56KVIS2012398 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 13:31:18 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56KVIWJ001971 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 13:31:18 -0700
Date: Mon, 6 Jun 2005 13:29:15 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1118089030.15166.502.camel@hurina>
Message-ID: <Pine.WNT.4.64.0506061317520.3720@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>  <1118084910.15166.468.camel@hurina>  <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com> <1118089030.15166.502.camel@hurina>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Timo Sirainen wrote:
> No, you're assuming that the client is actually interested in reading
> the new message. I think more commonly the timestamp change would just
> cause a STATUS query to request the new number of messages in the
> mailbox (updating a counter in screen - that's why so many clients do
> the STATUS now for all mailboxes).

So, you're saying that, given the very unlikely scenario that you 
describe, it is somehow a problem that, instead of saying "2 new 
messages", it only says "1 new message"?

The fact that there is at least one new message is still shown.

> So, what needs to happen within one
> second is:

s/second/timestamp interval

We haven't defined what the timestamp interval should be.

> 1) new mail is written
> 2) client issues LIST
> 3) client sees new timestamp, issues STATUS command
> 4) server sends reply to STATUS
> 5) another mail is written
> So it really requires just 1,5 RTTs for the problem to occur. The actual
> LIST and STATUS processing and replying is also near instantaneous
> usually.

Even if that was all true (and it isn't -- there is no reason to believe 
that all this will happen in a timestamp interval and you would have to go 
to considerable effect even to contrive it to happen), the only "bad 
effect" is that it says "1 new message" instead of "2 new messages".

It still shows that there are new messages.

Also, there is no way of knowing anything about the message content until 
the mailbox is actually examined.  Perhaps both are spam.

You have not demonstrated any benefit that outweighs the substantial cost.

>> But, assuming that we actually have to fix this (perhaps with a massive
>> speedup of RTTs),
> I think even with your example it could easily happen nowadays. It's
> maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
> pings to my ISP's mail server. That's 125 RTTs within one second.

There's a substantial difference between ping RTTs and RTTs in a higher 
level protocol such as IMAP.

>> we can require that any delivery of new messages write
>> an increased timestamp.
> That would be one implementation-specific way to handle the issue (the
> "timestamp" could be just some 32/64bit increasing number), although I
> don't see how you could actually get it to work with mboxes (and other
> legacy stores) as the mails are usually written using tons of different
> programs that aren't going to be increasing the timestamp.


Not at all.  The mail delivery agent is simply required to ensure that the 
timestamp changes, and if necessary delays delivery so that the new 
message(s) get an updated timestamp.

> Timestamps are also metadata. They just happen to be somewhat more
> easily accessible to you.

s/you/everyone

Even databases have faster access to the operating system's file timestamp 
than they do to a datum contained within the database.

> And I wouldn't want to be polling timestamps either. I think server
> should be notifying IDLEing clients of changes in them automatically.

Then why are you talking about using UIDNEXT at all?

You should be talking about notifications.

The entire discussion at hand has to do with determining which mailboxes 
require synchronization.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KHLfZ066784; Mon, 6 Jun 2005 13:17:21 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56KHLwq066783; Mon, 6 Jun 2005 13:17:21 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56KHGvE066770 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:17:17 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 68EFF1C351D2; Mon,  6 Jun 2005 23:17:10 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Mark Crispin <MRC@CAC.Washington.EDU>
Cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina> <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-hOSZDezp+WU7UNs3iVAh"
Date: Mon, 06 Jun 2005 23:17:10 +0300
Message-Id: <1118089030.15166.502.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-hOSZDezp+WU7UNs3iVAh
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Mon, 2005-06-06 at 12:38 -0700, Mark Crispin wrote:
> On Mon, 6 Jun 2005, Timo Sirainen wrote:
> > Unless you're going to make the "last updated" timestamp accurate up to
> > microseconds, there's a chance that the mailbox has had new messages
> > since your last check if another message was added just after you
> > checked it. Are you going to require that?
>=20
> RTTs in most of today's networking are measured in milliseconds, not=20
> microseconds.  I doubt very much that we will need millisecond level=20
> resolution.

I was going to say milliseconds first, but then thought maybe it's not
enough in future, and besides I think most OSes/filesystems support
either second or microsecond accuracy, but not millisecond.

> In any case, the timing race that you describe is unlikely.  In order to=20
> have a false negative, the following all needs to happen within the same=20
> interval of time:
>   1) server gets a new message
>   2) client polls and sees that there is a new message
>   3) client select mailbox
>   4) client processes new message
>   5) client closes mailbox
>   6) server gets a second new message
>=20
> That's quite a lot to have happen in one interval.

No, you're assuming that the client is actually interested in reading
the new message. I think more commonly the timestamp change would just
cause a STATUS query to request the new number of messages in the
mailbox (updating a counter in screen - that's why so many clients do
the STATUS now for all mailboxes). So, what needs to happen within one
second is:

1) new mail is written
2) client issues LIST
3) client sees new timestamp, issues STATUS command
4) server sends reply to STATUS
5) another mail is written

So it really requires just 1,5 RTTs for the problem to occur. The actual
LIST and STATUS processing and replying is also near instantaneous
usually.

> But, assuming that we actually have to fix this (perhaps with a massive=20
> speedup of RTTs),=20

I think even with your example it could easily happen nowadays. It's
maybe 4-6 RTTs. Using my standard ADSL connection I get about 8msec
pings to my ISP's mail server. That's 125 RTTs within one second.

> we can require that any delivery of new messages write=20
> an increased timestamp. =20

That would be one implementation-specific way to handle the issue (the
"timestamp" could be just some 32/64bit increasing number), although I
don't see how you could actually get it to work with mboxes (and other
legacy stores) as the mails are usually written using tons of different
programs that aren't going to be increasing the timestamp.

> > I think the most important thing people want is to see when new mail
> > appears in different mailboxes. They want to see it as fast as possible
> > and without the notifications getting randomly lost. Getting UIDNEXT
> > value is good enough for that.
>=20
> Polling metadata is absolutely the wrong way to accomplish this.  This is=
,=20
> in effect, what getting the UIDNEXT is.

Timestamps are also metadata. They just happen to be somewhat more
easily accessible to you.

And I wouldn't want to be polling timestamps either. I think server
should be notifying IDLEing clients of changes in them automatically.

--=-hOSZDezp+WU7UNs3iVAh
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCpK9FyUhSUUBViskRAjkwAKCPYdifuRm0A1OdBWH9rwkLpUazbQCeOrPV
LTRAf99KQ3RqC3maQUhjwJI=
=TOA/
-----END PGP SIGNATURE-----

--=-hOSZDezp+WU7UNs3iVAh--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56Jcc8F062635; Mon, 6 Jun 2005 12:38:38 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56JccYc062634; Mon, 6 Jun 2005 12:38:38 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56JcbLl062628 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 12:38:37 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56JcTKx027145 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 6 Jun 2005 12:38:29 -0700
Received: from Shimo-Tomobiki.panda.com ([65.122.177.186]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j56JcPGp023310 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Mon, 6 Jun 2005 12:38:28 -0700
Date: Mon, 6 Jun 2005 12:38:21 -0700
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1118084910.15166.468.camel@hurina>
Message-ID: <Pine.WNT.4.63.0506061223160.1256@Shimo-Tomobiki.panda.com>
References: <06480646801216@PD.gordano.com>  <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU> <1118084910.15166.468.camel@hurina>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Mon, 6 Jun 2005, Timo Sirainen wrote:
> Unless you're going to make the "last updated" timestamp accurate up to
> microseconds, there's a chance that the mailbox has had new messages
> since your last check if another message was added just after you
> checked it. Are you going to require that?

RTTs in most of today's networking are measured in milliseconds, not 
microseconds.  I doubt very much that we will need millisecond level 
resolution.

In any case, the timing race that you describe is unlikely.  In order to 
have a false negative, the following all needs to happen within the same 
interval of time:
  1) server gets a new message
  2) client polls and sees that there is a new message
  3) client select mailbox
  4) client processes new message
  5) client closes mailbox
  6) server gets a second new message

That's quite a lot to have happen in one interval.

But, assuming that we actually have to fix this (perhaps with a massive 
speedup of RTTs), we can require that any delivery of new messages write 
an increased timestamp.  Although that would seem to limit the rate of new 
message delivery, that is ameliorated by MULTIAPPEND; it isn't necessary 
to require a new timestamp per message, as long as multiple message 
delivery is atomic.

Alternatively, we can require that a SELECT/FETCH/close cycle occupy more 
than one interval.

> I think the most important thing people want is to see when new mail
> appears in different mailboxes. They want to see it as fast as possible
> and without the notifications getting randomly lost. Getting UIDNEXT
> value is good enough for that.

Polling metadata is absolutely the wrong way to accomplish this.  This is, 
in effect, what getting the UIDNEXT is.

This may be my personal problem, but in my opinion people who can't wait a 
minute to be told of newly-arrived spam are badly in need of a life...

> Other indicators would be fine as well,
> as long as they are reliable.

Yes.  Reliable notification.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56J8dpn060388; Mon, 6 Jun 2005 12:08:39 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56J8dLN060387; Mon, 6 Jun 2005 12:08:39 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56J8bcn060377 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 12:08:38 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id F03981C351D2; Mon,  6 Jun 2005 22:08:30 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: Mark Crispin <MRC@cac.washington.edu>
Cc: Simon Tyler <simont@gordano.com>, IMAP Extensions WG <ietf-imapext@imc.org>
In-Reply-To: <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com> <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-b+b3Crv/3bMDDok3XqXR"
Date: Mon, 06 Jun 2005 22:08:30 +0300
Message-Id: <1118084910.15166.468.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-b+b3Crv/3bMDDok3XqXR
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Fri, 2005-06-03 at 15:11 -0700, Mark Crispin wrote:
> On Fri, 3 Jun 2005, Simon Tyler wrote:
> > It does not detract from the fact that a client needs UIDVALIDITY on al=
l=20
> > synchronised mailboxes which is why I wanted it added to the LIST=20
> > response.
>=20
> Isn't it really the case that you only need UIDVALIDITY/UIDNEXT on=20
> mailboxes which have a changed "last updated" date/time from the last=20
> time?  Do you really need to be told UIDVALIDITY/UIDNEXT on mailboxes=20
> which are unchanged?  If so, why?

Unless you're going to make the "last updated" timestamp accurate up to
microseconds, there's a chance that the mailbox has had new messages
since your last check if another message was added just after you
checked it. Are you going to require that?

I think the most important thing people want is to see when new mail
appears in different mailboxes. They want to see it as fast as possible
and without the notifications getting randomly lost. Getting UIDNEXT
value is good enough for that. Other indicators would be fine as well,
as long as they are reliable.


--=-b+b3Crv/3bMDDok3XqXR
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCpJ8uyUhSUUBViskRAkxDAJ9q4IfgcNlFg2G7VVulJLA06XOLwACdG8eL
6XHp9ao0udFiI8fFzbOkWxU=
=ObUT
-----END PGP SIGNATURE-----

--=-b+b3Crv/3bMDDok3XqXR--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HrGGv054643; Mon, 6 Jun 2005 10:53:16 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56HrGQt054642; Mon, 6 Jun 2005 10:53:16 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HrFWo054633 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 10:53:15 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n293en1.watson.ibm.com (sp1n293en1.watson.ibm.com [129.34.20.41]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j56HsSOx012302 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:54:28 -0400
Received: from sp1n293en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n293en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j56Hr9s691436 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:53:09 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n293en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j56Hr8H636476 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 13:53:08 -0400
Received: from saturn ([9.2.43.75]) by mars.watson.ibm.com ID IMFd1118080387.27; Mon, 06 Jun 2005 13:53:07 -0400
Date: Mon, 06 Jun 2005 13:53:07 -0400
From: Barry Leiba <leiba@watson.ibm.com>
To: IMAP extensions Mailing List <ietf-imapext@imc.org>
Subject: Re: draft-ietf-imapext-list-extensions-13
Message-ID: <39E0A4D90F19C7E40F6E94B7@saturn.watson.ibm.com>
In-Reply-To: <42A4897C.1080609@oceana.com>
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com> <42A4897C.1080609@oceana.com>
X-Mailer: Mulberry/4.0.0 (Win32)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I've made the editorial changes; thanks.

> - section 6:  should we restrict selection options to be simple atoms
> as we have in this draft, or should we allow for them to be a list as
> well?   I'm thinking that some future selection option might take one
> or more arguments.  For example, the non-standard SCAN extension used
> by Pine and UW-IMAP could be reworked to look something like this:
> 
> x LIST ((SCAN "foo")) "" "bar/%"
> 
> or better yet:
> 
> x LIST ((SCAN "utf-8" "foo")) "" "bar/%"

I have no objection, myself, to extending things further in this way,
but I point out that it *is* a significant additional extension, not
just a simple issue of grammar.  Now there may be many more things to
test in the selection mechanism, for one thing -- it's not just
inspecting whether the mailbox matches some "attribute" (like REMOTE or
SUBSCRIBED), but also whether it meets some other open-ended criteria.
The grammar may also be open-ended (as in your second example; why stop
at two astrings?), which can cause us problems.

Also also, we now have a complication with boolean conditions.  What
might "(SUBSCRIBED (SCAN "foo") (SCAN "bar"))" mean?

We should do this right, while we're doing it.  I'm just not sure, for
this situation, what "right" is.  My sense, from talking with some of
the client authors (mostly with Cyrus), is that this goes beyond what
they need this for.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HbMet053642; Mon, 6 Jun 2005 10:37:22 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j56HbMMr053641; Mon, 6 Jun 2005 10:37:22 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from eagle.oceana.com (eagle.oceana.com [208.17.123.12]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j56HbLHM053632 for <ietf-imapext@imc.org>; Mon, 6 Jun 2005 10:37:21 -0700 (PDT) (envelope-from ken@oceana.com)
Received: from [192.168.10.26] (KEN.oceana.com [192.168.10.26]) by eagle.oceana.com (8.13.2/8.13.2) with ESMTP id j56HbCOU013783; Mon, 6 Jun 2005 13:37:12 -0400 (EDT)
Message-ID: <42A4897C.1080609@oceana.com>
Date: Mon, 06 Jun 2005 13:35:56 -0400
From: Ken Murchison <ken@oceana.com>
Organization: Oceana Matrix Ltd.
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7) Gecko/20040514
X-Accept-Language: en,pdf
MIME-Version: 1.0
To: Barry Leiba <leiba@watson.ibm.com>
CC: IMAP extensions Mailing List <ietf-imapext@imc.org>, Alexey Melnikov <Alexey.Melnikov@isode.com>
Subject: Re: draft-ietf-imapext-list-extensions-13
References: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>
In-Reply-To: <011E4AD30D4DAC128B4561A2@saturn.watson.ibm.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0 (BAYES_00)
X-Scanned-By: MIMEDefang 2.39
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

A couple comments/questions:

- section 3, para 1, sentence 2:  "A mailbox name match a list..." -> "A 
mailbox name matches a list..."

- section 3.3, para 1: "...that the mailbox has st least one..." -> 
"...that the mailbox has at least one..."

