Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 16:08:00 +0100 From: Josh Howlett Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Message-ID: <6ED388AA006C454BA35B0098396B9BFB0716652D@uxsrvr20.atlas.ukerna.ac.uk> > >> Josh's idea to get around the RADIUSS MTU issue is to include a > >> compressed authentication assertion in the RADIUS response and > >> for the SP to make a SAML attribute query. The authentication > >> assertion would effectively serve as a token for the=20 > SP to prove > >> it was entitled to attributes surrounding the subject. >=20 > Scott> That's a bit of a complex approach and requires=20 > implementing > Scott> WS-Security-based SOAP queries, which is not something I'd > Scott> take on without a really good reason. >=20 > How does authentication for attribute queries actually happen=20 > then today? I think we might have our wires crossed slightly. My suggestion was that the SP uses the Subject from the assertion within the AttributeQuery, as normally happens with attribute query, not the entire assertion. I agree with Scott that using an assertion as a token would be rather complex. However, I think I initially misunderstood Sam's proposal. I assumed that the attribute query and response would be "signed" using the key sent to the SP over RADIUS. josh. JANET(UK) is a trading name of The JNT Association, a company limited by guarantee which is registered in England under No. 2881024=20 and whose Registered Office is at Lumen House, Library Avenue, Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 15:12:54 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <00ad01cae899$231c0680$69541380$@2@osu.edu> > There's another way though. The IDP and RADIUS server already share a > secret. Use that key and a secure key derivation function (NIST SP > 800-56?) to construct the key. That way you don't actually need to > spend any bytes on key transport. Right, that makes sense. > I probably prefer explicit key transport because it's easier to see what > its security properties are. The KDF is fine though from a security > standpoint and definitely avoids any concern about length of identifiers. A key + an alg OID or something compact would probably fit without much trouble, I just wanted to note the possibility. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 15:03:03 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: If we run into size limits, there are several approaches. Note that we need to trasport between 16 and 32 bytes of key material plus an algorithm identifier if we want to transport the key. There's another way though. The IDP and RADIUS server already share a secret. Use that key and a secure key derivation function (NIST SP 800-56?) to construct the key. That way you don't actually need to spend any bytes on key transport. I probably prefer explicit key transport because it's easier to see what its security properties are. The KDF is fine though from a security standpoint and definitely avoids any concern about length of identifiers. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 14:54:46 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <00a801cae896$9a24ec90$ce6ec5b0$@2@osu.edu> > What I think we can take away from this is that > > 1) RADIUS needs to be able to express this at the cross -domain level > (trivially true) > 2) Our architecture needs to permit deployments where this distinction > is reflected into separate SP SAML identities. That's my preference, yes. Thanks, -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 14:51:52 +0100 From: Josh Howlett Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Message-ID: <6ED388AA006C454BA35B0098396B9BFB071664D8@uxsrvr20.atlas.ukerna.ac.uk> > In the scope of the beyond web use case, I'm not concerned=20 > about SPs needing metadata for the IDP; only for the IDP to=20 > be provisioned with information about each SP. Why aren't you concerned about SPs needing metadata for the IdP for the beyond web SSO use cases? > I think we should carefully explore how true that is for the=20 > use cases motivating beyond web SSO. Is it really true for=20 > the outsourcing providers that servers are registered at the=20 > individual host level? It might be, but it doesn't have to be. You can abstract a host of hosts behind a single SAML SP entity. > That seems surprising to me. Also, I think both for the HPC=20 > and outsourcing use case, allowing each SP to dynamically=20 > manage their servers will be important. I agree. It's just not the way that we interact with customers today (of course, that's not a reason not to change). > A third observation that comes to me: the complexity of all=20 > of this argues for doing as much as we can in the RADIUS=20 > exchange so that we do not need multiple trust infrastructures. Right. Ultimately, that's the point of the "scalable trust" use-case category: to seek a single mechanism to support the exchange of claims about things in different types of systems, without getting distracted by the idiosyncracies of the systems in question. josh. JANET(UK) is a trading name of The JNT Association, a company limited by guarantee which is registered in England under No. 2881024=20 and whose Registered Office is at Lumen House, Library Avenue, Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 14:51:20 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: Scott> As organizations get large, this distinction matters more, Scott> and as they get small, the benefits of the gateway decrease. >> Ah. I was sort of assuming you'd want to release to "blah mail >> servers," or to "Quux's HPC cluster." The rationale is that I >> might be willing to release far more information to MIT's HPC >> cluster with which I have a business arrangement than I would to >> a random dorm computer at MIT. Scott> Yes, that's my point. I can't do that meaningfully if the Scott> only distinction that exists is behind the RADIUS server at Scott> MIT. I can trust the RADIUS server to do the right thing, but Scott> I have no direct influence. What I think we can take away from this is that 1) RADIUS needs to be able to express this at the cross -domain level (trivially true) 2) Our architecture needs to permit deployments where this distinction is reflected into separate SP SAML identities. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 14:02:34 +0100 From: Josh Howlett Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Message-ID: <6ED388AA006C454BA35B0098396B9BFB071664AB@uxsrvr20.atlas.ukerna.ac.uk> Sam, > Josh and Scott have been assuming that we would have the same=20 > level of provisioning today for relying parties. ... > I also don't think it is desirable to couple the Moonshot=20 > trust management tightly enough to the beyond SSO stuff in=20 > order to require the trust management to make this work. Agreed. I think Scott and I were going further, in the sense that we shouldn't have a coupling with /any/ particular trust management scheme. I also agree that we should seek optimisations to facilitate deployment (but in seeking optimisations, we should be careful not to neglect the composability of the architecture). > I wonder whether we could provision a key that is conveyed to=20 > the IDP in the subject with the RADIUS exchange. That way,=20 > if the SP didn't have other credentials to use, it could=20 > authenticate as the holder of that key. The key could be=20 > transported to the SP in a RADIUS attribute, and transported=20 > to the IDP in the subject. Interesting idea! Some observations: First, the provisioning that we use today isn't just about key provisioning. It is also used to provide entities with information about the location of service endpoints; for example, the URI needed by the SP to approach the IdP with an attribute query. We normally provision this information alongside the key(s) in metadata. We could consider an approach whereby the entities concerned obtain this metadata from each other directly, and establish trust in the metadata by extending your mechanism to supply a proof. Unfortunately this mechanism would not support the current means by which we (as federation operators) make statements about entities that are frequently used by other entities to inform trust decisions (such as the LoA supported by an IdP). We could change the way in which we (as federation operators) make these statements, but this would require distributing signed assertions that contain these claims to the entity operators for them to include in their metadata. This would necessitate some fairly signicant changes to operational practise. I note that Moonshot trust management, when applied to SAML metadata, also implies significant changes to operational practise that might not be dissimilar to that described above, and so it might be worth exploring whether this is actually a useful transition step. As in all things related to trust management, I need to think this over some more. My second observation is that this reduces operational complexity at the expense of additional implementation effort (I don't think it would be significant effort, and I will immediately defer to Scott's opinion!). However, the initial deployment community that is the target of the proof-of-concept already bears the costs of operating distinct RADIUS and SAML systems, and so this optimisation doesn't really help. I understand, however, that it would help other communities. best regards, josh. JANET(UK) is a trading name of The JNT Association, a company limited by guarantee which is registered in England under No. 2881024=20 and whose Registered Office is at Lumen House, Library Avenue, Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 12:33:42 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <008d01cae882$e58fa7b0$b0aef710$@2@osu.edu> > Scott> Even if you really did imitate Kerberos and send the SP both > Scott> the key and the key encrypted to the IdP, what would be the > Scott> mechanism by which the latter ends up back at the IdP? > > > I can see a couple of options. > > 1) The RADIUS server generates metadata and feeds it to the IDP True, I hadn't considered that. > 2) What I recommend we implement: the key is included in the encrypted > subject. Maybe, but I don't think it would be an EncryptedID in the SAML sense of the construct, I don't think there's a way to embed a key like this. There *is* a way via holder of key assertions, which is why I assumed that's what you meant literally. You can create encrypted, MAC'd strings that carry data to the IdP via transient NameIDs, but there are some size limits. We used to encrypt the principal name, expiration time, SP name, and a MAC, and they were ok, but not that far from the limit. Of course, a custom Format can be used for a larger NameID to get around the limit, but the limit's there for practical reasons (logging/auditing, that kind of thing). > As far as the acceptor is concerned, we're saying that as a local > matter, the RADIUS server and IDP need to make sure that the IDP knows > the key somehow. That doesn't allow us to escape specifying a mechanism > so you can manage to have an IDP of your choice work with an arbitrary > RADIUS server, but it does minimize the number of parties in that > mechanism. Yes, I just wasn't sure whether you had a particular strategy defined that I wasn't seeing, or were deferring the details. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 12:21:03 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <008501cae881$212887d0$63799770$@2@osu.edu> > Scott> As organizations get large, this distinction matters more, > Scott> and as they get small, the benefits of the gateway decrease. > Ah. I was sort of assuming you'd want to release to "blah mail > servers," or to "Quux's HPC cluster." The rationale is that I might be > willing to release far more information to MIT's HPC cluster with which > I have a business arrangement than I would to a random dorm computer at > MIT. Yes, that's my point. I can't do that meaningfully if the only distinction that exists is behind the RADIUS server at MIT. I can trust the RADIUS server to do the right thing, but I have no direct influence. But queries of course never provide anything but an illusion of such control anyway, since all you get is the key and no control over where the data is actually sent back to. And location-based control on the front channel is contingent on vetting of locations, and most federations devolve that to organizational control anyway. But I think in terms of architecturally what's possible vs. what's true in common practice. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 12:12:29 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: >> In this instance, I think the gateway significantly helps trust. >> It means that the credential that the IDP knows about is housed >> on a fairly trusted box rather than say a bunch of edge servers. >> It means that the SP domain has much better options for dealing >> with credential rotation and key compromise for the credentials >> that are on the edge servers. Scott> Yes, but it also means that as an IdP I'm effectively Scott> releasing data not to the edge servers hosting actual Scott> resources or services but to a central system that may have Scott> nothing to do with my users, or the use case. And I argue as Scott> a result that the SP should never be portrayed as anything Scott> other than "Ohio State" or "MIT" or "Microsoft", or whatever Scott> the case may be. Scott> As organizations get large, this distinction matters more, Scott> and as they get small, the benefits of the gateway decrease. Ah. I was sort of assuming you'd want to release to "blah mail servers," or to "Quux's HPC cluster." The rationale is that I might be willing to release far more information to MIT's HPC cluster with which I have a business arrangement than I would to a random dorm computer at MIT. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 12:04:26 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007e01cae87e$cef07970$6cd16c50$@2@osu.edu> > In this instance, I think the gateway significantly helps trust. It > means that the credential that the IDP knows about is housed on a fairly > trusted box rather than say a bunch of edge servers. It means that the > SP domain has much better options for dealing with credential rotation > and key compromise for the credentials that are on the edge servers. Yes, but it also means that as an IdP I'm effectively releasing data not to the edge servers hosting actual resources or services but to a central system that may have nothing to do with my users, or the use case. And I argue as a result that the SP should never be portrayed as anything other than "Ohio State" or "MIT" or "Microsoft", or whatever the case may be. As organizations get large, this distinction matters more, and as they get small, the benefits of the gateway decrease. > We could do this all end-to-end, but I think it would actually decrease > the robustness and security of the system. If I'm wrong , then we can > probably simplify things a lot by treating it end-to-end. One of the differences I'll grant is that because of the dependency on RADIUS regardless, this is not something that a run of the mill developer can simply instantiate without regard for his broader organizational infrastructure. That's a concern to me, but it's not a problem that your suggestion exacerbates, I admit. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:54:23 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: Scott> So does assigning everything its own IP address, but there Scott> are understood to be benefits. Gateways harm trust, Scott> accountability, privacy, etc. In this instance, I think the gateway significantly helps trust. It means that the credential that the IDP knows about is housed on a fairly trusted box rather than say a bunch of edge servers. It means that the SP domain has much better options for dealing with credential rotation and key compromise for the credentials that are on the edge servers. >> This discussion is reenforcing my belief that we're far better >> off if everything happens in the RADIUS exchange. Scott> Is it the local RADIUS hop that proves its identity Scott> cross-domain then? And it's left to that system to decide Scott> which GSS acceptors are allowed to logically act in the Scott> capacity that the RADIUS system represents to the outside Scott> world? Yes. There's some discussion of this in the feasibility analysis. We could do this all end-to-end, but I think it would actually decrease the robustness and security of the system. If I'm wrong , then we can probably simplify things a lot by treating it end-to-end. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:48:52 +0100 From: Josh Howlett Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Probably no audio archive for the bar bof Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Message-ID: <6ED388AA006C454BA35B0098396B9BFB07166448@uxsrvr20.atlas.ukerna.ac.uk> > I was hoping to post a link to the IETF bar bof audio.=20 > Unfortunately, as best I can tell,while the audio was=20 > streamed, only normal sessions were archived. > So, it looks like we will not have an audio archive for te bar bof. I was listening on iTunes. I am probably clutching at straws but does anyone know if iTunes caches streams locally? A quick Google search suggests that I am, indeed, straw-clutching but I might not be using the right search terms. josh. JANET(UK) is a trading name of The JNT Association, a company limited by guarantee which is registered in England under No. 2881024=20 and whose Registered Office is at Lumen House, Library Avenue, Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:46:52 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: Scott> I know that's the underlying exchange, but you were Scott> describing an in-band communication of the proof key from Scott> RADIUS to the IdP >> yes. Scott> But as I understand it, the IdP isn't supposed to have to Scott> know anything about RADIUS, so outside of using a holder of Scott> key assertion, how else would it be done? Scott> Even if you really did imitate Kerberos and send the SP both Scott> the key and the key encrypted to the IdP, what would be the Scott> mechanism by which the latter ends up back at the IdP? I can see a couple of options. 1) The RADIUS server generates metadata and feeds it to the IDP 2) What I recommend we implement: the key is included in the encrypted subject. As far as the acceptor is concerned, we're saying that as a local matter, the RADIUS server and IDP need to make sure that the IDP knows the key somehow. That doesn't allow us to escape specifying a mechanism so you can manage to have an IDP of your choice work with an arbitrary RADIUS server, but it does minimize the number of parties in that mechanism. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:36:55 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007d01cae87a$f65dd880$e3198980$@2@osu.edu> > Scott> I know that's the underlying exchange, but you were > Scott> describing an in-band communication of the proof key from > Scott> RADIUS to the IdP > yes. But as I understand it, the IdP isn't supposed to have to know anything about RADIUS, so outside of using a holder of key assertion, how else would it be done? Even if you really did imitate Kerberos and send the SP both the key and the key encrypted to the IdP, what would be the mechanism by which the latter ends up back at the IdP? I thought the general flow of the SAML would be: - RADIUS sends back authn assertion to SP along with key it should use. - SP performs query to IdP signed with key using std SOAP binding. So how does the key binding get communicated to the IdP, which isn't party to the first step? -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:30:32 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007c01cae87a$122529c0$366f7d40$@2@osu.edu> > Scott> My suggestion is to avoid bringing the end to end vs. gateway > Scott> discussion into the architecture. If people want gateways, > Scott> they can proxy the interactions as they choose. > > I have a couple of concerns about that approach. > > First, we're building an implementation not just a standard. So, I > don't think we can escape it. Well, I escape it in Shibboleth by building for end to end and letting people that want gateways get them somewhere else, but that's my bias. > Secondly, I don't see any reason to believe that the gateway and GSS > acceptor are guaranteed to be from the same vendor. At least as the > IETF looks at things, that means you need a spec to get interoperability. Yes, but that's not a lot of work to specify, at least with regard to SOAP queries. > Finally, I think ignoring this will increase deployment complexity > significantly. So does assigning everything its own IP address, but there are understood to be benefits. Gateways harm trust, accountability, privacy, etc. > This discussion is reenforcing my belief that we're far better off if > everything happens in the RADIUS exchange. Is it the local RADIUS hop that proves its identity cross-domain then? And it's left to that system to decide which GSS acceptors are allowed to logically act in the capacity that the RADIUS system represents to the outside world? -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:24:24 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: >> Ah, I think I might have been two precise in my terminology. >> server TLS plus xml signature on the query with the SP learning >> the key to use for a MAC from the RADIUS exchange falls within >> what I was talking about. Scott> I know that's the underlying exchange, but you were Scott> describing an in-band communication of the proof key from Scott> RADIUS to the IdP yes. Scott> via a holder of key assertion, which is Scott> where the complexity comes in. Ah, no. I only intended to draw an analogy to holder of key. I was just planning to use RADIUS's existing "mechanisms" [1] for sending the key. [1] They are fairly primitive. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:21:49 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007b01cae878$da8a0b80$8f9e2280$@2@osu.edu> > Ah, I think I might have been two precise in my terminology. server TLS > plus xml signature on the query with the SP learning the key to use for > a MAC from the RADIUS exchange falls within what I was talking about. I know that's the underlying exchange, but you were describing an in-band communication of the proof key from RADIUS to the IdP via a holder of key assertion, which is where the complexity comes in. My question remains, though, how does the SP prove that it should receive that key from the RADIUS infrastructure? I thought it did so by authenticating itself with something that the IdP-affiliated RADIUS would trust. That amounts to the same issue with the SP having credentials trusted cross-domain. I'm sure it's just my RADIUS and EAP ignorance here. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:21:33 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: >> From a SAML standpoint this is true. I'm concerned though about >> how this works at an implementation level with GSS. In >> particular, it seems very undesirable to force all the >> cross-domain credentials out to endpoints. Scott> My suggestion is to avoid bringing the end to end vs. gateway Scott> discussion into the architecture. If people want gateways, Scott> they can proxy the interactions as they choose. I have a couple of concerns about that approach. First, we're building an implementation not just a standard. So, I don't think we can escape it. Secondly, I don't see any reason to believe that the gateway and GSS acceptor are guaranteed to be from the same vendor. At least as the IETF looks at things, that means you need a spec to get interoperability. Finally, I think ignoring this will increase deployment complexity significantly. This discussion is reenforcing my belief that we're far better off if everything happens in the RADIUS exchange. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:17:07 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: >> How does authentication for attribute queries actually happen >> then today? Scott> Commercially, you'll see TLS, signing, and even basic-auth Scott> (commercial products tending to care only about point to Scott> point federation). Scott> The SP supports a few esoteric methods, but the two that are Scott> actually widely supported are TLS or XML Signatures over the Scott> SAML request in some combination, with the trust layer based Scott> on the metadata corresponding to the SAML message issuer. Scott> In practice, it's usually mutual TLS, but client signature + Scott> server TLS would be the other one that you'd see for cases Scott> where the TLS doesn't terminate at the IdP. Ah, I think I might have been two precise in my terminology. server TLS plus xml signature on the query with the SP learning the key to use for a MAC from the RADIUS exchange falls within what I was talking about. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:10:41 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007a01cae877$4cc6ce10$e6546a30$@2@osu.edu> > How does authentication for attribute queries actually happen then today? Commercially, you'll see TLS, signing, and even basic-auth (commercial products tending to care only about point to point federation). The SP supports a few esoteric methods, but the two that are actually widely supported are TLS or XML Signatures over the SAML request in some combination, with the trust layer based on the metadata corresponding to the SAML message issuer. In practice, it's usually mutual TLS, but client signature + server TLS would be the other one that you'd see for cases where the TLS doesn't terminate at the IdP. We implemented delegation support at the IdP via an extension that supports client TLS as a key proof for a holder of key assertion in a WS-Security header, but it's handled only in the extension for a single profile we needed and hasn't been generalized out to the IdP's security layer for use with other profiles. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 11:04:02 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007901cae876$5e9812d0$1bc83870$@2@osu.edu> > From a SAML standpoint this is true. I'm concerned though about how > this works at an implementation level with GSS. In particular, it seems > very undesirable to force all the cross-domain credentials out to > endpoints. My suggestion is to avoid bringing the end to end vs. gateway discussion into the architecture. If people want gateways, they can proxy the interactions as they choose. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:51:08 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: >> Josh's idea to get around the RADIUSS MTU issue is to include a >> compressed authentication assertion in the RADIUS response and >> for the SP to make a SAML attribute query. The authentication >> assertion would effectively serve as a token for the SP to prove >> it was entitled to attributes surrounding the subject. Scott> That's a bit of a complex approach and requires implementing Scott> WS-Security-based SOAP queries, which is not something I'd Scott> take on without a really good reason. How does authentication for attribute queries actually happen then today? Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:49:00 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: Scott> Furthermore, in the query case specifically, offloading the Scott> query to a shared system that handles the load for a range of Scott> actual app services is fairly simple. That's merely a twist Scott> on the usual commercial model of federation, which is to have Scott> a gateway visible to the IdPs and everything else implemented Scott> with proprietary SSO behind that. Again, to standardize this with GSS, and more particularly to implement this, we need to actually specify all the components. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:47:06 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Scott" == Scott Cantor writes: Scott> An SP is a wholly logical construct. The mapping to physical Scott> systems is entirely arbitrary. It's also not based on Scott> hostnames like SSL is. Hostnames only come into play when Scott> they have to, to enforce security provisions that are Scott> relevant to the IdP, such as with web site endpoints. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:37:17 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <007401cae872$a2019950$e604cbf0$@2@osu.edu> > Josh's idea to get around the RADIUSS MTU issue is to include a > compressed authentication assertion in the RADIUS response and for the > SP to make a SAML attribute query. The authentication assertion would > effectively serve as a token for the SP to prove it was entitled to > attributes surrounding the subject. That's a bit of a complex approach and requires implementing WS-Security-based SOAP queries, which is not something I'd take on without a really good reason. In SAML, that's essentially delegation. We know how to do it and we have code that does a big chunk of it, but it's not implemented for securing attribute queries on either end right now. > Josh and Scott have been assuming that we would have the same level of > provisioning today for relying parties. I'm very uncomfortable with that. Well, for reasons I stated in the other message, this comes up a lot but I think is largely a non-issue in practice. > In order to support our use cases I think it is critical that SPs should > only need to be provisioned within their own domain, not globally. An SP is a wholly logical construct. The mapping to physical systems is entirely arbitrary. It's also not based on hostnames like SSL is. Hostnames only come into play when they have to, to enforce security provisions that are relevant to the IdP, such as with web site endpoints. > For example, I think we'd want to say something roughly like "Google is > permitted to run mail servers," globally, "Google's mail servers get the > e-mail address of XYZ's users," in XYZ's attribute release policies, but > only Google needs to know what Google's mail servers actually are. That just means they all have to share one SAML identity and some set of credentials bound to that identity. > I wonder whether we could provision a key that is conveyed to the IDP in > the subject with the RADIUS exchange. That way, if the SP didn't have > other credentials to use, it could authenticate as the holder of that > key. The key could be transported to the SP in a RADIUS attribute, and > transported to the IDP in the subject. Isn't it the case that the RADIUS exchange is such that the SP is being authenticated anyway? I thought the whole system was predicated on the SP being trusted to participate in the exchange. If it's trusted, then it has a credential anyway, right? > So, under this model, the subject would be very similar to a Kerberos > ticket; in an encrypted structure it would include: Notwithstanding my questions, yes, it's possible to do most of that with SubjectConfirmation. But I think the implementation complexity coupled with the questionable benefits makes it a bad idea. Plus which, I thought you were arguing for using transient IDs that would act as pesudo-bearer credentials anyway? -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:34:03 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Josh" == Josh Howlett writes: >> In the scope of the beyond web use case, I'm not concerned about >> SPs needing metadata for the IDP; only for the IDP to be >> provisioned with information about each SP. Josh> Why aren't you concerned about SPs needing metadata for the Josh> IdP for the beyond web SSO use cases? Not entirely sure. Some of it is an assumption that I'm making that there are more SPs than IDPs. >> I think we should carefully explore how true that is for the use >> cases motivating beyond web SSO. Is it really true for the >> outsourcing providers that servers are registered at the >> individual host level? Josh> It might be, but it doesn't have to be. You can abstract a Josh> host of hosts behind a single SAML SP entity. This is much less true for GSS. AT least, GSS tends to be aware of the individual host. Also, from a security standpoint, I'm more comfortable with the cross-domain credentials living on domain AAA servers than on each SP. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 10:20:23 -0400 From: Scott Cantor Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Organization: The Ohio State University MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-ID: <006c01cae870$45baed60$d130c820$@2@osu.edu> >> I think we should carefully explore how true that is for the >> use cases motivating beyond web SSO. Is it really true for >> the outsourcing providers that servers are registered at the >> individual host level? > > It might be, but it doesn't have to be. You can abstract a host of hosts > behind a single SAML SP entity. Correct. It's also not necessarily even "host"-based when you're talking about non-web services. There may *be* hosts of course, but those hosts may not be addressable in the same way and the metadata may not need to care what they are. What ultimately matters for queries is generally an identity and a credential (a key). The reason for using metadata in those cases and not just PKI are varied and I won't waste the time on it here. Furthermore, in the query case specifically, offloading the query to a shared system that handles the load for a range of actual app services is fairly simple. That's merely a twist on the usual commercial model of federation, which is to have a gateway visible to the IdPs and everything else implemented with proprietary SSO behind that. >> That seems surprising to me. Also, I think both for the HPC >> and outsourcing use case, allowing each SP to dynamically >> manage their servers will be important. > > I agree. It's just not the way that we interact with customers today (of > course, that's not a reason not to change). Most of the time server farms like that are sitting behind load balanced names that don't change often anyway. So even on the front channel, it's not *that* big a deal. And even if the names do change, as long as the metadata processing is automated, all the SP does is make their registration changes with a federation and wait a day. It just isn't that big a deal. It's a common perception that it's a big deal. -- Scott Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 09:11:21 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: holder of key for encrypted subject MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Josh" == Josh Howlett writes: >> I wonder whether we could provision a key that is conveyed to the >> IDP in the subject with the RADIUS exchange. That way, if the SP >> didn't have other credentials to use, it could authenticate as >> the holder of that key. The key could be transported to the SP >> in a RADIUS attribute, and transported to the IDP in the subject. Josh> Interesting idea! Some observations: Josh> First, the provisioning that we use today isn't just about key Josh> provisioning. It is also used to provide entities with Josh> information about the location of service endpoints; for Josh> example, the URI needed by the SP to approach the IdP with an Josh> attribute query. We normally provision this information Josh> alongside the key(s) in metadata. We could consider an Josh> approach whereby the entities concerned obtain this metadata Josh> from each other directly, and establish trust in the metadata Josh> by extending your mechanism to supply a proof. Unfortunately Josh> this mechanism would not support the current means by which we Josh> (as federation operators) make statements about entities that Josh> are frequently used by other entities to inform trust Josh> decisions (such as the LoA supported by an IdP). We could Josh> change the way in which we (as federation operators) make Josh> these statements, but this would require distributing signed Josh> assertions that contain these claims to the entity operators Josh> for them to include in their metadata. This would necessitate Josh> some fairly signicant changes to operational practise. In the scope of the beyond web use case, I'm not concerned about SPs needing metadata for the IDP; only for the IDP to be provisioned with information about each SP. In the case of the trust management use cases I think that this sort of exploration is very valuable. Josh> My second observation is that this reduces operational Josh> complexity at the expense of additional implementation effort Josh> (I don't think it would be significant effort, and I will Josh> immediately defer to Scott's opinion!). However, the initial Josh> deployment community that is the target of the Josh> proof-of-concept already bears the costs of operating distinct Josh> RADIUS and SAML systems, and so this optimisation doesn't Josh> really help. I understand, however, that it would help other Josh> communities. I think we should carefully explore how true that is for the use cases motivating beyond web SSO. Is it really true for the outsourcing providers that servers are registered at the individual host level? That seems surprising to me. Also, I think both for the HPC and outsourcing use case, allowing each SP to dynamically manage their servers will be important. A third observation that comes to me: the complexity of all of this argues for doing as much as we can in the RADIUS exchange so that we do not need multiple trust infrastructures. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 07:10:36 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: holder of key for encrypted subject Message-ID: Scott, Josh and I have been discussing the SAML interactions between the RADIUS server and IDP. Josh's idea to get around the RADIUSS MTU issue is to include a compressed authentication assertion in the RADIUS response and for the SP to make a SAML attribute query. The authentication assertion would effectively serve as a token for the SP to prove it was entitled to attributes surrounding the subject. Josh and Scott have been assuming that we would have the same level of provisioning today for relying parties. I'm very uncomfortable with that. In order to support our use cases I think it is critical that SPs should only need to be provisioned within their own domain, not globally. For example, I think we'd want to say something roughly like "Google is permitted to run mail servers," globally, "Google's mail servers get the e-mail address of XYZ's users," in XYZ's attribute release policies, but only Google needs to know what Google's mail servers actually are. I also don't think it is desirable to couple the Moonshot trust management tightly enough to the beyond SSO stuff in order to require the trust management to make this work. I wonder whether we could provision a key that is conveyed to the IDP in the subject with the RADIUS exchange. That way, if the SP didn't have other credentials to use, it could authenticate as the holder of that key. The key could be transported to the SP in a RADIUS attribute, and transported to the IDP in the subject. So, under this model, the subject would be very similar to a Kerberos ticket; in an encrypted structure it would include: * identifier for subject known to IDP * identifier for SP * Any group information the RADIUS server needs to community to know what attribute release policy the subject is in * Key that has been shared with the SP * Expiration time --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Fri, 30 Apr 2010 06:19:41 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Probably no audio archive for the bar bof Message-ID: I was hoping to post a link to the IETF bar bof audio. Unfortunately, as best I can tell,while the audio was streamed, only normal sessions were archived. So, it looks like we will not have an audio archive for te bar bof. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Wed, 21 Apr 2010 21:08:49 +0100 From: Josh Howlett Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Brief update Content-Transfer-Encoding: quoted-printable Message-ID: <6ED388AA006C454BA35B0098396B9BFB07082B4A@uxsrvr20.atlas.ukerna.ac.uk> 1. Closure of Phase 1 We have now completed the first major phase of Project Moonshot. This was in part a technical sanity check, and in part a toe-in-the-water to test the community's receptiveness of (1) the importance of solving our use-cases and (2) the approach that we are proposing to addressing these. We have satisfied ourselves that the reception to these has been sufficiently positive to warrant progression to Phase 2. I would like to thank everyone who has contributed their opinions and ideas. This feedback has been tremendously useful. 2. Progression to Phase 2 Phase 2 of Project Moonshot begins. In this phase we are focusing on the planning the implementation of a proof-of-concept test-bed that demonstrates the use-cases. This phase will run from now until end July 2010. We expect to have a detailed complete project plan by the end of May.=20=20 We will also be preparing the ground for a new IETF Working Group, which we feel is the most appropriate vehicle for addressing the problems under discussion. We definitely do *not* want this to be a "Moonshot" working group; we think it would be better if the WG were chartered to consider the general problem of application of federated identity to non-web protocols, and the range of possible technical approaches. Phase 2 will close with after the end of IETF 78. If we have satisfied ourselves that we have sufficient community support, a credible plan, and the resources to execute it, we expect to move onto Phase 3 which is when we start to build the rocket. 3. GEANT participation The GEANT project has agreed to commit some significant resources to Project Moonshot. This has allowed us to obtain access to some scarce talents, which is a fantastic development. More on this when the project plan is complete. 4. More rocket scientists required! We warmly welcome the participation of other parties. If you or your organisation has an interest, please drop me a line. best regards, josh. JANET(UK) is a trading name of The JNT Association, a company limited by guarantee which is registered in England under No. 2881024=20 and whose Registered Office is at Lumen House, Library Avenue, Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Mon, 19 Apr 2010 21:30:23 +0200 From: Rafa Marin Lopez Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: EAP over Kerberos Mime-Version: 1.0 (Apple Message framework v1078) Message-ID: <1E00E82A-4A8E-4D33-990C-983E2A439624@um.es> --Apple-Mail-4--807036317 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Dear all, my colleague Gabi and I thought about the idea (see attached slides) of = transporting EAP over Kerberos. That is, EAP is used as a = pre-authentication method for Kerberos (transported in pa-data field). = In other words, Kerberos would act as EAP lower-layer (see slide 3). In = this model, KDC is acting as EAP authenticator and can interface with a = backend AAA infrastructure. With the key material obtained after EAP = authentication (e.g. MSK) it would be possible to generate a shared = secret between Kerberos client and KDC. After thinking about this idea, we realized that moonshot document = talked also about the alternative of transporting EAP over Kerberos. = That is the reason of discussing this idea in this mailing list. Moreover, taking into account EAP-GSS, we thought about another = alternative (see slide 6). In this alternative, Kerberos uses GSS-API to = obtain pre-authentication data. In particular, the GSS TOKEN could carry = EAP as EAP-GSS approach specifies. We would like to share these ideas in order to discuss them in the = context of Moonshot. In fact, some discussions have been already = happened with Sam and Josh and some possible issues still need = discussions: for example, we believe that if KDC acts as EAP = authenticator, KDC will have to be stateful. Thus, any comment is highly appreciated. Best regards. --Apple-Mail-4--807036317 Content-Disposition: inline; filename=eap_over_krb_saml.pdf Content-Type: application/pdf; x-unix-mode=0644; name="eap_over_krb_saml.pdf" Content-Transfer-Encoding: base64 JVBERi0xLjQKJcOkw7zDtsOfCjIgMCBvYmoKPDwvTGVuZ3RoIDMgMCBSL0ZpbHRlci9GbGF0ZURl Y29kZT4+CnN0cmVhbQp4nH1Ty2rlMAzd5yu8LjQjKX4FgiHJvXcxu0JgFkN3fS3aQrvp71cPJ+nc 9g4BRbYlnaNjGVp0H82bA3cN7CYSG7zY93v358q9Nujke39sQA7cS8NB6j1LuJP/am33qXm40pLy cd60NBhTG1wHme1y536d0Pnsloe/AyBQ6QfowBfCAQJE/lJR1xcMA2ToYYTJojJMMMtBx6tyu/xu jktzcw6XkRsg79t4DncoyKWPcFIMhOKHjYBtYcGkSyRlxRQQ8CIWhb4lhym1ecWiqFjoIZQ4YCgE 3E25RqHMJqqbMNZNLxaT2lzPeqEy4lSz+gHnywxIGPi4d7syOBgC0o4D43ciP4ApPQuNu8VjJXmB Sopt73Jsu42INyKnDZRRWPKJPQKDFxeLlvWqFsm9bITJyIeSB+rsAHOdEMxkt0bBTiy4pswcN0Iu 4TwbO8s/cD53yWEqh5oqkp5o5WTklaKZQBcFQOx5xEPa57wqQH2hKDPNF0xMjYdcNRg3DSq5UFZN /mkJZeztDewsNUzK1KqjKZzrUhY0a42DSkFf8mWsO31X/2XBUzCv77BihL35G/cJ/jfnfAplbmRz dHJlYW0KZW5kb2JqCgozIDAgb2JqCjQzNwplbmRvYmoKCjUgMCBvYmoKPDwvTGVuZ3RoIDYgMCBS L0ZpbHRlci9GbGF0ZURlY29kZT4+CnN0cmVhbQp4nDWMzQrCQAyE73mKORdak3S1CMuC25+Dt8KC B/GmrQcV7MXXd1ORgS+TZBKuBB96g1Fyto0at8643HAq8CKBaZmJbYEn5dDqHhaH1T9/0ztNxfrS lO9iIuVdpajFmK7YDALnkKaz154PHHkvzG2uTRDPXZDMPpSZQ6jZayuddRo1hks6Up9opBFfkQgn +AplbmRzdHJlYW0KZW5kb2JqCgo2IDAgb2JqCjE0MwplbmRvYmoKCjggMCBvYmoKPDwvTGVuZ3Ro IDkgMCBSL0ZpbHRlci9GbGF0ZURlY29kZT4+CnN0cmVhbQp4nMV9y47lSJLdPr7irgVUiP52BwoJ VHV3CZjdqBvQQtAqpBlhECmgezO/L7o93I7xkcmsbsxUAZE03sPjfBw6zY3mxu09vP797a+v7fXT ti+2OP+WPP/+7f+8/sd/ef2/t/Ca///tX9+2+cPr69sOoqXPCX/Nf/Uvr/2/b//yX96297SN/tre y1b6/LflENqkSTm9l1fe4k7Fy2nkfbOUIyyHw/LEK4aWgUeaFLLe3odRt7Fbi3xZ4WzRdpM2mvVC Tt9MS+/dmqllt1Yzywpni7ZbzbD1Qk7fTBnvzZopYbdWM8sKF9bcbjXD1gs5fTO5rGbyK6X2Xq2Z ZQX3W1jbrWbYeiEnNDM3DvvGq5k4YdwMWuFs0XbczLJeyOmbCW3feDWzjfdszSwrnC3abjXD1gs5 