- section 5, example 2: response for "Vegetable" mailbox is missing

- section 6:  should we restrict selection options to be simple atoms as 
we have in this draft, or should we allow for them to be a list as well? 
  I'm thinking that some future selection option might take one or more 
arguments.  For example, the non-standard SCAN extension used by Pine 
and UW-IMAP could be reworked to look something like this:

x LIST ((SCAN "foo")) "" "bar/%"

or better yet:

x LIST ((SCAN "utf-8" "foo")) "" "bar/%"


-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j54GrpNb027961; Sat, 4 Jun 2005 09:53:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j54GrpoG027960; Sat, 4 Jun 2005 09:53:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j54GrpTR027945 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 09:53:51 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id BD4FB142275 for <ietf-imapext@imc.org>; Sat,  4 Jun 2005 09:52:37 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 22154-06 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 09:52:37 -0700 (PDT)
Received: from [192.168.1.100] (unknown [198.144.201.116]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 08C0D142274 for <ietf-imapext@imc.org>; Sat,  4 Jun 2005 09:52:36 -0700 (PDT)
Mime-Version: 1.0 (Apple Message framework v622)
To: IMAP Extensions WG <ietf-imapext@imc.org>
Message-Id: <356c352e4f4d75ec6531c13c305ce835@osafoundation.org>
Content-Type: multipart/alternative; boundary=Apple-Mail-6-10380674
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Fwd: Last Call: 'Email Submission Between Independent Networks' to BCP 
Date: Sat, 4 Jun 2005 09:53:37 -0700
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--Apple-Mail-6-10380674
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	format=flowed

FYI

Begin forwarded message:

> From: The IESG <iesg-secretary@ietf.org>
> Date: June 3, 2005 6:47:09 AM PDT
> To: IETF-Announce <ietf-announce@ietf.org>
> Subject: Last Call: 'Email Submission Between Independent Networks' to 
> BCP
> Reply-To: iesg@ietf.org
>
> The IESG has received a request from an individual submitter to 
> consider the
> following document:
>
> - 'Email Submission Between Independent Networks '
>    <draft-hutzler-spamops-04.txt> as a BCP
>
> The IESG plans to make a decision in the next few weeks, and solicits
> final comments on this action.  Please send any comments to the
> iesg@ietf.org or ietf@ietf.org mailing lists by 2005-07-01.
>
> The file can be obtained via
> http://www.ietf.org/internet-drafts/draft-hutzler-spamops-04.txt
>
>
> _______________________________________________
> IETF-Announce mailing list
> IETF-Announce@ietf.org
> https://www1.ietf.org/mailman/listinfo/ietf-announce
>

--Apple-Mail-6-10380674
Content-Transfer-Encoding: 7bit
Content-Type: text/enriched;
	charset=US-ASCII

FYI


Begin forwarded message:


<excerpt><bold><color><param>0000,0000,0000</param>From:
</color></bold>The IESG <<iesg-secretary@ietf.org>

<bold><color><param>0000,0000,0000</param>Date: </color></bold>June 3,
2005 6:47:09 AM PDT

<bold><color><param>0000,0000,0000</param>To:
</color></bold>IETF-Announce <<ietf-announce@ietf.org>

<bold><color><param>0000,0000,0000</param>Subject: </color>Last Call:
'Email Submission Between Independent Networks' to BCP 

<color><param>0000,0000,0000</param>Reply-To:
</color></bold>iesg@ietf.org


The IESG has received a request from an individual submitter to
consider the 

following document:


- 'Email Submission Between Independent Networks '

   <<draft-hutzler-spamops-04.txt> as a BCP


The IESG plans to make a decision in the next few weeks, and solicits

final comments on this action.  Please send any comments to the

iesg@ietf.org or ietf@ietf.org mailing lists by 2005-07-01.


The file can be obtained via

http://www.ietf.org/internet-drafts/draft-hutzler-spamops-04.txt



_______________________________________________

IETF-Announce mailing list

IETF-Announce@ietf.org

https://www1.ietf.org/mailman/listinfo/ietf-announce


</excerpt>
--Apple-Mail-6-10380674--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j549smhk045066; Sat, 4 Jun 2005 02:54:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j549smo9045064; Sat, 4 Jun 2005 02:54:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from foon.sendmail.com (tls.sendmail.com [209.246.26.40]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j549smVS045045 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:48 -0700 (PDT) (envelope-from guenther@sendmail.com)
Received: from lab.smi.sendmail.com ([10.210.100.93]) by foon.sendmail.com (Switch-3.1.7/Switch-3.1.7) with ESMTP id j549sjNs013833 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=OK) for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:45 -0700
X-DomainKeys: Sendmail DomainKeys Filter v0.2.7 foon.sendmail.com j549sjNs013833
DomainKey-Signature: a=rsa-sha1; s=tls; d=sendmail.com; c=nofws; q=dns; b=pw2UXH1vH0IX7hOpsi/HHNXtQPBuQk+i0FLv7u5DbM5nLLaKYIihiOxJ+pEDJ0dtq +oDL786BGXYvWKkO3J6yeKX1x5NvmcKOSQjwNPesJAvGDeXuum+dgzInMOX7JCpspIY X6owaDTcxVaI4q740jDzYAZQRgDytb8feg2JR9M=
Received: from lab.smi.sendmail.com (localhost [127.0.0.1]) by lab.smi.sendmail.com (8.12.11/8.12.11) with ESMTP id j549sjlS015443 for <ietf-imapext@imc.org>; Sat, 4 Jun 2005 02:54:45 -0700 (PDT) (envelope-from guenther@lab.smi.sendmail.com)
Message-Id: <200506040954.j549sjlS015443@lab.smi.sendmail.com>
To: IMAP Extensions WG <ietf-imapext@imc.org>
From: Philip Guenther <guenther+imap@sendmail.com>
Subject: comments on IMAP-ABNF extensions draft -02
Date: Sat, 04 Jun 2005 02:54:45 -0700
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I'm not convinced that this draft is taking the best course for defining
the syntax for extensions.  I see this document as a place to tweak how
the IMAP grammar is expressed to make extensions easier to express,
*without actually actually extending it*.  That is, extensions should be
able express their syntaxes as amending the grammar in RFC 3501 and this
draft by incremental alternative ("=/") only and without replacing
grammar elements defined in RFC 3501 or this draft, preferably without
having to repeat redundant tokens.

And yes, this goal is of course limited by our ability to predict how
commands will be extended, but the key point is what it excludes:
servers are *never* required to accept syntax for extensions that they
don't support**.  This contrasts against the (few) places where the RFC
3501 grammar *requires* clients to unilaterally accept extensible data
with open-ended semantics: body-extension, flag-extension, and
resp-text-code.  As currently written, several of the rules in this
draft have an open-ended syntax that servers would apparently have to
parse, only so that they can reject them with a BAD.

I also disagree with almost all uses of RFC 2119 keywords.  The only two
that make sense to me are both in section 3, in the lead in prose
("Implementations MUST accept...") and in the comment following the
definition of the 'search' rule ("...to SEARCH MUST be registered...").
Of the other nine uses of "MUST", five of them are directed at the
authors of future RFCs and not at clients or servers and are therefore
neither testable nor enforcable: a future RFC that wanted to violate
such a requirement could simply amend/updated _this_ doc to permit
whatever syntax it wanted.***  The other four require servers to reject
with a BAD response uses of extension syntax that they don't
understand, but that can be dropped if the syntax wasn't defined in the
open-ended manner it currently is.

(If a client sends stuff that doesn't meet the syntax of RFC 3501 plus
the RFCs for the extensions supported the server, then the server can do
d**n well whatever it pleases.  If you don't believe that, please
consider how existing servers *do* react when a client is non-compliant
in such ways as including a CRNL in a quoted string as an argument for
a (possibly unknown) command.  GIGO!)


So, for example, consider the append-opt-* rules.  As currently written,
they are an open-ended syntax that servers have to parse, then reject
with a BAD.  Rather than do that, I think this would be better expressed
as a combination of prose and a rule that is *not* referenced on the
left-hand-side of any other rules.  I.e., something like:

     append          = "APPEND" SP mailbox 1*append-message
                         ;; more than one append-message may appear only
                         ;; if MULTIAPPEND [MULTIAPPEND] capability is
                         ;; present

     append-data     = literal / literal8
                         ;; literal8 may appear only if BINARY [BINARY]
                         ;; capability is present.
                         ;; If append-data is extended, it is strongly
                         ;; recommended that the extension match a
                         ;; subset of the tagged-ext rule syntax

     append-ext      =   <for extension only>
                         ;; This rule exists solely to be augmented by
                         ;; extensions via incremental alternative
                         ;; ("=/") rules.  It is strongly recommended
                         ;; that such extensions match a subset of the
                         ;; tagged-ext rule syntax

     append-message  = append-opts SP append-data

     append-opts     = [SP flag-list] [SP date-time] *(SP append-ext)
...
     tagged-ext      = atom SP tagged-ext-val
                         ;; recommended overarching syntax for
                         ;; extensions

     tagged-ext-val  = astring /
                       "(" tagged-ext-val *(SP tagged-ext-val) ")"
                         ;; extensions that follow this general syntax
                         ;; should use nstring instead of astring when
                         ;; appropriate in the context of the extension
<< or should this just say astring / nstring / "(" ...? >>


The key difference between this and the current draft is that where this
syntax permits something beyond RFC 3501, it's because of (and tied to)
a previously published extension, period.  There's no need of a "MUST be
rejected with a BAD" clause for unknown extensions, because they don't
match the syntax, period.


IMHO, the tagged-ext syntax (or something similarly simple) is Good
Enough to be referenced by FETCH, STORE, SELECT, and SEARCH too.  It
certainly appears to cover the already scoped extensions for CONDSTORE,
ANNOTATE, and several others. Using a single syntax for extensions to
commands would make the syntax part of this draft easier to read.  If it
*isn't* sufficient, then I suspect the Right Thing might be to simply
say that extensions should consist of a atom as a tag, followed by
exactly one argument with balanced parenthesis.


On the issue of what this draft should include to make future extensions
easier to express: it seems unfortunate that RFC 3501's response-data
rule includes the leading "*" SP and the trailing CRLF, such that
extensions like ANNOTATEMORE have to duplicate them, ala:

       annotate-data     = "ANNOTATION" SP mailbox SP entry-list
                           ; empty string for mailbox implies
                           ; server annotation.
...
       response-data     /= "*" SP annotate-data CRLF
                           ; adds to original IMAP data responses

instead of just

       response-foo      /= "ANNOTATION" SP mailbox SP entry-list
                           ; empty string for mailbox implies
                           ; server annotation.

RFC 3501 got it right with resp-text-code; can this draft fix things for
response-data?  Perhaps:

   response-data = "*" SP resp-data CRLF
   resp-data     = resp-cond-state / resp-cond-bye / mailbox-data /
		   message-data / capability-data

Okay, so that's not the greatest pair of names for rules, but you get
the point.

**  It could be argued that x-command breaks this rule, but that syntax
    rule isn't actually defined unambiguously.  IMHO, it's appearance in
    the grammer is a mistake.

*** Whether you can place constraints on your successors is a pretty
    subtle issue in the theory of law and rule-making societies.  I
    haven't been participating in the IETF for very long, but to my eye,
    the _only_ constraint that a substantive RFC (i.e., one that isn't
    actually amending the IETF's own procedures) can place on those that
    seek to amend or replace it is the requirement that they (also) be
    on the standards track, or maybe that they be of the same or higher
    level on the standard track.

    If you enjoy pondering these sorts of issues, I highly recommend
    that you track down and read "The Paradox of Self-Amendment" by
    Peter Suber and consider his discussion of "partially
    self-entrenched" laws.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j544xHMj034886; Fri, 3 Jun 2005 21:59:17 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j544xH6x034885; Fri, 3 Jun 2005 21:59:17 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from foon.sendmail.com (tls.sendmail.com [209.246.26.40]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j544xEbq034879 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 21:59:16 -0700 (PDT) (envelope-from guenther@sendmail.com)
Received: from lab.smi.sendmail.com ([10.210.100.93]) by foon.sendmail.com (Switch-3.1.7/Switch-3.1.7) with ESMTP id j544xCLm011391 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=OK); Fri, 3 Jun 2005 21:59:12 -0700
X-DomainKeys: Sendmail DomainKeys Filter v0.2.7 foon.sendmail.com j544xCLm011391
DomainKey-Signature: a=rsa-sha1; s=tls; d=sendmail.com; c=nofws; q=dns; b=FPrxh9mf9qRelwq7myzPVP1u5Z9OGI97eCwnkpq8C9jNaOZL0kt1PukNHbIheJO5W FLIYiKY1sxlCFjBXlhQhYCvtSvNrKA4UHwhqee8c4PKXyWdJAQBOiujnRuGBpsBxkkY 6Ck1D+N6OCfRSYlAd+1O5oQCS0JhyyBH8yL9yBs=
Received: from lab.smi.sendmail.com (localhost [127.0.0.1]) by lab.smi.sendmail.com (8.12.11/8.12.11) with ESMTP id j544xBhm090612; Fri, 3 Jun 2005 21:59:12 -0700 (PDT) (envelope-from guenther@lab.smi.sendmail.com)
Message-Id: <200506040459.j544xBhm090612@lab.smi.sendmail.com>
To: Cyrus Daboo <daboo@isamet.com>
Cc: IMAP Extensions WG <ietf-imapext@imc.org>
From: Philip Guenther <guenther+imap@sendmail.com>
Subject: Re: ANNOTATEMODE uses string rule? 
In-reply-to: <9BCBAA2CA31A73366FE5D0A2@ninevah.local> 
References: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U>  <9BCBAA2CA31A73366FE5D0A2@ninevah.local> 
Date: Fri, 03 Jun 2005 21:59:11 -0700
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Cyrus Daboo <daboo@isamet.com> writes:
>--On June 3, 2005 3:26:14 PM -0700 Mark Crispin <MRC@cac.washington.edu> 
>wrote:
>> I just learned that the ANNOTATEMORE extension uses the string rule,
>> rather than astring or nstring, for several of its rules.
>>
>> This creates an unnecessary new grammar for clients to send and servers
>> to process, both for tokens and the list of tokens.  Please give serious
>> consideration to using astring instead, which will permit the use of
>> existing grammar generators and processors.
...
>Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
>rather than astring. Is there anyone else that feels the same way as Mark?

I basicly agree.  To be precise, I think that 'att-match' and
'entry-match' should be defined as 'list-mailbox', while 'attrib' and
'entry' should be defined as 'astring' (i.e., mirror/reuse the syntax
categories used for mailbox names).


Philip Guenther



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NeNli016880; Fri, 3 Jun 2005 16:40:23 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53NeNl7016879; Fri, 3 Jun 2005 16:40:23 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NeMhw016873 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 16:40:22 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53NeJ8D010118 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 3 Jun 2005 16:40:19 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53NeJic001822 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Fri, 3 Jun 2005 16:40:19 -0700
Date: Fri, 3 Jun 2005 16:38:16 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Cyrus Daboo <daboo@isamet.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: ANNOTATEMODE uses string rule?
In-Reply-To: <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
Message-ID: <Pine.WNT.4.64.0506031636510.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U> <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Fri, 3 Jun 2005, Cyrus Daboo wrote:
> Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
> rather than astring. Is there anyone else that feels the same way as Mark?

I apologize for not noticing this earlier.  My guess is that this use of 
string rather than astring is an influence from ACAP.  However, it is a 
major shift from past IMAP precedent; IMAP uses either astring or nstring.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NG4pp015323; Fri, 3 Jun 2005 16:16:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53NG4Ym015322; Fri, 3 Jun 2005 16:16:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from darius.cyrusoft.com (darius.cyrusoft.com [63.163.82.2]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53NG2Kc015312 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 16:16:02 -0700 (PDT) (envelope-from daboo@isamet.com)
Received: from user-152-3-214-115.wireless.duke.edu (pool-151-201-34-147.pitt.east.verizon.net [151.201.34.147]) (authenticated bits=0) by darius.cyrusoft.com (8.12.9/8.12.9) with ESMTP id j53MoNIC008921 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 3 Jun 2005 18:50:24 -0400
Date: Fri, 03 Jun 2005 19:15:57 -0400
From: Cyrus Daboo <daboo@isamet.com>
To: Mark Crispin <MRC@cac.washington.edu>, IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: ANNOTATEMODE uses string rule?
Message-ID: <9BCBAA2CA31A73366FE5D0A2@ninevah.local>
In-Reply-To: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.EDU>
References:  <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.ED U>
X-Mailer: Mulberry/4.0.0 (Mac OS X)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Spam-Status: No, hits=0.0 tests=none
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Hi Mark,

--On June 3, 2005 3:26:14 PM -0700 Mark Crispin <MRC@cac.washington.edu> 
wrote:

>
> I just learned that the ANNOTATEMORE extension uses the string rule,
> rather than astring or nstring, for several of its rules.
>
> This creates an unnecessary new grammar for clients to send and servers
> to process, both for tokens and the list of tokens.  Please give serious
> consideration to using astring instead, which will permit the use of
> existing grammar generators and processors.
>
> The existing IMAP grammar for clients to send is remarkably small; there
> are only about 10 basic types.  So, adding string (as opposed to astring)
> and list of string is a 20% increase in the client-generated grammar.

Both ANNOTATE and ANNOTATEMORE using string for the entry and attributes 
rather than astring. Is there anyone else that feels the same way as Mark?

-- 
Cyrus Daboo



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MSIFi011800; Fri, 3 Jun 2005 15:28:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53MSIj3011799; Fri, 3 Jun 2005 15:28:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MSHH2011793 for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:18 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MSHoG029581 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:17 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MSGQg022501 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT) for <ietf-imapext@IMC.ORG>; Fri, 3 Jun 2005 15:28:16 -0700
Date: Fri, 3 Jun 2005 15:26:14 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: ANNOTATEMODE uses string rule?
Message-ID: <Pine.WNT.4.64.0506031516290.5000@Tomobiki-Cho.CAC.Washington.EDU>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I just learned that the ANNOTATEMORE extension uses the string rule, 
rather than astring or nstring, for several of its rules.

This creates an unnecessary new grammar for clients to send and servers to 
process, both for tokens and the list of tokens.  Please give serious 
consideration to using astring instead, which will permit the use of 
existing grammar generators and processors.

The existing IMAP grammar for clients to send is remarkably small; there 
are only about 10 basic types.  So, adding string (as opposed to astring) 
and list of string is a 20% increase in the client-generated grammar.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MDxMn011006; Fri, 3 Jun 2005 15:13:59 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53MDxUu011005; Fri, 3 Jun 2005 15:13:59 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53MDwcM010998 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 15:13:58 -0700 (PDT) (envelope-from MRC@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MDrAg026747 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 3 Jun 2005 15:13:54 -0700
Received: from Tomobiki-Cho.CAC.Washington.EDU (tomobiki-cho.cac.washington.edu [128.95.135.58]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j53MDrTF019805 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT); Fri, 3 Jun 2005 15:13:53 -0700
Date: Fri, 3 Jun 2005 15:11:50 -0700 (Pacific Daylight Time)
From: Mark Crispin <MRC@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: IMAP Extensions WG <ietf-imapext@imc.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <06480646801216@PD.gordano.com>
Message-ID: <Pine.WNT.4.64.0506031510090.5000@Tomobiki-Cho.CAC.Washington.EDU>
References: <06480646801216@PD.gordano.com>
Organization: Networks & Distributed Computing
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Fri, 3 Jun 2005, Simon Tyler wrote:
> It does not detract from the fact that a client needs UIDVALIDITY on all 
> synchronised mailboxes which is why I wanted it added to the LIST 
> response.

Isn't it really the case that you only need UIDVALIDITY/UIDNEXT on 
mailboxes which have a changed "last updated" date/time from the last 
time?  Do you really need to be told UIDVALIDITY/UIDNEXT on mailboxes 
which are unchanged?  If so, why?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j53A042Z076791; Fri, 3 Jun 2005 03:00:04 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j53A04bF076790; Fri, 3 Jun 2005 03:00:04 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j53A02mS076765 for <ietf-imapext@imc.org>; Fri, 3 Jun 2005 03:00:04 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 501791A745; Fri,  3 Jun 2005 12:00:00 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 863FC1A6B9; Fri,  3 Jun 2005 11:59:53 +0200 (CEST)
Message-Id: <U1Bslde1KJ+IzN/W9M8M+g.md5@prosecco.oryx.com>
Date: Fri, 3 Jun 2005 11:59:26 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>, Mark Crispin <MRC@CAC.Washington.EDU>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
In-Reply-To: <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Mark Crispin writes:
> Synchronization is difficult.  There is no doubt about that fact.  
> We've spent a decade talking about it with no published 
> specification.  Why don't we roll up our sleeves and talk about how 
> to do it right, rather than consider expensive (to some) 
> half-solutions that are at best "good enough" (to others) but don't 
> quite solve the problem?

The best is the enemy of the good.

A decade has passed, as you say. I think that, after a decade, it's 
about time to lower our aims and be content with a "good enough" 
solution.

My two cents. And I'm not necessarily content with Simon's proposed solution.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536dnba004478; Thu, 2 Jun 2005 23:39:49 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j536dn0Q004477; Thu, 2 Jun 2005 23:39:49 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536dksC004450 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 23:39:46 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id tcuudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:48:06 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id scuudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:48:06 +0100
Date: Fri, 3 Jun 2005 07:48:06 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "IMAP Extensions WG" <ietf-imapext@imc.org>
Message-Id: <06480646801216@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_06480646801217/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_06480646801217/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_06480646801218/PD.gordano.com==_"


--==_06480646801218/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

=20
 =20
(Sorry, resent as I have a horrible habit of not typing the word "not" when=
 I need to!)=20
 =20
CONDSTORE would be a very useful addition for us (although more painful to=
 implement :-( ). Although again, I would like to see some of it's=
 information returned through a LIST response - perhaps the HIGHESTMODSEQ=
 to allow a client to detect changed folders without having to select=
 them.=20
 =20
It does not detract from the fact that a client needs UIDVALIDITY on all=
 synchronised mailboxes which is why I wanted it added to the LIST=
 response.=20
 =20
Simon=20

---- Message from mailto:<mrc@cac.washington.edu Mark Crispin=
 <mrc@cac.washington.edu> at 02-Jun-2005 12:19:44 ------=20


At our WG chair's suggestion (thank you Lisa!), I wish to clarify the=20
following statement:
>  This is the direction of cooperation, rather than trying to boost your=20
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's=20
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into.=20
It isn't just a matter of conscious intent; it is also how it is=
 perceived=20
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST=20
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this=20
does not mean that we have to deploy the first tool is proposed.  We=20
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon=20
last-modified date/time.  We have a start in that direction with=20
CONDSTORE.  Is there a problem with this strawman that I missed?  Of=20
course, it's less painful for my server; but I also believe that it's=20
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.








--==_06480646801218/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><FONT face=3DArial size=3D2>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P>(Sorry, resent as I have a horrible habit of not typing the word "not"=
 when I need to!)</P>
<P>&nbsp;</P>
<P><FONT face=3DArial size=3D2>CONDSTORE would be a very useful addition=
 for us (although more painful to implement :-( ). Although again, I would=
 like to see some of it's information returned through a LIST response -=
 perhaps the HIGHESTMODSEQ to allow a client to detect changed folders=
 without having to select them.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>It does not detract from the fact that a=
 client needs UIDVALIDITY on all synchronised mailboxes which is why I=
 wanted it added to the LIST response.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@cac.washington.edu"=
 target=3D_blank>Mark Crispin &lt;mrc@cac.washington.edu&gt;</A> at=
 02-Jun-2005 12:19:44 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>At our WG chair's suggestion (thank you Lisa!), I=
 wish to clarify the <BR>following statement:<BR>&gt; &nbsp;This is the=
 direction of cooperation, rather than trying to boost your <BR>&gt; own=
 product at the expense of others.<BR><BR>as following:<BR><BR>This is the=
 direction of cooperation, rather than trying to boost one's <BR>own=
 product at the expense of others.<BR><BR>Nobody is doing such boosting=
 now, but it is an easy trap to fall into. <BR>It isn't just a matter of=
 conscious intent; it is also how it is perceived <BR>and the impact upon=
 others.<BR><BR>All of us MUST be sensitive to the problems that others=
 face, and we MUST <BR>NOT be dismissive when someone says "this causes me=
 a problem."<BR><BR>We can all agree that IMAP is lacking in=
 synchronization tools. &nbsp;But this <BR>does not mean that we have to=
 deploy the first tool is proposed. &nbsp;We <BR>should consider=
 alternatives.<BR><BR>I've suggested one such alternative as a strawman: a=
 mechanism based upon <BR>last-modified date/time. &nbsp;We have a start in=
 that direction with <BR>CONDSTORE. &nbsp;Is there a problem with this=
 strawman that I missed? &nbsp;Of <BR>course, it's less painful for my=
 server; but I also believe that it's <BR>technically superior.<BR><BR>--=
 Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science does not emerge=
 from voting, party politics, or public debate.<BR>Si vis pacem, para=
 bellum.<BR><BR><BR><BR></BLOCKQUOTE></FONT><BR><BR></FONT></body></html>

--==_06480646801218/PD.gordano.com==_--
--==_06480646801217/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536Uosf001265; Thu, 2 Jun 2005 23:30:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j536Uoju001264; Thu, 2 Jun 2005 23:30:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j536UlcO001228 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 23:30:48 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id dauudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:39:05 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id cauudaaa for ietf-imapext@imc.org; Fri, 3 Jun 2005 07:39:05 +0100
Date: Fri, 3 Jun 2005 07:39:05 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "IMAP Extensions WG" <ietf-imapext@imc.org>
Message-Id: <06390556201197@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
In-Reply-To: <Pine.OSX.4.63.0506021204100.266@pangtzu.panda.com>
References: <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com> <1d61bc9cc57fa16d7c700383fe6e7b98@osafoundation.org> <Pine.OSX.4.63.0506021148170.266@pangtzu.panda.com> <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_06390556201198/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_06390556201198/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_06390556201199/PD.gordano.com==_"


--==_06390556201199/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
CONDSTORE would be a very useful addition for us (although more painful to=
 implement :-( ). Although again, I would like to see some of it's=
 information returned through a LIST response - perhaps the HIGHESTMODSEQ=
 to allow a client to detect changed folders without having to select=
 them.=20
 =20
It does detract from the fact that a client needs UIDVALIDITY on all=
 synchronised mailboxes which is why I wanted it added to the LIST=
 response.=20
 =20
Simon=20

---- Message from mailto:<mrc@cac.washington.edu Mark Crispin=
 <mrc@cac.washington.edu> at 02-Jun-2005 12:19:44 ------=20


At our WG chair's suggestion (thank you Lisa!), I wish to clarify the=20
following statement:
>  This is the direction of cooperation, rather than trying to boost your=20
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's=20
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into.=20
It isn't just a matter of conscious intent; it is also how it is=
 perceived=20
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST=20
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this=20
does not mean that we have to deploy the first tool is proposed.  We=20
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon=20
last-modified date/time.  We have a start in that direction with=20
CONDSTORE.  Is there a problem with this strawman that I missed?  Of=20
course, it's less painful for my server; but I also believe that it's=20
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.






--==_06390556201199/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>CONDSTORE would be a very useful addition=
 for us (although more painful to implement :-( ). Although again, I would=
 like to see some of it's information returned through a LIST response -=
 perhaps the HIGHESTMODSEQ to allow a client to detect changed folders=
 without having to select them.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>It does detract from the fact that a client=
 needs UIDVALIDITY on all synchronised mailboxes which is why I wanted it=
 added to the LIST response.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@cac.washington.edu">Mark=
 Crispin &lt;mrc@cac.washington.edu&gt;</A> at 02-Jun-2005 12:19:44=
 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>At our WG chair's suggestion (thank you Lisa!), I=
 wish to clarify the <BR>following statement:<BR>&gt; &nbsp;This is the=
 direction of cooperation, rather than trying to boost your <BR>&gt; own=
 product at the expense of others.<BR><BR>as following:<BR><BR>This is the=
 direction of cooperation, rather than trying to boost one's <BR>own=
 product at the expense of others.<BR><BR>Nobody is doing such boosting=
 now, but it is an easy trap to fall into. <BR>It isn't just a matter of=
 conscious intent; it is also how it is perceived <BR>and the impact upon=
 others.<BR><BR>All of us MUST be sensitive to the problems that others=
 face, and we MUST <BR>NOT be dismissive when someone says "this causes me=
 a problem."<BR><BR>We can all agree that IMAP is lacking in=
 synchronization tools. &nbsp;But this <BR>does not mean that we have to=
 deploy the first tool is proposed. &nbsp;We <BR>should consider=
 alternatives.<BR><BR>I've suggested one such alternative as a strawman: a=
 mechanism based upon <BR>last-modified date/time. &nbsp;We have a start in=
 that direction with <BR>CONDSTORE. &nbsp;Is there a problem with this=
 strawman that I missed? &nbsp;Of <BR>course, it's less painful for my=
 server; but I also believe that it's <BR>technically superior.<BR><BR>--=
 Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science does not emerge=
 from voting, party politics, or public debate.<BR>Si vis pacem, para=
 bellum.<BR><BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_06390556201199/PD.gordano.com==_--
--==_06390556201198/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52LxBpv034445; Thu, 2 Jun 2005 14:59:11 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52LxB9f034444; Thu, 2 Jun 2005 14:59:11 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52Lx9E4034435 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 14:59:10 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 0C0771A75D; Thu,  2 Jun 2005 23:59:06 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id A9FBE1A6E0; Thu,  2 Jun 2005 23:59:00 +0200 (CEST)
Message-Id: <Jqf492O+yMCM5gT0I8WDlg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 23:58:37 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com> <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com> <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com> <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52LxAE4034438
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> So your calculation is a best case calculation where all the STATUS 
> commands are pipelined and mine is the worst case calculation where 
> none of them are.   The answer is likely to be somewhere in between.

IMO, it's likely to be close enough to full pipelining that the 
roundtrip cost becomes insignificant compared to the traffic cost.

...
> So it is your belief that it is easy to do the command batching than 
> to code fallbacks?

Once the cost of testing is added: Yes, I think so, as a rule of thumb. 
Not always. It depends on how many additional test cases the fallback 
creates, etc.

> Perhaps.   I expect there to be widely variant opinions on that as 
> well, but as we are talking about implementation again, it just 
> doesn't seem worth the discussion.

We can't really keep implementation discussion completely away. Unless 
we want to end up like CORBA, of course.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52JJp4i021441; Thu, 2 Jun 2005 12:19:51 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52JJpRX021440; Thu, 2 Jun 2005 12:19:51 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52JJoGb021433 for <ietf-imapext@IMC.ORG>; Thu, 2 Jun 2005 12:19:50 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52JJlS9028919 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 12:19:48 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52JJjYt018310 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 12:19:47 -0700
Date: Thu, 2 Jun 2005 12:19:44 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: IMAP Extensions WG <ietf-imapext@imc.org>
cc: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
Message-ID: <Pine.OSX.4.63.0506021204100.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com> <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com> <1d61bc9cc57fa16d7c700383fe6e7b98@osafoundation.org> <Pine.OSX.4.63.0506021148170.266@pangtzu.panda.com> <fddec3aeebdd030ddd1e120a8ad87364@osafoundation.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

At our WG chair's suggestion (thank you Lisa!), I wish to clarify the 
following statement:
>  This is the direction of cooperation, rather than trying to boost your 
> own product at the expense of others.

as following:

This is the direction of cooperation, rather than trying to boost one's 
own product at the expense of others.

Nobody is doing such boosting now, but it is an easy trap to fall into. 
It isn't just a matter of conscious intent; it is also how it is perceived 
and the impact upon others.

All of us MUST be sensitive to the problems that others face, and we MUST 
NOT be dismissive when someone says "this causes me a problem."

We can all agree that IMAP is lacking in synchronization tools.  But this 
does not mean that we have to deploy the first tool is proposed.  We 
should consider alternatives.

I've suggested one such alternative as a strawman: a mechanism based upon 
last-modified date/time.  We have a start in that direction with 
CONDSTORE.  Is there a problem with this strawman that I missed?  Of 
course, it's less painful for my server; but I also believe that it's 
technically superior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IgeIr018107; Thu, 2 Jun 2005 11:42:40 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Ige46018106; Thu, 2 Jun 2005 11:42:40 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout5.cac.washington.edu (mxout5.cac.washington.edu [140.142.32.135]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IgepF018100 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:42:40 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout5.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Igd0U008082 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:42:39 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Igc9r009519 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:42:39 -0700
Date: Thu, 2 Jun 2005 11:42:37 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021137450.266@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> You're acting like this is the first time that this
> has been asked for and it's not.

With all due respect Steve, you also are sounding like a broken record. 
It isn't as if you've been exactly neutral with respect to your own 
products either.

Can we please leave this baiting aside, and discuss the real issues?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Ia7LF017717; Thu, 2 Jun 2005 11:36:07 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Ia7bL017716; Thu, 2 Jun 2005 11:36:07 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Ia6wm017710 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:36:06 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Ia3ZA016385 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:36:03 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Ia1ee007861 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:36:03 -0700
Date: Thu, 2 Jun 2005 11:36:01 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021126490.266@pangtzu.panda.com>
References: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>    <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>    <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com> <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> The STATUS command was added as a means to acquire folder metadata.
> The reason it was added was because a large number of client implementors
> wanted to markup their displays with this information because their
> customers wanted to see it.

That was not why STATUS was added.  Remember that I'm the person who came 
up with it.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IQERu017058; Thu, 2 Jun 2005 11:26:14 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52IQEYk017057; Thu, 2 Jun 2005 11:26:14 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IQEOS017050 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:26:14 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IQApZ009592 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:26:10 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IQ8uq025304 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:26:09 -0700
Date: Thu, 2 Jun 2005 11:26:07 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021120090.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> Hogwash.  If you are writing a general purpose client then you have to
> implement fallbacks for the various optional features that are already in
> play and it isn't that hard.

Not so.  Clients exist today that depend upon an optional feature, and 
declare the server to be broken that lack that optional feature.

> This particular requirement is an attempt to make MUCH more difficult
> tasks in the client (synchronization, display markup, user notification)
> easier and perform better.

Except that the proposal in question doesn't really accomplish that 
directly.  It is a possible solution that may be acceptable in some 
circumstances, but it dances around the overall problem (for example, it 
can't detect external deletions).

I have a general dislike of quick and dirty "solutions" that don't solve 
the general problem and impose potentially high costs.

Synchronization is difficult.  There is no doubt about that fact.  We've 
spent a decade talking about it with no published specification.  Why 
don't we roll up our sleeves and talk about how to do it right, rather 
than consider expensive (to some) half-solutions that are at best "good 
enough" (to others) but don't quite solve the problem?

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IJiP1016688; Thu, 2 Jun 2005 11:19:44 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52IJiq4016687; Thu, 2 Jun 2005 11:19:44 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52IJhV4016680 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:19:43 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IJehP015155 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 11:19:41 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52IJdaO023835 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 11:19:40 -0700
Date: Thu, 2 Jun 2005 11:19:39 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Steve Hole <steve.hole@MessagingDirect.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
Message-ID: <Pine.OSX.4.63.0506021058300.266@pangtzu.panda.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Steve Hole wrote:
> If the feature is useful, then it will be implemented by clients and
> servers that wish to implement it.   If this causes feature pressure on
> implementations, then too bad.   The feature will either have merit and be
> useful or it won't.   It is simply one of a set of features that
> contribute to the selection of one implementation over another.   Denying
> features in optional services just because some servers can't implement it
> efficiently, but some clearly can, it ludicrous.   It smacks of software
> socialism :-)

The contrary argument is that certain other protocols are littered with 
pie-in-the-sky features that someone added, but were never widely 
implemented.  The result is that, when reading the protocol specification, 
it is impossible to determine what features need to be implemented and 
what features can be omitted.

This problem has been with us since the beginning; consider TELNET and 
FTP.  TELNET and FTP are full of crud that someone thought was important 
30 years ago, yet are inadequate to do today's needs.  I've heard all 
sorts of horror stories about the FTP meetings 30+ years ago.

This has an additional harmful effect.  There is at least one server 
vendor which believes that SEARCH is optional to implement on the server. 
This vendor realized that nobody implements every aspect of every IMAP 
document, and assumed that this gave him the discretion to apply his own 
razor...to the base specification.

Feature bloat is not a good thing.  It is something to be avoided at all 
cost.  Every new feature proposal must justify its own existance.  It 
can't just be a good idea; it has to be an overwhelmingly good idea above 
and beyond any harm that it will cause.

It has already been pointed out that this proposed LIST extension is 
really a way to determine, for synchronization purposes, what has changed 
since the last synchronization.  Rather than take the first idea that 
comes along for how to do this, we should carefully consider how best to 
do this.

One thing that immediately comes to mind are tools that use a "last 
changed" date/time.  That is basically an extension of the \Marked and 
\UnMarked functionality, only now we're talking about synchronizing a 
disconnected client rather than showing an online client what's new.

Not only does this have less potential cost, it may also lead in new 
desirable directions.  For example, maybe LIST could have a means to 
filter mailboxes based upon "last changed" date/time.

Thinking in the direction of "how to do new things that are good for 
everybody" would be more productive.  This is the direction of 
cooperation, rather than trying to boost your own product at the expense 
of others.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Hte5C014979; Thu, 2 Jun 2005 10:55:40 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HtemF014978; Thu, 2 Jun 2005 10:55:40 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout1.cac.washington.edu (mxout1.cac.washington.edu [140.142.32.134]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HtdjG014971 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:55:39 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout1.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52HtcTn005075 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 10:55:39 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52HtS6g029860 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 10:55:38 -0700
Date: Thu, 2 Jun 2005 10:55:27 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Barry Leiba <leiba@watson.ibm.com>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
Message-ID: <Pine.OSX.4.63.0506021044240.266@pangtzu.panda.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <429F24BC.6000908@watson.ibm.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Barry Leiba wrote:
> But the protocol MUST do <y>, though THAT is expensive to do with some legacy 
> message stores (that he doesn't care about).  (An example of <y> that comes 
> to mind is "return an exact length for RFC822.SIZE, though clients can easily 
> work around that.)  Again, let's not spend a lot of time talking about 
> RFC822.SIZE now -- I want to make the point that we all bring our 
> implementation biases with us.

For the record, if I had known that there was an issue for any potential 
server with lengths, I would have designed IMAP differently.

When the requirements for RFC822.SIZE were stated, I did not have a clue 
that it could possibly be expensive for Exchange.  Exchange did not exist 
at that time, and Microsoft was a relatively small company that produced 
DOS and hired programmers based on how well they came up with an algorithm 
to do tic-tac-toe with bit arithmetic.

The client can NOT easily work around non-exact lengths.  The purpose of 
exact lengths is for the client to allocate space; the same is the case 
with literals.  If the client does partial fetching, it MUST have exact 
lengths.

I knew about UNIX's use of newline-only newlines, but newline conversion 
had a long history in the network and it was asserted repeatedly that this 
was trivial.  The implementation complaint from the UNIX guys was that 
this IMAP proposal of mine provided for multiple simultaneous access which 
was "impossible" to do on UNIX.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HJ1od010842; Thu, 2 Jun 2005 10:19:01 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HJ19x010841; Thu, 2 Jun 2005 10:19:01 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HJ0WP010825 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:19:00 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 11:20:43 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 11:18:54 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com>
References: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com> <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>    <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>    <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>    <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
Message-ID: <EXECMAIL.20050602111854.F@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52HJ0WP010828
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 18:49:28 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> No. Of sending all the STATUS commands at the same time. Perfectly legal 
> in IMAP, works with all servers:
> 
> C: a STATUS asfd (UIDNEXT UIDVALIDITY)
> C: b STATUS qwre (UIDNEXT UIDVALIDITY)
> C: c STATUS zxvc (UIDNEXT UIDVALIDITY)
> S: * STATUS asfd (UIDNEXT 124 UIDVALIDITY 43)
> S: a OK
> S: * STATUS qwre (UIDNEXT 1 UIDVALIDITY 1)
> S: b OK
> S: * STATUS zxvc (UIDNEXT 2 UIDVALIDITY 3)
> S: c OK

So your calculation is a best case calculation where all the STATUS 
commands are pipelined and mine is the worst case calculation where none 
of them are.   The answer is likely to be somewhere in between.   Many 
clients will batch to some point, so I'll grant that.   They probably 
should all do that.   I wonder how many actually do.   
 
> More difficult to implement in a client than an RPC-like interface, but 
> IMO less than twice as hard, so quite likely simpler than implementing 
> both with and without an extension.

So it is your belief that it is easy to do the command batching than to 
code fallbacks?   Perhaps.   I expect there to be widely variant opinions 
on that as well, but as we are talking about implementation again, it just
doesn't seem worth the discussion.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HAI8v009984; Thu, 2 Jun 2005 10:10:18 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52HAITn009983; Thu, 2 Jun 2005 10:10:18 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52HAHVc009971 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 10:10:17 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 11:11:57 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 11:10:08 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com>
References: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>    <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>    <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Message-ID: <EXECMAIL.20050602111008.E@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52HAHVc009978
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 18:05:46 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> One excellent way to placate me is to explain how the proposal outlined 
> in this thread relates to this:
> 
> > This particular requirement is an attempt to make MUCH more difficult 
> > tasks in the client (synchronization, display markup, user 
> > notification) easier and perform better.

The STATUS command was added as a means to acquire folder metadata.   
The reason it was added was because a large number of client implementors 
wanted to markup their displays with this information because their 
customers wanted to see it.    At that time there was discussion on 
returning the metadata as part of the LIST response.   It was rejected 
because LIST is mandatory to implement in the base specification and MUST 
be fast.    Adding the requirement for returning metadata to a base 
command puts an unreasonable load on implementations that may not be able 
to bear it.

That doesn't obviate the requirement for the metadata ... thus the STATUS 
command.

Roundtrips are painful from a client standpoint.    This is especially 
true when your client is in fact a service -- ie. a J2EE or Web Service.  
If you need the information for your display then you currently have to 
get it by issuing multiple STATUS commands.   This is mandated by the 
protocol.    It doesn't matter how fast your server is at answering the 
question, because the roundtrips and associated IO overwhelm all other 
performance issues.

If you optionally add the metadata to the LIST response you eliminate the 
roundtrips.   Then, to get the expected performance, your server must be 
able perform the requested operation efficiently.   Many can.   If you are
building service class infrastructure, then you will select one that does.

But, most importantly, you have eliminated the protocol inefficiency that 
causes services delays regardless of how good or bad your store 
implementation is.

To summarize.   The requirement has some merit.   It was a clear 
performance and feature win in a proprietary form for us both with service 
clients and desktop clients.   It is just one of several useful features 
that have been discussed wrt LIST extensions.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Gpd75008560; Thu, 2 Jun 2005 09:51:39 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Gpd05008559; Thu, 2 Jun 2005 09:51:39 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52GpUcI008545 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:51:39 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 10:53:12 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 10:51:24 -0700
To: Mark Crispin <mrc@cac.washington.edu>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com>
References: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
Message-ID: <EXECMAIL.20050602105124.D@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52GpdcI008554
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 06:14:03 -0700 (PDT) Mark Crispin 
<mrc@cac.washington.edu> wrote:

> 
> On Thu, 2 Jun 2005, Simon Tyler wrote:
> > Whether or not a server stores UIDNEXT or calculates it is a server 
> > implementation decision. It is irrelevant when designing a protocol.
> 
> This is completely wrong.

No he's not.   His request is just difficult for you to implement.

> For better or worse, Internet protocols are molded to accomodate existing 
> implementations.  You will receive strong pushback on any proposed 
> protocol or protocol enhancement whose impact is perceived as harmful to 
> one or more implementations.  Complaints that "it's a poor specification" 
> if your demands are not fufilled will fall on deaf ears.

This is an argument that has been seriously overused and has prevented 
some very useful features from being added.   Make it optional and 
don't implement it.   If people *really* want it, then you will have to 
figure out a way to make it work.   If they don't, then continue to not 
implement it.
 
> In fact, what you want is to have IMAP molded to accomodate *your* client 
> implementation.  Rather than fix your client so it doesn't do something as 
> ill-conceived as a STATUS on every LISTed mailbox, you want to change the 
> protocol so the server does the stupid thing for you automatically, all to 
> save your client some RTTs.  It doesn't address the underlying problem.

Also not true Mark.   You're acting like this is the first time that this 
has been asked for and it's not.   In fact it has come up repeatedly over 
the years.   It is a legitmate client requirement.   Yes, it requires more
work from the server.    That's why it is called a "server".

> The entire point of the Internet protocol architecture is a set of 
> interoperable protocols which can be implemented across a wide variety of 
> systems, including systems with legacy infrastructure.

Which definitely applies to base protocols but does not necessarily apply 
to extensions.  Extensions are optional.

I believe that the problem you've had with the LIST extensions all along 
is that you believe client implementors will implement them -- in droves.
Because they are useful.  And that will put pressure on server 
implementations that have difficulty in providing these services.

> The design of protocols for a perfect world, in which you never need to 
> concern yourself with the other guy's problems, is not the realm of 
> Internet.  It is the realm of proprietary protocols and systems.

> In the Internet world, interoperability is king, and everything else must 
> submit to the demands of that king.

And I submit that you are currently hurting interoperability wrt to LIST 
extensions.   The metadata features are so useful that many people are 
implementing proprietary extensions none of which have a chance of 
interoperating.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Go0eW008430; Thu, 2 Jun 2005 09:50:00 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Go0wB008429; Thu, 2 Jun 2005 09:50:00 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52GnwoW008423 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:49:59 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 30C341A74F; Thu,  2 Jun 2005 18:49:58 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id BF55C1A707; Thu,  2 Jun 2005 18:49:51 +0200 (CEST)
Message-Id: <K9ZXh2aufUDwaYxEIvyeEg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 18:49:28 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com> <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> On Thu, 2 Jun 2005 10:12:42 +0200 Arnt Gulbrandsen
> <arnt@gulbrandsen.priv.no> wrote:
>> Alternative 1 has a cost of two roundtrips, plus traffic costs for 
>> the LIST responses and for the responses to STATUS, plus CPU costs 
>> for LIST, STATUS and quite a bit of parsing.
>
> Almost right.   Alternative 1 has a cost of 1 roundtrip for the LIST 
> and 1 roundtrip for each STATUS per interesting mailbox.  Perhaps you 
> are thinking of a STATUS extension that accepts multiple mailboxes as 
> arguments?

No. Of sending all the STATUS commands at the same time. Perfectly legal 
in IMAP, works with all servers:

C: a STATUS asfd (UIDNEXT UIDVALIDITY)
C: b STATUS qwre (UIDNEXT UIDVALIDITY)
C: c STATUS zxvc (UIDNEXT UIDVALIDITY)
S: * STATUS asfd (UIDNEXT 124 UIDVALIDITY 43)
S: a OK
S: * STATUS qwre (UIDNEXT 1 UIDVALIDITY 1)
S: b OK
S: * STATUS zxvc (UIDNEXT 2 UIDVALIDITY 3)
S: c OK

More difficult to implement in a client than an RPC-like interface, but 
IMO less than twice as hard, so quite likely simpler than implementing 
both with and without an extension.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52GbiP0007212; Thu, 2 Jun 2005 09:37:44 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52GbiEk007211; Thu, 2 Jun 2005 09:37:44 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Gbhmu007203 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:37:43 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 10:39:20 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 10:37:32 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>
References: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>    <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Message-ID: <EXECMAIL.20050602103732.C@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52Gbhmu007205
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 10:12:42 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> 
> Let me step back. AFAICT, there are two approaches to this:
> 
> 1. Client sends LIST. When an interesting LIST response arrives, it 
> sends a STATUS command. As the STATUS commands' responses arrive, the 
> client has the information it wants.
> 
> 2. Client sends LIST with some extension. For each mailbox matching the 
> supplied pattern, the server does the same work that STATUS would do, 
> and sends the result in a LIST response.
> 
> Do I understand it correctly?
> 
> Alternative 1 has a cost of two roundtrips, plus traffic costs for the 
> LIST responses and for the responses to STATUS, plus CPU costs for 
> LIST, STATUS and quite a bit of parsing.

Almost right.   Alternative 1 has a cost of 1 roundtrip for the LIST and
1 roundtrip for each STATUS per interesting mailbox.  Perhaps you are 
thinking of a STATUS extension that accepts multiple mailboxes as 
arguments?

> Alternative 2 has a cost of one roundtrip, plus traffic for the LIST 
> responses, plus CPU cost for LIST including STATUS.
>
> If we assume that there's 100 LIST responses and 10 interesting 
> mailboxes, then 1 costs one roundtrip more and 2 has 90 unnecessary 
> STATUS processings and 90 unnecessary transmitted UIDNEXT/UIDVALIDITY 
> pairs. (And of course, a client that implements 2 has to implement 1 as 
> well, or else it won't work with plain servers.)

Based on the roundtrip calculation correction:

1 costs 1 LIST + 10 STATUS roundtrips.
2 costs 1 LIST roundtrip.

I believe the response processing to be roughly equivalent in both cases.

>From a client point of view, avoiding the 10 additional roundtrips is very
useful -- especially in a highly latent situation.

> Hm. I suppose I must misunderstand terribly, because I cannot see any 
> appeal in 2 at all.

You've missed the roundtrip cost of multiple STATUS commands.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G6Fb7004555; Thu, 2 Jun 2005 09:06:15 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52G6F4K004554; Thu, 2 Jun 2005 09:06:15 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52G6ESx004545 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:06:14 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id F34BD1A753; Thu,  2 Jun 2005 18:06:12 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id A83BA1A6F5; Thu,  2 Jun 2005 18:06:06 +0200 (CEST)
Message-Id: <v/lUFQVFRh9J0a1GfTG0mw.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 18:05:46 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> Sorry Arnt, it appears that I'm picking on your messages in 
> particular,  and I'm not really.

I am much offended. You must now placate me. ;)

One excellent way to placate me is to explain how the proposal outlined 
in this thread relates to this:

> This particular requirement is an attempt to make MUCH more difficult 
> tasks in the client (synchronization, display markup, user 
> notification) easier and perform better.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G1cHn003828; Thu, 2 Jun 2005 09:01:38 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52G1cth003827; Thu, 2 Jun 2005 09:01:38 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52G1YbD003818 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:01:35 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id insudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 17:09:45 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3335/SQ2499.01.1aa407eb) with ESMTP id hnsudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 17:09:45 +0100
Date: Thu, 2 Jun 2005 17:09:45 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: <ietf-imapext@imc.org>
Message-Id: <16094559300656@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3335
X-Originating-IP: [62.172.232.189]
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
References: <08173804605816@PD.gordano.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_16094560900657/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_16094560900657/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_16094560900658/PD.gordano.com==_"


--==_16094560900658/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I do seem to have opened a can of worms here. Apologies to all concerned -=
 I'm rapidly learning how this process works.=20
 =20
I agree with pretty much all of what you have said. =20
 =20
The only points I would like to make are:=20
 =20
1: In my experience IMAP clients appear to do a LIST followed by a STATUS=
 for every folder. It may be nasty, it may be expensive but they already do=
 it. =20
 =20
2. The idea that users wish to only synchronise a subset of folders is not=
 correct. Again, in my experience most people use IMAP to obtain an offline=
 store whilst retaining the files server side for other uses (WebMail, home=
 access, whatever). Obviously there are exceptions to this - more=
 sophistaticated users\clients or mobile device access but these are the=
 exception rather than the rule. If you are in the business of=
 synchronising a folder then you need UIDVALIDITY so you have to do a=
 STATUS. If you are synchronising all folders then you have to do a STATUS=
 on all folders.=20
 =20
If we come up with extensions that are useful to client developers then=
 they will be pressured to use them. I work for a company that primarily=
 writes servers but we need a client now. The real driver here is to make=
 the end user experience as quick and as easy as possible. If one client=
 works better than another then the worse client will have to improve or=
 die.=20
 =20
Simon=20

---- Message from mailto:<leiba@watson.ibm.com Barry Leiba=
 <leiba@watson.ibm.com> at 02-Jun-2005 11:24:44 ------=20


I don't agree with Mark on every aspect of the IMAP protocol, nor even=20
on every aspect of protocol design in general.  I agree with him on this=20
point.

First, one thing (and I don't mean to get into an extended discussion=20
here; I just want to make a point and move on): there is some=20
inconsistency in Mark's stand.  We can't have the protocol do <x>=20
because it is expensive to do with some legacy message stores (at least=20
one of which he needs to implement for).  But the protocol MUST do <y>,=20
though THAT is expensive to do with some legacy message stores (that he=20
doesn't care about).  (An example of <y> that comes to mind is "return=20
an exact length for RFC822.SIZE, though clients can easily work around=20
that.)  Again, let's not spend a lot of time talking about RFC822.SIZE=20
now -- I want to make the point that we all bring our implementation=20
biases with us.

I also think Simon is right that it's easy to add an extension that's=20
cheap to implement in some (most?) servers, and that it's nice to give=20
that optimization to the client.  We did talk, in fact, about adding=20
STATUS items to the LISTEXT response selections.  I would side with=20
Simon here IF (and this is the key point) we could rely on clients to=20
ask for the information when they need it, and to interpret the server's=20
lack of support for it as an indication that this is an expensive thing=20
to do on this server... and to use a different approach.

What I'm saying is that the client would say "list all the mailboxes,=20
and while you're at it give me <this other information>".  If the server=20
didn't not return the other information, the client would understand=20
that it's because the server can't do that efficiently, and the client=20
would proceed *without* the information, using a different method of=20
processing.

We can't rely on that.  What will, instead, happen, is that clients=20
(perhaps not yours, Simon, but others) will fall back to extracting the=20
information the old, slow way, incurring all the overhead.  Which they=20
can do anyway, of course, but...

My usual approach is to say that if we can provide clients with=20
information they want in a way that the server might be able to=20
optimize, the server might as well support it because the clients will=20
get it anyway.  In this case, though, I agree with Mark that the client=20
does NOT need the information always, and should instead be more=20
selective about obtaining it.  So the remainder of the "but..." sentence=20
above is, "...but it's best not to encourage behaviour that is sometimes=20
bad, and generally unnecessary."

I actually LIKE the idea of the server's having an option of responding=20
"this is a very expensive operation", and wish we could rely on clients=20
to pay attention to that.  Alas, we can't.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam






--==_16094560900658/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I do seem to have opened a can of worms=
 here. Apologies to all concerned - I'm rapidly learning how this process=
 works.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I agree with pretty much all of what you=
 have said.&nbsp;</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>The only points I would like to make=
 are:</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>1:&nbsp;In my experience IMAP clients appear=
 to do a LIST followed by a STATUS for every folder. It may be nasty, it=
 may be expensive but they already do it. </FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>2. The idea that users wish to only=
 synchronise a subset of folders is not correct. Again, in my experience=
 most people use IMAP to obtain an offline store whilst retaining the files=
 server side for other uses (WebMail, home access, whatever). Obviously=
 there are exceptions to this - more sophistaticated users\clients or=
 mobile device access but these are the exception rather than the rule. If=
 you are in the business of synchronising a folder then you need=
 UIDVALIDITY so you have to do a STATUS. If you are synchronising all=
 folders then you have to do a STATUS on all folders.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>If we come up with extensions that are=
 useful to client developers then they will be pressured to use them. I=
 work for a company that primarily writes servers but we need a client now.=
 The real driver here is to make the end user experience as quick and as=
 easy as possible. If one client works better than another then the worse=
 client will have to improve or die.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<leiba@watson.ibm.com">Barry=
 Leiba &lt;leiba@watson.ibm.com&gt;</A> at 02-Jun-2005 11:24:44 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid"><BR>I don't agree with Mark on every aspect of the IMAP=
 protocol, nor even <BR>on every aspect of protocol design in general.=
 &nbsp;I agree with him on this <BR>point.<BR><BR>First, one thing (and I=
 don't mean to get into an extended discussion <BR>here; I just want to=
 make a point and move on): there is some <BR>inconsistency in Mark's=
 stand. &nbsp;We can't have the protocol do &lt;x&gt; <BR>because it is=
 expensive to do with some legacy message stores (at least <BR>one of which=
 he needs to implement for). &nbsp;But the protocol MUST do &lt;y&gt;,=
 <BR>though THAT is expensive to do with some legacy message stores (that=
 he <BR>doesn't care about). &nbsp;(An example of &lt;y&gt; that comes to=
 mind is "return <BR>an exact length for RFC822.SIZE, though clients can=
 easily work around <BR>that.) &nbsp;Again, let's not spend a lot of time=
 talking about RFC822.SIZE <BR>now -- I want to make the point that we all=
 bring our implementation <BR>biases with us.<BR><BR>I also think Simon is=
 right that it's easy to add an extension that's <BR>cheap to implement in=
 some (most?) servers, and that it's nice to give <BR>that optimization to=
 the client. &nbsp;We did talk, in fact, about adding <BR>STATUS items to=
 the LISTEXT response selections. &nbsp;I would side with <BR>Simon here IF=
 (and this is the key point) we could rely on clients to <BR>ask for the=
 information when they need it, and to interpret the server's <BR>lack of=
 support for it as an indication that this is an expensive thing <BR>to do=
 on this server... and to use a different approach.<BR><BR>What I'm saying=
 is that the client would say "list all the mailboxes, <BR>and while you're=
 at it give me &lt;this other information&gt;". &nbsp;If the server=
 <BR>didn't not return the other information, the client would understand=
 <BR>that it's because the server can't do that efficiently, and the client=
 <BR>would proceed *without* the information, using a different method of=
 <BR>processing.<BR><BR>We can't rely on that. &nbsp;What will, instead,=
 happen, is that clients <BR>(perhaps not yours, Simon, but others) will=
 fall back to extracting the <BR>information the old, slow way, incurring=
 all the overhead. &nbsp;Which they <BR>can do anyway, of course,=
 but...<BR><BR>My usual approach is to say that if we can provide clients=
 with <BR>information they want in a way that the server might be able to=
 <BR>optimize, the server might as well support it because the clients will=
 <BR>get it anyway. &nbsp;In this case, though, I agree with Mark that the=
 client <BR>does NOT need the information always, and should instead be=
 more <BR>selective about obtaining it. &nbsp;So the remainder of the=
 "but..." sentence <BR>above is, "...but it's best not to encourage=
 behaviour that is sometimes <BR>bad, and generally unnecessary."<BR><BR>I=
 actually LIKE the idea of the server's having an option of responding=
 <BR>"this is a very expensive operation", and wish we could rely on=
 clients <BR>to pay attention to that. &nbsp;Alas, we=
 can't.<BR><BR>Barry<BR><BR>--<BR>Barry Leiba, Pervasive Computing=
 Technology=
 &nbsp;(leiba@watson.ibm.com)<BR>http://www.research.ibm.com/people/l/leiba=
<BR>http://www.research.ibm.com/spam<BR><BR><BR><BR></BLOCKQUOTE></FONT></b=
ody></html>

--==_16094560900658/PD.gordano.com==_--
--==_16094560900657/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FqDNZ002900; Thu, 2 Jun 2005 08:52:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FqDHg002899; Thu, 2 Jun 2005 08:52:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j52FqBxE002892 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:52:12 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 90CED1A747; Thu,  2 Jun 2005 17:52:10 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 8E18F1A6EF; Thu,  2 Jun 2005 17:52:04 +0200 (CEST)
Message-Id: <G54Annq/kZ1hROmZzY1qKQ.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 17:51:44 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Steve Hole <steve.hole@MessagingDirect.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
In-Reply-To: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Steve Hole writes:
> We are in the IMAP *Extensions* working group.   Presumably we are 
> developing extensions that meet the requirements of the community.
>
> A requirement has been expressed and we have devolved, almost 
> instantly, into a discussion on the relative abilities of some 
> servers to implement the feature that meets the requirement.    This 
> is just wrong.

So it is. Stupid of me to use that argument; it's not central to my position.

I'm not opposed to extensions as such. I do, however, think that 
extensions need to provide some minimum level of utility: Each 
extension brings with it an increased interoperability and testing 
burden, and its utility should be worth that.

In the case of this discussion, I don't see that proposed extension 
does. Perhaps that's because it doesn't, or perhaps that's because the 
people who like the idea aren't pointing out its utility.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52Fpms9002871; Thu, 2 Jun 2005 08:51:48 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52Fpmvu002870; Thu, 2 Jun 2005 08:51:48 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from smtp.osafoundation.org (laweleka.osafoundation.org [204.152.186.98]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FpmaS002852 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:51:48 -0700 (PDT) (envelope-from lisa@osafoundation.org)
Received: from localhost (localhost [127.0.0.1]) by smtp.osafoundation.org (Postfix) with ESMTP id BFC0A142275; Thu,  2 Jun 2005 08:50:53 -0700 (PDT)
Received: from smtp.osafoundation.org ([127.0.0.1]) by localhost (laweleka.osafoundation.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 29352-09; Thu, 2 Jun 2005 08:50:53 -0700 (PDT)
Received: from [152.3.214.109] (user-152-3-214-109.wireless.duke.edu [152.3.214.109]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by smtp.osafoundation.org (Postfix) with ESMTP id 26F05142274; Thu,  2 Jun 2005 08:50:53 -0700 (PDT)
In-Reply-To: <429F24BC.6000908@watson.ibm.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <429F24BC.6000908@watson.ibm.com>
Mime-Version: 1.0 (Apple Message framework v622)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <2562b68792998dcd3baff37762c3480c@osafoundation.org>
Content-Transfer-Encoding: 7bit
Cc: ietf-imapext@imc.org
From: Lisa Dusseault <lisa@osafoundation.org>
Subject: Re: IMAP4 List Command Extensions
Date: Thu, 2 Jun 2005 08:51:36 -0700
To: Barry Leiba <leiba@watson.ibm.com>
X-Mailer: Apple Mail (2.622)
X-Virus-Scanned: by amavisd-new and clamav at osafoundation.org
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Thanks Steve and Barry -- I'm glad this discussion naturally got to the 
point of exploring solutions and costs.

  - Client implementors, is this important enough to really use an 
optional extension in the area? Would client implementors actually 
implement both halves of the "IF (supports optional feature)... ELSE 
..."?

  - Are there any other solutions, workarounds or approaches that would 
address the problem described, and be implementable by most (not 
necessarily all) servers?

thx,
Lisa

On Jun 2, 2005, at 8:24 AM, Barry Leiba wrote:

>
> I don't agree with Mark on every aspect of the IMAP protocol, nor even 
> on every aspect of protocol design in general.  I agree with him on 
> this point.
>
> First, one thing (and I don't mean to get into an extended discussion 
> here; I just want to make a point and move on): there is some 
> inconsistency in Mark's stand.  We can't have the protocol do <x> 
> because it is expensive to do with some legacy message stores (at 
> least one of which he needs to implement for).  But the protocol MUST 
> do <y>, though THAT is expensive to do with some legacy message stores 
> (that he doesn't care about).  (An example of <y> that comes to mind 
> is "return an exact length for RFC822.SIZE, though clients can easily 
> work around that.)  Again, let's not spend a lot of time talking about 
> RFC822.SIZE now -- I want to make the point that we all bring our 
> implementation biases with us.
>
> I also think Simon is right that it's easy to add an extension that's 
> cheap to implement in some (most?) servers, and that it's nice to give 
> that optimization to the client.  We did talk, in fact, about adding 
> STATUS items to the LISTEXT response selections.  I would side with 
> Simon here IF (and this is the key point) we could rely on clients to 
> ask for the information when they need it, and to interpret the 
> server's lack of support for it as an indication that this is an 
> expensive thing to do on this server... and to use a different 
> approach.
>
> What I'm saying is that the client would say "list all the mailboxes, 
> and while you're at it give me <this other information>".  If the 
> server didn't not return the other information, the client would 
> understand that it's because the server can't do that efficiently, and 
> the client would proceed *without* the information, using a different 
> method of processing.
>
> We can't rely on that.  What will, instead, happen, is that clients 
> (perhaps not yours, Simon, but others) will fall back to extracting 
> the information the old, slow way, incurring all the overhead.  Which 
> they can do anyway, of course, but...
>
> My usual approach is to say that if we can provide clients with 
> information they want in a way that the server might be able to 
> optimize, the server might as well support it because the clients will 
> get it anyway.  In this case, though, I agree with Mark that the 
> client does NOT need the information always, and should instead be 
> more selective about obtaining it.  So the remainder of the "but..." 
> sentence above is, "...but it's best not to encourage behaviour that 
> is sometimes bad, and generally unnecessary."
>
> I actually LIKE the idea of the server's having an option of 
> responding "this is a very expensive operation", and wish we could 
> rely on clients to pay attention to that.  Alas, we can't.
>
> Barry
>
> --
> Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
> http://www.research.ibm.com/people/l/leiba
> http://www.research.ibm.com/spam
>



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FZqJE000929; Thu, 2 Jun 2005 08:35:52 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FZq9B000928; Thu, 2 Jun 2005 08:35:52 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FZq9N000921 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:35:52 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA; Thu, 2 Jun 2005 09:37:29 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 09:35:41 -0700
To: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
Subject: Re: IMAP4 List Command Extensions
Cc: ietf-imapext@imc.org
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
Message-ID: <EXECMAIL.20050602093541.B@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52FZq9N000922
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Sorry Arnt, it appears that I'm picking on your messages in particular, 
and I'm not really.   You simply authored some conveniently short messages 
with summary positions that it was easier to do reply editing on.    
Please note that I'm addressing the whole thread rather than a single 
message or viewpoint. 

On Thu, 2 Jun 2005 10:16:20 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> It's difficult to implement efficiently for 
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.) 

So?   Make it an optional feature, and then implement it if you can.

> and not too easy for clients, either, because they have to implement a
> fallback in addition, and test both.

Hogwash.  If you are writing a general purpose client then you have to 
implement fallbacks for the various optional features that are already in 
play and it isn't that hard.   If you are writing a single purpose client 
(say a dedicated webmail service) then you just implement the features you
want and make sure that your backend complies.

I've always found the arguments for "difficult to implement in the client"
to be interesting when the requirements come from client implementors.    
This particular requirement is an attempt to make MUCH more difficult 
tasks in the client (synchronization, display markup, user notification) 
easier and perform better.  The cost of fallback code in the low level 
protocol drivers is laughably trivial compared to these other issues.

Cheers.
---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FOsEx099938; Thu, 2 Jun 2005 08:24:54 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FOsd3099937; Thu, 2 Jun 2005 08:24:54 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from igw2.watson.ibm.com (igw2.watson.ibm.com [129.34.20.6]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FOrtF099922 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:24:54 -0700 (PDT) (envelope-from leiba@watson.ibm.com)
Received: from sp1n294en1.watson.ibm.com (sp1n294en1.watson.ibm.com [129.34.20.40]) by igw2.watson.ibm.com (8.13.1/8.13.1/8.13.1-2005-04-25 igw) with ESMTP id j52FQ5lr022843 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:26:05 -0400
Received: from sp1n294en1.watson.ibm.com (localhost [127.0.0.1]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_2) with ESMTP id j52FOlk94748 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:24:47 -0400
Received: from mars (mars.watson.ibm.com [9.2.40.64]) by sp1n294en1.watson.ibm.com (8.11.7-20030924/8.11.7/01-14-2004_1) with ESMTP id j52FOld94746 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 11:24:47 -0400
Received: from SATURN ([9.12.241.7]) by mars.watson.ibm.com ID IMFd1117725886.26; Thu, 02 Jun 2005 11:24:46 -0400
Message-ID: <429F24BC.6000908@watson.ibm.com>
Date: Thu, 02 Jun 2005 11:24:44 -0400
From: Barry Leiba <leiba@watson.ibm.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
In-Reply-To: <08173804605816@PD.gordano.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

I don't agree with Mark on every aspect of the IMAP protocol, nor even 
on every aspect of protocol design in general.  I agree with him on this 
point.

First, one thing (and I don't mean to get into an extended discussion 
here; I just want to make a point and move on): there is some 
inconsistency in Mark's stand.  We can't have the protocol do <x> 
because it is expensive to do with some legacy message stores (at least 
one of which he needs to implement for).  But the protocol MUST do <y>, 
though THAT is expensive to do with some legacy message stores (that he 
doesn't care about).  (An example of <y> that comes to mind is "return 
an exact length for RFC822.SIZE, though clients can easily work around 
that.)  Again, let's not spend a lot of time talking about RFC822.SIZE 
now -- I want to make the point that we all bring our implementation 
biases with us.

I also think Simon is right that it's easy to add an extension that's 
cheap to implement in some (most?) servers, and that it's nice to give 
that optimization to the client.  We did talk, in fact, about adding 
STATUS items to the LISTEXT response selections.  I would side with 
Simon here IF (and this is the key point) we could rely on clients to 
ask for the information when they need it, and to interpret the server's 
lack of support for it as an indication that this is an expensive thing 
to do on this server... and to use a different approach.

What I'm saying is that the client would say "list all the mailboxes, 
and while you're at it give me <this other information>".  If the server 
didn't not return the other information, the client would understand 
that it's because the server can't do that efficiently, and the client 
would proceed *without* the information, using a different method of 
processing.

We can't rely on that.  What will, instead, happen, is that clients 
(perhaps not yours, Simon, but others) will fall back to extracting the 
information the old, slow way, incurring all the overhead.  Which they 
can do anyway, of course, but...

My usual approach is to say that if we can provide clients with 
information they want in a way that the server might be able to 
optimize, the server might as well support it because the clients will 
get it anyway.  In this case, though, I agree with Mark that the client 
does NOT need the information always, and should instead be more 
selective about obtaining it.  So the remainder of the "but..." sentence 
above is, "...but it's best not to encourage behaviour that is sometimes 
bad, and generally unnecessary."

I actually LIKE the idea of the server's having an option of responding 
"this is a very expensive operation", and wish we could rely on clients 
to pay attention to that.  Alas, we can't.

Barry

--
Barry Leiba, Pervasive Computing Technology  (leiba@watson.ibm.com)
http://www.research.ibm.com/people/l/leiba
http://www.research.ibm.com/spam



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FNKR2099814; Thu, 2 Jun 2005 08:23:20 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52FNK3P099813; Thu, 2 Jun 2005 08:23:20 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from klee.esys.ca (klee.esys.ca [198.161.92.168]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52FNJtw099807 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 08:23:20 -0700 (PDT) (envelope-from steve.hole@messagingdirect.com)
Received: from edmholes.am.tsacorp.com (gallileo.esys.ca [198.161.92.72])  by klee.esys.ca via TCP (smtp internal) with ESMTPA  for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 09:25:04 -0600
From: Steve Hole <steve.hole@MessagingDirect.com>
Date: Thu, 2 Jun 2005 09:23:14 -0700
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com>    <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
Message-ID: <EXECMAIL.20050602092314.A@edmholes.am.tsacorp.com>
X-Mailer: Execmail for Win32 6.0.0 alpha Build (1)
MIME-Version: 1.0
Content-Type: Text/Plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by above.proper.com id j52FNKtw099808
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005 10:16:20 +0200 Arnt Gulbrandsen 
<arnt@gulbrandsen.priv.no> wrote:

> It's easy to implement efficiently for an IMAP server that has exclusive 
> control over its storage. It's difficult to implement efficiently for 
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.) and not 
> too easy for clients, either, because they have to implement a fallback 
> in addition, and test both.

We are in the IMAP *Extensions* working group.   Presumably we are 
developing extensions that meet the requirements of the community.

A requirement has been expressed and we have devolved, almost instantly, 
into a discussion on the relative abilities of some servers to implement 
the feature that meets the requirement.    This is just wrong.

The trick, as always, is to design protocol so that useful features may be
provided by servers and consumed by clients.   There has been a *long* 
history of denying useful protocol features in IMAP, particularly in the 
folder management set, based on the premise that:

 "If we provide an optional feature, then certain clients will depend
  on it and not work properly with servers that don't implement it."

And, this is exactly what happens.  If it's useful, then it's useful and 
people will use it.   The reason that there is resistance to this, 
primarily by server vendors, is that this causes pressure to implement 
something that, possibly, their server just can't do very well.

Oh well!

If the feature is useful, then it will be implemented by clients and 
servers that wish to implement it.   If this causes feature pressure on 
implementations, then too bad.   The feature will either have merit and be
useful or it won't.   It is simply one of a set of features that 
contribute to the selection of one implementation over another.   Denying
features in optional services just because some servers can't implement it
efficiently, but some clearly can, it ludicrous.   It smacks of software 
socialism :-)

The craft of good protocol design is to identify useful but potentially 
inefficient features and make sure that:

1.  Either the feature can be removed from the larger optional set (an 
optional feature within a larger optional feature set).

or

2.  Make the optional features disjoint (no dependencies).

If it was easy, everyone would do protocol design.

As to the specifics of this last proposal, this is perhaps the 5th or 6th 
time that I have see this requirement with more or less the same rhetoric
following the expression of the requirement.

The fact that it has been raised this many times should tell you something 
about the *requirement* at least.   An efficient means of acquiring the 
status/metadata (contents to be defined) of a collection of folders would 
be an *extremely* useful feature.   So useful in fact, that I know of 
several instances of private extensions in this area (ours being one of 
them).  No interoperbility of course, because they are non-standard.

So, I encourage those who are interested in proposing a standard for such a
feature to do so.   Don't be shouted down on the basis of a subset of 
implementations that don't support the underlying infrastructure.  
Discussions on the merit of the requirement/feature should be based in
terms of its general usefulness and applicability to existing feature 
context.   This is software engineering process 101. 

I encourage those who have lots of protocol design experience to make the 
feature conform with the rest of IMAP and be all the other things that we 
want in a protocol: simple, elegant, easy to use.  If it is a feature that
cannot be implemented in your server, then clearly to need to lobby for 
the feature being optional in some appropriate form.

Cheers.

---
Steve Hole
Chief Technical Officer - Electronic Billing and Payment Systems
ACI Worldwide

Email: holes@aciworldwide.com
Phone: 780 424 4922




Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52DECm5086026; Thu, 2 Jun 2005 06:14:12 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52DECfU086025; Thu, 2 Jun 2005 06:14:12 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout4.cac.washington.edu (mxout4.cac.washington.edu [140.142.33.19]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52DE8uw085997 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 06:14:09 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.33.9]) by mxout4.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52DE5xl016312 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 06:14:05 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52DE3vj007934 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 06:14:04 -0700
Date: Thu, 2 Jun 2005 06:14:03 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: Alexey Melnikov <alexey.melnikov@isode.com>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <07455621805752@PD.gordano.com>
Message-ID: <Pine.OSX.4.63.0506020510000.266@pangtzu.panda.com>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Simon Tyler wrote:
> Whether or not a server stores UIDNEXT or calculates it is a server 
> implementation decision. It is irrelevant when designing a protocol.

This is completely wrong.

For better or worse, Internet protocols are molded to accomodate existing 
implementations.  You will receive strong pushback on any proposed 
protocol or protocol enhancement whose impact is perceived as harmful to 
one or more implementations.  Complaints that "it's a poor specification" 
if your demands are not fufilled will fall on deaf ears.

In fact, what you want is to have IMAP molded to accomodate *your* client 
implementation.  Rather than fix your client so it doesn't do something as 
ill-conceived as a STATUS on every LISTed mailbox, you want to change the 
protocol so the server does the stupid thing for you automatically, all to 
save your client some RTTs.  It doesn't address the underlying problem.

The entire point of the Internet protocol architecture is a set of 
interoperable protocols which can be implemented across a wide variety of 
systems, including systems with legacy infrastructure.

The design of protocols for a perfect world, in which you never need to 
concern yourself with the other guy's problems, is not the realm of 
Internet.  It is the realm of proprietary protocols and systems.

In the Internet world, implementors MUST consider the impact of what their 
application does; they can not pretend that they are in a vacuum where 
only they count.  There will be strong pushback against any proposal that 
is perceived as allowing one vendor to count coup against another.

In the proprietary world, you don't have to worry about such things.  You 
control both the client and the server, and any other implementor with the 
temerity to implement your protocol must follow your rules.  At any time, 
you can change the protocol to hurt the other implementor, and you're 
motivated to do so because they are your competitor.

In the Internet world, you have to limit the protocol to what all 
implementations can do well, so that everybody can play ball on a level 
playing field.

In the proprietary world, you can fine-tune the protocol to optimize your 
implementation.  It's your ball and your field.

In the Internet world, interoperability is king, and everything else must 
submit to the demands of that king.

In the proprietary world, the owner is king, and that king receives 
similar submission.

You seem to want the best of both worlds, while paying the price of 
neither.  It doesn't work that way.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52CfDcx073693; Thu, 2 Jun 2005 05:41:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j52CfD17073691; Thu, 2 Jun 2005 05:41:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.166]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j52CfCC2073670 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 05:41:12 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout3.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Cf6SA023084 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 2 Jun 2005 05:41:06 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j52Cf4m9015257 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 2 Jun 2005 05:41:05 -0700
Date: Thu, 2 Jun 2005 05:41:04 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Timo Sirainen <tss@iki.fi>
cc: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <1117701978.15166.414.camel@hurina>
Message-ID: <Pine.OSX.4.63.0506020528510.266@pangtzu.panda.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>  <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com> <1117701978.15166.414.camel@hurina>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Thu, 2 Jun 2005, Timo Sirainen wrote:
> I wouldn't mind implementing combined LIST+STATUS reply. The client I
> use in fact already does STATUS for all mailboxes and I like it that
> way, it's all over in less than a second (~300MB in 30 mboxes, with
> non-Dovecot aware procmail appending mails).

30 is not very many mailboxes.

Consider what happens when there are 300 mailboxes, or 3000 mailboxes, or 
30,000 mailboxes.

There are users who have that many!  And we're not talking about tiny 10MB 
mailboxes either.

Plenty of seemingly-good ideas fail when they are scaled upwards a few 
orders of magnitude more than originally considered.

I can guarantee that those orders of magnitude higher scaling *will* 
occur.  It's a miracle that IMAP, a protocol designed at a time when total 
mailbox size was limited to 1MB, has managed to survive nearly 20 years. 
IMHO, a large reason for IMAP's durability was because IMAP was designed 
to be scalable.

> Anyway, I agree that doing a LIST and sending STATUS to non-\Unmarked
> mailboxes is better.

Indeed.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528kQtU092453; Thu, 2 Jun 2005 01:46:26 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528kQZM092452; Thu, 2 Jun 2005 01:46:26 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from danu.procontrol.fi (kone17.procontrol.vip.fi [212.149.71.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528kP8m092410 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:46:25 -0700 (PDT) (envelope-from tss@iki.fi)
Received: from hurina (82-203-177-3.dsl.gohome.fi [82.203.177.3]) by danu.procontrol.fi (Postfix) with ESMTP id 110C41C134C6 for <ietf-imapext@imc.org>; Thu,  2 Jun 2005 11:46:19 +0300 (EEST)
Subject: Re: IMAP4 List Command Extensions
From: Timo Sirainen <tss@iki.fi>
To: ietf-imapext@imc.org
In-Reply-To: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com> <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-dbyS/srAGiLHWT2++WS8"
Date: Thu, 02 Jun 2005 11:46:18 +0300
Message-Id: <1117701978.15166.414.camel@hurina>
Mime-Version: 1.0
X-Mailer: Evolution 2.0.4 
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

--=-dbyS/srAGiLHWT2++WS8
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Thu, 2005-06-02 at 10:16 +0200, Arnt Gulbrandsen wrote:
> Simon Tyler writes:
> > I would like to see quality in both realms. I don't see that my=20
> > UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds=20
> > to it. I think it is sufficiently easy to implement efficiently that=20
> > it can be included. It is also optional, so a server that cannot=20
> > implement efficienty it can ignore it.
>=20
> It's easy to implement efficiently for an IMAP server that has exclusive=20
> control over its storage. It's difficult to implement efficiently for=20
> other IMAP servers (uw-imapd, dovecot, binc, courier et al.)

I'd just like to point out that while it may be difficult, it's not
impossible. Dovecot has taken ages to get into v1.0 and it's still not
there, partially because it has attempted to implement support for
legacy stores in very efficient way.

I wouldn't mind implementing combined LIST+STATUS reply. The client I
use in fact already does STATUS for all mailboxes and I like it that
way, it's all over in less than a second (~300MB in 30 mboxes, with
non-Dovecot aware procmail appending mails).

Anyway, I agree that doing a LIST and sending STATUS to non-\Unmarked
mailboxes is better. That could however be implemented to the LIST
+STATUS extension if needed. Wouldn't save more than one roundtrip and a
few bytes of traffic though.

Better yet would be instant notification of mailboxes which have
changed, and client could then do STATUS for them if it wanted. Mark
said a year ago he was working on something for this, but I don't
remember anything coming out of it?


--=-dbyS/srAGiLHWT2++WS8
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBCnsdayUhSUUBViskRArHdAKCIqG8r8Snh55hhM5qMxi03xN29gQCfd96E
Ey5MFs/qQti/te/l1SGLfws=
=stuj
-----END PGP SIGNATURE-----

--=-dbyS/srAGiLHWT2++WS8--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528Goo1082401; Thu, 2 Jun 2005 01:16:50 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528GoCv082400; Thu, 2 Jun 2005 01:16:50 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j528Gn0w082385 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:16:49 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 675001A6E0; Thu,  2 Jun 2005 10:16:48 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 102881A6EC; Thu,  2 Jun 2005 10:16:38 +0200 (CEST)
Message-Id: <lAHQRhCM2Kn6nWiEHrIDNw.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:16:20 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <08173804605816@PD.gordano.com>
In-Reply-To: <08173804605816@PD.gordano.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Simon Tyler writes:
> I would like to see quality in both realms. I don't see that my 
> UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds 
> to it. I think it is sufficiently easy to implement efficiently that 
> it can be included. It is also optional, so a server that cannot 
> implement efficienty it can ignore it.

It's easy to implement efficiently for an IMAP server that has exclusive 
control over its storage. It's difficult to implement efficiently for 
other IMAP servers (uw-imapd, dovecot, binc, courier et al.) and not 
too easy for clients, either, because they have to implement a fallback 
in addition, and test both.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j528DDgV081268; Thu, 2 Jun 2005 01:13:13 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j528DDUT081267; Thu, 2 Jun 2005 01:13:13 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j528DCjg081252 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:13:13 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id B3D411A757; Thu,  2 Jun 2005 10:13:11 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id D65CD1A6E0; Thu,  2 Jun 2005 10:13:00 +0200 (CEST)
Message-Id: <HmEGAiGU/6PhDnXVx7FtDA.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:12:42 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com> <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
In-Reply-To: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

Let me step back. AFAICT, there are two approaches to this:

1. Client sends LIST. When an interesting LIST response arrives, it 
sends a STATUS command. As the STATUS commands' responses arrive, the 
client has the information it wants.

2. Client sends LIST with some extension. For each mailbox matching the 
supplied pattern, the server does the same work that STATUS would do, 
and sends the result in a LIST response.

Do I understand it correctly?

Alternative 1 has a cost of two roundtrips, plus traffic costs for the 
LIST responses and for the responses to STATUS, plus CPU costs for 
LIST, STATUS and quite a bit of parsing.

Alternative 2 has a cost of one roundtrip, plus traffic for the LIST 
responses, plus CPU cost for LIST including STATUS.

If we assume that there's 100 LIST responses and 10 interesting 
mailboxes, then 1 costs one roundtrip more and 2 has 90 unnecessary 
STATUS processings and 90 unnecessary transmitted UIDNEXT/UIDVALIDITY 
pairs. (And of course, a client that implements 2 has to implement 1 as 
well, or else it won't work with plain servers.)

Hm. I suppose I must misunderstand terribly, because I cannot see any 
appeal in 2 at all.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5289O6u080036; Thu, 2 Jun 2005 01:09:24 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5289OP4080035; Thu, 2 Jun 2005 01:09:24 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5289LCZ080004 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:09:22 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id gbqudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 09:17:38 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id fbqudaaa; Thu, 2 Jun 2005 09:17:38 +0100
Date: Thu, 2 Jun 2005 09:17:38 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Arnt Gulbrandsen" <arnt@gulbrandsen.priv.no>, <ietf-imapext@imc.org>
Cc: "Alexey Melnikov" <alexey.melnikov@isode.com>, "Mark Crispin" <mrc@cac.washington.edu>
Message-Id: <08173804605816@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
References: <07455621805752@PD.gordano.com> <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_08173804605817/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_08173804605817/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_08173804605818/PD.gordano.com==_"


--==_08173804605818/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I would like to see quality in both realms. I don't see that my=
 UIDNEXT\UIDVALIDITY detracts from protocol quality but  actually adds to=
 it. I think it is sufficiently easy to implement efficiently that it can=
 be included. It is also optional, so a server that cannot implement=
 efficienty it can ignore it.=20

Simon=20

---- Message from mailto:<arnt@gulbrandsen.priv.no Arnt Gulbrandsen=
 <arnt@gulbrandsen.priv.no> at 02-Jun-2005 10:01:52 ------=20

You seem to be pushing "quality" into the server implementation realm=20
and pretending that a protocol itself has no quality.

But it has. A protocol can good, in which case it's easy to implement=20
efficiently, and it can be bad, in which case it usually is difficult=20
to implement efficiently.

Arnt





--==_08173804605818/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P>&nbsp;</P>
<P>I would like to&nbsp;see quality in both realms. I don't see that my=
 UIDNEXT\UIDVALIDITY detracts from protocol quality but&nbsp; actually adds=
 to it. I think it is sufficiently easy to implement efficiently that it=
 can be included. It is also optional, so a server that cannot implement=
 efficienty&nbsp;it can ignore it.</P>
<P><BR>Simon</P>
<P><FONT face=3DArial size=3D2><BR>---- Message from <A=
 href=3D"mailto:<arnt@gulbrandsen.priv.no">Arnt Gulbrandsen=
 &lt;arnt@gulbrandsen.priv.no&gt;</A> at 02-Jun-2005 10:01:52 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">You seem to be pushing "quality" into the server=
 implementation realm <BR>and pretending that a protocol itself has no=
 quality.<BR><BR>But it has. A protocol can good, in which case it's easy=
 to implement <BR>efficiently, and it can be bad, in which case it usually=
 is difficult <BR>to implement=
 efficiently.<BR><BR>Arnt<BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_08173804605818/PD.gordano.com==_--
--==_08173804605817/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5282Q4t077811; Thu, 2 Jun 2005 01:02:26 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5282Qsr077810; Thu, 2 Jun 2005 01:02:26 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from bra.gulbrandsen.priv.no (bra.gulbrandsen.priv.no [212.125.101.197]) by above.proper.com (8.12.11/8.12.9) with SMTP id j5282Lw8077775 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 01:02:23 -0700 (PDT) (envelope-from arnt@gulbrandsen.priv.no)
Received: by bra.gulbrandsen.priv.no (Postfix, from userid 1000) id 657251A752; Thu,  2 Jun 2005 10:02:20 +0200 (CEST)
Received: from prosecco.oryx.com (prosecco.oryx.com [217.19.171.140]) by bra.gulbrandsen.priv.no (Postfix) with SMTP id 44DFA1A6E7; Thu,  2 Jun 2005 10:02:10 +0200 (CEST)
Message-Id: <cjQPpOVsSxzY672NQe6DWg.md5@prosecco.oryx.com>
Date: Thu, 2 Jun 2005 10:01:52 +0200
From: Arnt Gulbrandsen <arnt@gulbrandsen.priv.no>
To: ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
Cc: Simon Tyler <simont@gordano.com>, Alexey Melnikov <alexey.melnikov@isode.com>, Mark Crispin <mrc@cac.washington.edu>
References: <08261932806426@PD.gordano.com> <07455621805752@PD.gordano.com>
In-Reply-To: <07455621805752@PD.gordano.com>
Content-Type: text/plain; format=flowed
MIME-Version: 1.0
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

You seem to be pushing "quality" into the server implementation realm 
and pretending that a protocol itself has no quality.

But it has. A protocol can good, in which case it's easy to implement 
efficiently, and it can be bad, in which case it usually is difficult 
to implement efficiently.

Arnt



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j527bfhM069164; Thu, 2 Jun 2005 00:37:41 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j527bfMw069163; Thu, 2 Jun 2005 00:37:41 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j527baua069131 for <ietf-imapext@imc.org>; Thu, 2 Jun 2005 00:37:37 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id xxpudaaa for ietf-imapext@imc.org; Thu, 2 Jun 2005 08:45:56 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id wxpudaaa; Thu, 2 Jun 2005 08:45:56 +0100
Date: Thu, 2 Jun 2005 08:45:56 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Mark Crispin" <mrc@CAC.Washington.EDU>
Cc: "Alexey Melnikov" <alexey.melnikov@isode.com>, <ietf-imapext@imc.org>
Message-Id: <07455621805752@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <08394718706473@PD.gordano.com>
References: <08261932806426@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_07455623405753/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_07455623405753/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_07455623405754/PD.gordano.com==_"


--==_07455623405754/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
I'm afraid I still disagree on this.=20
 =20
Whether or not a server stores UIDNEXT or calculates it is a server=
 implementation decision. It is irrelevant when designing a protocol. A=
 well designed server will implement the specification in an efficient way,=
 if calculation is expensive then it will store it as metadata, if=
 accessing the metadata is expensive it will cache the metadata. Moulding a=
 specification to a perceived implementation issue (such as performance of=
 obtaining UIDNEXT) will simply result in a poor specification. If we=
 follow your logic then we can equally argue that the SUBSCRIBED option in=
 the draft command extension can also be expense - how expensive is it to=
 decide if a folder is subscribed? I would argue it is implementation=
 dependent.=20
 =20
Simon=20

---- Message from mailto:<mrc@CAC.Washington.EDU Mark Crispin=
 <mrc@CAC.Washington.EDU> at 01-Jun-2005 11:44:10 ------=20

On Wed, 1 Jun 2005, Simon Tyler wrote:
> My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I=20
> fully understand Mark's issue with performance - getting the number of=20
> recent messages back could be inefficient. Perhaps if we just limited=
 to=20
> the two UID values then these performance issue would go away?

It won't.

UIDNEXT is not necessarily in a fixed metadata value; if new mail has=
 been=20
appended then it may be necessary to go through a UID assignment step in=20
order to calculate UIDNEXT.  [This same problem afflicts the UIDPLUS=20
extension; but in a COPY or APPEND command it may be alright if it takes=20
a little bit longer.]

UIDVALIDITY is normally constant, but it could be affected by the UIDNEXT=20
step if it's determined that something is fouled up in the mailbox.

The bottom line is that unless the mail store has an index of all=
 messages=20
for the mailbox (and legacy mail stores do not), these are expensive=20
calculations.

It gets worse.

Even opening a file to sniff at metadata can be costly.  The very first=20
versions of UW imapd did that 15 years ago, to discriminate between=20
mailbox files and non-mailbox files.  This was removed because it=
 resulted=20
in a huge performance penalty.  It may be only a fraction of a second,=
 but=20
when doing a LIST command that fraction of a section per mailbox adds up=20
to several minutes.  Or worse.

This isn't as much of a problem on modern filesystems as in the past, but=20
the problem has not been exterminated.

Even inode metadata from stat() can be expensive.  For example, a timing=20
test on my home directory shows that "ls -RF" takes about twice as long=
 as=20
"ls -R", even when all the directories are in the buffer cache.  That's=20
much better than 10 years ago, but still is not good.

Small systems with few users have less of a problem that large systems=20
with tens or hundreds of thousands of users.

Dedicated mail stores designed for IMAP only use, with its own internal=20
hierarchy that does not attempt to mirror a filesystem hierarchy, also=20
have less of a problem.  [The problem with mirroring a filesystem=20
hierarchy, especially on systems used for more than just IMAP, is that=20
users make symlinks that create filename loops.]

But saying that "the problem can be reduced in some circumstances" does=20
not equate to "the problem does not exist."

A well-designed client works well in a less than ideal world.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.





--==_07455623405754/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>I'm afraid I still disagree on=
 this.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Whether or not a server stores UIDNEXT or=
 calculates it&nbsp;is a server implementation decision. It is irrelevant=
 when designing a protocol. A well designed server will implement the=
 specification in an efficient way, if calculation is expensive then it=
 will store it as metadata, if accessing the metadata is expensive&nbsp;it=
 will cache the metadata.&nbsp;Moulding a specification to a perceived=
 implementation issue (such as performance of obtaining UIDNEXT) will=
 simply result in a poor specification. If we follow your logic then we can=
 equally argue that the SUBSCRIBED option in the draft command extension=
 can also be expense - how expensive is it to decide if a folder is=
 subscribed? I would argue it is implementation dependent.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A href=3D"mailto:<mrc@CAC.Washington.EDU">Mark=
 Crispin &lt;mrc@CAC.Washington.EDU&gt;</A> at 01-Jun-2005 11:44:10=
 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">On Wed, 1 Jun 2005, Simon Tyler wrote:<BR>&gt; My main=
 driver behind this is to get UIDVALIDITY and UIDNEXT back. I <BR>&gt;=
 fully understand Mark's issue with performance - getting the number of=
 <BR>&gt; recent messages back could be inefficient. Perhaps if we just=
 limited to <BR>&gt; the two UID values then these performance issue would=
 go away?<BR><BR>It won't.<BR><BR>UIDNEXT is not necessarily in a fixed=
 metadata value; if new mail has been <BR>appended then it may be necessary=
 to go through a UID assignment step in <BR>order to calculate UIDNEXT.=
 &nbsp;[This same problem afflicts the UIDPLUS <BR>extension; but in a COPY=
 or APPEND command it may be alright if it takes <BR>a little bit=
 longer.]<BR><BR>UIDVALIDITY is normally constant, but it could be affected=
 by the UIDNEXT <BR>step if it's determined that something is fouled up in=
 the mailbox.<BR><BR>The bottom line is that unless the mail store has an=
 index of all messages <BR>for the mailbox (and legacy mail stores do not),=
 these are expensive <BR>calculations.<BR><BR>It gets worse.<BR><BR>Even=
 opening a file to sniff at metadata can be costly. &nbsp;The very first=
 <BR>versions of UW imapd did that 15 years ago, to discriminate between=
 <BR>mailbox files and non-mailbox files. &nbsp;This was removed because it=
 resulted <BR>in a huge performance penalty. &nbsp;It may be only a=
 fraction of a second, but <BR>when doing a LIST command that fraction of a=
 section per mailbox adds up <BR>to several minutes. &nbsp;Or=
 worse.<BR><BR>This isn't as much of a problem on modern filesystems as in=
 the past, but <BR>the problem has not been exterminated.<BR><BR>Even inode=
 metadata from stat() can be expensive. &nbsp;For example, a timing=
 <BR>test on my home directory shows that "ls -RF" takes about twice as=
 long as <BR>"ls -R", even when all the directories are in the buffer=
 cache. &nbsp;That's <BR>much better than 10 years ago, but still is not=
 good.<BR><BR>Small systems with few users have less of a problem that=
 large systems <BR>with tens or hundreds of thousands of=
 users.<BR><BR>Dedicated mail stores designed for IMAP only use, with its=
 own internal <BR>hierarchy that does not attempt to mirror a filesystem=
 hierarchy, also <BR>have less of a problem. &nbsp;[The problem with=
 mirroring a filesystem <BR>hierarchy, especially on systems used for more=
 than just IMAP, is that <BR>users make symlinks that create filename=
 loops.]<BR><BR>But saying that "the problem can be reduced in some=
 circumstances" does <BR>not equate to "the problem does not=
 exist."<BR><BR>A well-designed client works well in a less than ideal=
 world.<BR><BR>-- Mark --<BR><BR>http://staff.washington.edu/mrc<BR>Science=
 does not emerge from voting, party politics, or public debate.<BR>Si vis=
 pacem, para bellum.<BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_07455623405754/PD.gordano.com==_--
--==_07455623405753/PD.gordano.com==_--



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j51IiJUV059686; Wed, 1 Jun 2005 11:44:19 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j51IiJS2059685; Wed, 1 Jun 2005 11:44:19 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from mxout2.cac.washington.edu (mxout2.cac.washington.edu [140.142.33.4]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j51IiG9V059675 for <ietf-imapext@imc.org>; Wed, 1 Jun 2005 11:44:16 -0700 (PDT) (envelope-from mrc@CAC.Washington.EDU)
Received: from smtp.washington.edu (smtp.washington.edu [140.142.32.139]) by mxout2.cac.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j51IiCtx014921 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 1 Jun 2005 11:44:12 -0700
Received: from pangtzu.panda.com (pangtzu.panda.com [206.124.149.117]) (authenticated authid=mrc) by smtp.washington.edu (8.13.4+UW05.04/8.13.4+UW05.05) with ESMTP id j51IiAXj002892 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Wed, 1 Jun 2005 11:44:11 -0700
Date: Wed, 1 Jun 2005 11:44:10 -0700 (PDT)
From: Mark Crispin <mrc@CAC.Washington.EDU>
To: Simon Tyler <simont@gordano.com>
cc: Alexey Melnikov <alexey.melnikov@isode.com>, ietf-imapext@imc.org
Subject: Re: IMAP4 List Command Extensions
In-Reply-To: <07144132803948@PD.gordano.com>
Message-ID: <Pine.OSX.4.63.0506011125420.266@pangtzu.panda.com>
References: <07363107802703@PD.gordano.com> <07062778102640@PD.gordano.com> <07062778102640@PD.gordano.com> <07363107802703@PD.gordano.com> <07144132803948@PD.gordano.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

On Wed, 1 Jun 2005, Simon Tyler wrote:
> My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I 
> fully understand Mark's issue with performance - getting the number of 
> recent messages back could be inefficient. Perhaps if we just limited to 
> the two UID values then these performance issue would go away?

It won't.

UIDNEXT is not necessarily in a fixed metadata value; if new mail has been 
appended then it may be necessary to go through a UID assignment step in 
order to calculate UIDNEXT.  [This same problem afflicts the UIDPLUS 
extension; but in a COPY or APPEND command it may be alright if it takes 
a little bit longer.]

UIDVALIDITY is normally constant, but it could be affected by the UIDNEXT 
step if it's determined that something is fouled up in the mailbox.

The bottom line is that unless the mail store has an index of all messages 
for the mailbox (and legacy mail stores do not), these are expensive 
calculations.

It gets worse.

Even opening a file to sniff at metadata can be costly.  The very first 
versions of UW imapd did that 15 years ago, to discriminate between 
mailbox files and non-mailbox files.  This was removed because it resulted 
in a huge performance penalty.  It may be only a fraction of a second, but 
when doing a LIST command that fraction of a section per mailbox adds up 
to several minutes.  Or worse.

This isn't as much of a problem on modern filesystems as in the past, but 
the problem has not been exterminated.

Even inode metadata from stat() can be expensive.  For example, a timing 
test on my home directory shows that "ls -RF" takes about twice as long as 
"ls -R", even when all the directories are in the buffer cache.  That's 
much better than 10 years ago, but still is not good.

Small systems with few users have less of a problem that large systems 
with tens or hundreds of thousands of users.

Dedicated mail stores designed for IMAP only use, with its own internal 
hierarchy that does not attempt to mirror a filesystem hierarchy, also 
have less of a problem.  [The problem with mirroring a filesystem 
hierarchy, especially on systems used for more than just IMAP, is that 
users make symlinks that create filename loops.]

But saying that "the problem can be reduced in some circumstances" does 
not equate to "the problem does not exist."

A well-designed client works well in a less than ideal world.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.



Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5176WN3097803; Wed, 1 Jun 2005 00:06:32 -0700 (PDT) (envelope-from owner-ietf-imapext@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j5176WBB097802; Wed, 1 Jun 2005 00:06:32 -0700 (PDT)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-imapext@mail.imc.org using -f
Received: from PD.gordano.com (id191.gordano.com [62.172.232.191]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j5176RmH097767 for <ietf-imapext@imc.org>; Wed, 1 Jun 2005 00:06:28 -0700 (PDT) (envelope-from simont@gordano.com)
Received: from [127.0.0.1] by PD.dev.ntmail.co.uk (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id fekudaaa for ietf-imapext@imc.org; Wed, 1 Jun 2005 08:14:41 +0100
Received: from [62.172.232.189] by PD.gordano.com (GMS 11.00.3334/SQ2499.01.1aa407eb) with ESMTP id eekudaaa; Wed, 1 Jun 2005 08:14:41 +0100
Date: Wed, 1 Jun 2005 08:14:41 +0100
From: "Simon Tyler" <simont@gordano.com>
Subject: Re: IMAP4 List Command Extensions
To: "Alexey Melnikov" <alexey.melnikov@isode.com>
Cc: <ietf-imapext@imc.org>
Message-Id: <07144132803948@PD.gordano.com>
X-Mailer: Gordano Messaging Suite v11.00.3334
X-Originating-IP: [62.172.232.189]
In-Reply-To: <Pine.OSX.4.63.0505310031220.305@pangtzu.panda.com>
References: <07363107802703@PD.gordano.com> <07062778102640@PD.gordano.com> <07062778102640@PD.gordano.com> <07363107802703@PD.gordano.com>
X-Sender: simont-webmail@dev.ntmail.co.uk
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==_07144132803949/PD.gordano.com==_"
X-AntiVirus: Checked for viruses by Gordano's AntiVirus Software
Sender: owner-ietf-imapext@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-imapext/mail-archive/>
List-ID: <ietf-imapext.imc.org>
List-Unsubscribe: <mailto:ietf-imapext-request@imc.org?body=unsubscribe>

This is a MIME-encapsulated message

--==_07144132803949/PD.gordano.com==_
Content-Type: multipart/alternative; boundary="==_07144132803950/PD.gordano.com==_"


--==_07144132803950/PD.gordano.com==_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


 =20
Looks like almost the same idea I had just with slightly different=
 syntax.=20
 =20
My main driver behind this is to get UIDVALIDITY and UIDNEXT back. I fully=
 understand Mark's issue with performance - getting the number of recent=
 messages back could be inefficient. Perhaps if we just limited to the two=
 UID values then these performance issue would go away?=20
 =20
Simon=20

---- Message from mailto:<alexey.melnikov@isode.com Alexey Melnikov=
 <alexey.melnikov@isode.com> at 31-May-2005 18:02:00 ------=20

Simon Tyler wrote:

> Indeed the protocol should consider a wide variety of implementations.=20
> Such as one which has no large overhead in obtaining the extra=20
> information I was proposing. Our client will clearly backdown to using=20
> a STATUS command should the extended LIST response not be found.
>
> =20
>
> I fully understand all the issues, that is why I felt it was a=20
> reasonable extension - servers which have issues returning the UID=20
> information in short timescales would obviously not choose to add in=20
> such an extension. Those that can obtain the information could do so.
>
Exactly.

I've posted a similar proposal some times ago:

<http://www.imc.org/ietf-imapext/mail-archive/msg02245.html>

which I intend to submit as a draft, once the LIST-EXTENDED is done.






--==_07144132803950/PD.gordano.com==_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html><head><style><!-- P{ margin-top:0;margin-bottom:0;} //=
 --></style></head><body><P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Looks like almost the same idea I had just=
 with slightly different syntax.</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>My main driver behind this is to get=
 UIDVALIDITY and UIDNEXT back. I fully understand Mark's issue with=
 performance - getting the number of recent messages back could be=
 inefficient. Perhaps if we just limited to the two UID values then these=
 performance issue would go away?</FONT></P>
<P><FONT face=3DArial size=3D2></FONT>&nbsp;</P>
<P><FONT face=3DArial size=3D2>Simon</P>
<P><BR>---- Message from <A=
 href=3D"mailto:<alexey.melnikov@isode.com">Alexey Melnikov=
 &lt;alexey.melnikov@isode.com&gt;</A> at 31-May-2005 18:02:00 ------</P>
<BLOCKQUOTE style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT:=
 #000000 2px solid">Simon Tyler wrote:<BR><BR>&gt; Indeed the protocol=
 should consider a wide variety of implementations. <BR>&gt; Such as one=
 which has no large overhead in obtaining the extra <BR>&gt; information I=
 was proposing. Our client will clearly backdown to using <BR>&gt; a STATUS=
 command should the extended LIST response not be found.<BR>&gt;<BR>&gt;=
 &nbsp;<BR>&gt;<BR>&gt; I fully understand all the issues, that is why I=
 felt it was a <BR>&gt; reasonable extension - servers which have issues=
 returning the UID <BR>&gt; information in short timescales would obviously=
 not choose to add in <BR>&gt; such an extension. Those that can obtain the=
 information could do so.<BR>&gt;<BR>Exactly.<BR><BR>I've posted a similar=
 proposal some times=
 ago:<BR><BR>&lt;http://www.imc.org/ietf-imapext/mail-archive/msg02245.html=
&gt;<BR><BR>which I intend to submit as a draft, once the LIST-EXTENDED is=
 done.<BR><BR><BR><BR></BLOCKQUOTE></FONT></body></html>

--==_07144132803950/PD.gordano.com==_--
--==_07144132803949/PD.gordano.com==_--