fTPb/KvNxFHe02rGrHCyeLvVDFsv5HTNxD72jVczPWAzy5r/oqXbaTNivZDTN9PKe7RmanvXGy6B Fc4WbbeaYeuFnNDM3DjsfctXtUrardXMssLZou24mWW9kNM3k5s1koY1kYZR4nJuhslNuhddj8Qp GXHM70C9rHC2UoJdSEaf0oE+9NV7TWtb3Qpa4cLqq/da1gs5fTNbXr1XfIVR3+3wzQrut7C2W82w 9UJOaGZuvK1uZbd6XN0KWuFs0XbczLJeyOmbaXV1K7tV++pW0Apni7ZbzbD1Qk7fTI2rW9mtkrGZ ZYWzRdutZth6IadvJvfVrUxrW70XWlPxaOl2qxm2Xsjpm0l5dSu7FevqVgJY4WzRdqsZtl7ICc3M jbfVrexWiKtbQSucLdqOm1nWCzl9M1uFZkZfvQoY4WjwRqsNtl5I6NoYcfVc4dXzaqDbFYFFQq99 iNJ5JfsFmFtfvNX8FTDC0Wh9gVpfzA3vv1dN2pW8irknwQy3SOjVcJLeJOn6xZqH9hyvfU+E1RaD WxwG0EVh+HxbjGn5HvsDQzsiWwy4mMoCrEVhEMYQ8n4A7F7yMruRISRYjodldiMZw+6l8cjBC5m4 l2KJCynky4pnS9xLQYpDiZy+GXEvxRIXUjZeVjxb4l7qDiX5LYOFzYh7qRa7kPvGsBzdL3FttRoR dxIZfSPiXO7Wch+pibUccXltIShxMZHL6JdLScviNNJmazn6ZXEXeQfYrTQeJBYnkpbFTaSN1nL0 y+IgEkYcSeNBYnEb57I6hnMjW45uWV1CbpxdR+MBYnUUeTkY8VoOu8zMissNZBS7i8aE1OIczis+ lgM4LxZa4WyJcyhIcQeNcTXBtOwaiiXun2y6rHC2xDXU3QnyWwILm8nNGknDmpDlcFzOzTBZdT/Q NVSyZMTi/slmywpnKyXYhWT0KR3oxTWcVl/u39wQrXCwxnINFdlXjzW8ayjE4hqype4fb2xWOFnq Gkoz4gwip2tGXUOxxP2TZpYVzpa4hrpDm3YUYGEz4hqKJe6fbLyscLbENRSkOIPI6ZsR11Ascf9k 42XNPhStvlxD3aGoXQhY2Iy4hmpt2vtM4mWFC6tLtxEXy2oGXcO5qi3XUCxx/6SZZYWzJa6hIMUZ RE7fjLiGYon7JxsvK5wtcQ11h7b3aM2gayirxDVkSxxA3nYZ4WioayhtiDOIhK4NcQ3Z6HpRqi4G WxsUvfYhSueV7JfFXNk15MVq/gQY4Wi0vkBNXYEKrqGQpdWZVHUDeaNlhKMhrqDsQFr+THVOIq8R N1GMAPzLCCdjIGy8297LL8Cf1KeYhjiIvOEywtFIBWBJ3Ypkv3wrcvnf/9vbX1/xfdtGev37a3v9 0yu8/u0txt213eatP5cSuwN/fvvnnSWOLYx98zz27SdN2v9jmmmUDjQ1UPyMbu4yiYhnLpVAgdZt PvBwmRG23fR+v9Xu3/6Vorzz/33x17+8lT4H7Xnvw8vrL//79V9/m2Oe11/+5X/+HH/7kn4O25ef 4s8xfvkp/Lyluby1Lz/t63/78lP2v5btl9C//K+//NPbn/6y78HF4aW0ycGFssmuzyU+oPnXliQa lnSJD8vvexxj+uYX+55ox/YdSl/2Pftl+/VLyLL/Mc89f3oUP6SCRH3HPlbgMShZiR8Sf77nokty 4sqk1N7peTIHP/wU4JNw1l/bVH/7EujvirnmeSsRmpe/ja+1UoB94msdFDb81r6UOjs3HnnxMvDP sVuVZ3HisLg8U5MGydHKZNFGsEgYo9IOAfgmO4+KpjWX+Ze4lifHx02r/MwXUoN/Bau4XWZru7TI fTg0E8Fy+57ZfUhBlheNo/84NW1br/3WQyFRhOl4sfX5Uqs4K19QXTdTzR+h3o5+no0kIgpEOv3Y VCNd+OleNmiQf+vU18pWu2o/gU3oZx8y75W4/xtiKHX2Z63nJB1AHTSenS4F3f9R7/8tfNlv/zjv 8XnX73/zXFFosUo3hl0V9Cr8CL5k5e2Ju9Pi2LtIaeaXufgrLf6BjT/6zvBbBxLqdZuT5C//9s2u aCeg81YyPcljrrK8n1taTrAceZnwuMwPcePRy7vIiPhdrTnmmUtZlubmHxfNEdWbUCnw61rusI+8 XE7Lh+MAHrePkdyAfaP58oQsJYhA+3FoTrdc+8g7PT2WWKePzNYOFSs5K7KCjQbpgU33dHbNJQn9 9HjS7tBEIpwRk436g1jncMqamr9UPg28TenkXRsb0ufIwb59+J3E2p+iYgVnbUxKW+Fy4PO82DTo B5TcwBBr3sppvfPsq6mPm4ajqQLhX8Hqbq/ZqpcWxScPzXSw3L5vHLPkM0yWEUXXxMepeWNY+66H M0lzaVN6ZO2PKbGCs7YLqutmol1WeUQTYDZT5e1rJOL5DEmp87OxBqCS3zLfg7Jd5nEacj7sauee BdJ5uexr7Q/1idTd/ol6w9/uu9ocstyvJ9If6Gof97KZokCXzf3x0GHfd7Zy7hJ7ZiXMHoat6biy 1ZzFz0vdjqzF0txv3R59SMzNdLHKftL4l7CWG3kF18138woQ/hWs7PadrXhpTapjMxkst++VBzI8 ziJLiep+NNjEx6l5Y1j7rodDpBQ/YWuHi9WcVS+orpuB0Id4rASYzVAKQ+L7sVAvn1Kmm6+U7hqN 0qlMh0+2kz4IOaWZf3799RUyqT/OPZuZNpJtM2/ceF4dMp1GirRM2hM6vWaPwqvfgqRPTJc3XXHv j0Vj3mTlPnghmgN2DjPjYk4zVrivfs3+5Yh9ZdnryZvmqDvFsuPr2rWFnavneHvxzjtjru6vC2x/ VTvokGZcfK5ur7F2DdDt1ddB78yZVu4DgelDn8D1NeyoA13ruXr3ucPaPYCXSVONnAZqcToe5Nie 8PlFMb7Fv73z7uyXMZ+vDK1Piz0OWTvfJp+vDa3PdgZCpMfyvn4fCtfz9aH1ZV2itxm4o0v0Cu3i Ck2OZtyFLtB8oITTcdLato5/Z850iTLFfc/XKFOgFuQdGb07buF8kWj99g7nhVyhOLvNGM9XidYH O/4ZD6y0vrxiOl8lWh/h9gmDrhK5uOV8ldj1XeyBRvp0xOu2deiZxgK3UKh0leaFi+18lWh9tbMQ QqGrNOWyj71P14nWN7iVQuYr9UrbCTsZ7EbiDpBOSQpnfdF664Te9DXaPPfnq0TrA3QA9AJirp+9 3/kq0foI99I26CpRKC6frxKtT3AvbeTpRU69O2uMs9yMnTyYSPlV7XyVaH2FnmArdJVmfIw8rBM+ veipt/gzHQ+d0XG+SrS+w/200f5Fetd8Uhi/pYb7iaJK+/oX5dwc0S9KB1jM8zhnZ0QZeAcsrY/W B4xBF2mqP+fzRaL1yQ5/dLpG86bL5XyNaH22O2mQQxsplNvOl4jWGzc9ayjURa8jTuAySVYvMDJd oHkseZwvEK3vdgIG622evrKdr0+kOKjdRYPTEedpL/HcBdD6zU7ACLJ6Xk0Q49jm2n2dSeitdzrl ga7EYu6N9DB7thzhwnUalkSKqsV1lL3QuaKAebST9dbpnqVONCc4zM7uw7w9c0omoc6Owrz5cyqA Z6dgHlXeZbs6r8bs+1Fmewy+WuMj2tcu5sYP9HmOci52PI0f3hsJbgml0Xv9uJGuNrv4jYaY1H/n Yl30q/Fzdz7R5yWxVvn5Ov/Ocdfir/wknQ+RXOGBVDut3Y+yWlf04iEipUHnCg+Ays/AbV9rHfdb pWddmLdrbqvLrXQPh3l75WY33IuH8WFKLTfo5io9o8KgW8U6qLdCz6J9/VQ/dCxlXpVET6rcox0R hQvm+v1IO9zRhZ4iYZ6H3O1efGN3mR7GucN9VBIf1X5XGDdFVel1cR7g69ENNdd3isevbovuy0Tx vzwG3NKdjmneE2UD94j6iLl+H+dsGfsLOqa5VdnMKXmjHmuun/c0uoLUt4Z5hkoAR4A6zxRk7LTW Bt7LfZ09dt9Wwu9wWBqqkbvkmQudgfkY9nvSaG2l/V5qzHztmhynsQ9pcx5nhLPLbSY6i3aFgrQ5 E6vwinLuSaBrZHrJ3Oa8nnDlSNHkkcyrP0BFjdZ3UoupkTxMcgxzX/d0oUgsuZy543200fmmNwgd 77ogrWbSut2lUVrdj7QVuGeStLofaVv3Ot2Zc22Y9527e6nV/S6Fnovu2BSmkuY9ve6LSr1xmNcv VxvpvGYEjDyv2V9gvxNp/eyPoN9tnCxeqD+yfjFJm7M3gqdoy9LmfpwlQJuMjrNntLPSmrQ5+1dw MBvfRXuL2dbxHZep1wV99o3OYKbHuTneb53v6Hkf5QRDhs53P70rSnAndu4tMj1f4M7t3LtMLec4 4HlUpd3puaxz3ru0uh8peeu6fkir+5HGiE9RbnW/VQc4FzSuphSWHMClG4nWzjv05P/wehgzjCyr 541rzoXBm6MZFGLW19FHt4Dfo4PQB0XNQ2LXZZzZ9/XmQ9BTNMx+J9eTK8brYZw2Bq2cEoJea4Hn gw6GQNsm66e4Ti4kr4c7bnci5UDDK1sIAfinGGHouUU+0lc+hyfmWnCseaBAjugGfblgaT0OmbdC a/t0iE+Xk/zTDY+Tb37qSODuN7y8kFx4crs4SHU65bTWYgk05Gi0fg4toD9feBqK2ABrGyR/6hzy +9FH5fXQF80MhUDrZ16i3UWGnxFeHBpGOdrp3Jz8awmxwaAmJDna2cqZfbpgcLSBH68zE/HkMFNO Ig4l+RHNyYXmGyma1g88UhqOcdLYeTjG6zseKQ32KCYT6/m60vqGR8odIaV/nAeTvL7CsUZ2YQKH CE6qpPXwNA2R12Z+NX9iz68I/X3g4V/gSRrnXobnioAqKcQw18cXeaEn/viK8NwIFMdKpJ7Qz/3M XB83Q1c+1ldo526GGOA4J3JeaOcbCXbjsA8c55D1/SpgxevhCSbhsLBRetq5l6H1BY4zcZc/b/gQ z5qk9cBOgwcK+4GvBuzlFeBRMLPTAq3Pr7CdNblxMM80yXHIMB8b4yzJuTpsIEkOcYbZEV/seXxh PIyDp2Ga9dzPTLPBzcdB2ZlBc+5jthdG8WhSZ6JuKZ/6F1oNQ7VA033m6g4x4JehO0aMA78C3OZd fjwZtBIi0SHTaHxjb/Z4Nmg1PEYkVL5p/3/ixtU8X3U/nzlRJDWPvnFw/G//+lbKfiTUic+0Gtld ziwBw1D6WuC7m4212XrtOHehbXPX9jHOqH4XBjKNxRQ4cVEwliu5ry7kKOgLqf2Y9mGZ0VGqpuNY RgHCgoT8qqzwm66+tZ6RsCBhQcImrz+CaMAo6/7fpNzlO6n73stX4VxoYG3GGoyT1G2c+ymkN3E7 Vb7gJPTiDMBIZ89QGTk7+SNF3hL2/Z/uOTNw8ktTNapjhYTB2Ct5uEUuVZ8X3ZE2JK1IOoAUD16y +EocFCPsMZWInHjoKKOoMpKHbEROvh8KZQROzpaL43yPxhpRS1HedSkO1JRE4mW/8uWatSBrRtbm WPvrlINJb1930v2Obp60IynoCa99RD0l0VHJu/CJcz+5yBlRT3Dl07vnBD0l0dF+3Te6TPu9GTwn 6Cm9O9bqWDEBVS56mVd/slKH5lgbslZkHY4V+qQkl73skq/XrANZQVT5He/ShKJKcoVK3bsRYp09 H7AmFBUTLSM7VhSVnM9SM11+6qocaUHSjKTNkaKm5HTO46dOat6tjrQjKWiqAGV2mpKjLrv26UpR 1wqcGTUFfWlxmspOU3zMpemxnyhBUgUlVZyksqUZ7qSyg2U+Pibr3mNlz1qRtSJrd6woKWErU0vE Oh8DjnUgazfW6iRVnKSEbR449VP0IADWgpKqKKnqJFWcpPY7kx7J8m8/dFMFJVVRUtVJqjhJjV0A RMr/nkk7koKkkLKipDI933lg2zWJBZGLD9mcnCq6D5kemzuf8M6TWjwneBAN9dTei2OtyDofnJN1 Og+TdT4CHWtF1oKs3bGCnjI95XZW2edOT0HHOpAV9NSdnhrqKdMTabLyPp9YG+qpo56601NDPWV6 fExW9vAo3d6RFiTNSNocKegpi+taZZc7PbIda0dWkAD2+90JSnzRSs/QfeC2HdyTjqKCXn84UXXs ozI7OlXOwSDf1FFCHzVQU8NpqjtNiV9a6SF6yVqRtSBrd6yoKXFNq5zacXiUdJTUAEnFzUlqOEmJ b1rl3A7ye4F1gKSEaBnZsaKkxDut4pkP8lgca0HWjKzNsaKmxDutfM0GuSyOtCOpSSrC4Ufnmmfx TqtcsSNnRNc8wsE715zeyQEne6eVnkyT0z+eCW2coKkYUFNxQ03pAKyKCga5Vo61ImtB1u5YB7LS 9amigqFTbBAMpKCpiJqKzj3XcU7lW3+QPw2k6J0LzzKyIwVJ6UCnirQuWAuyZmRtjrUjKzuolZ4l O6smSSIaWEFTeP3x4MU5rSLWQU46MMKxw5VPONqhN/nGKI5pFa0OcimREfWUYLQj08UNh3oSx7TK 7KpBPqVjrchakLU7VtSTeKZV7oEL1oGsICjnmkfnmqtL3uQeGOT4Ayu65hFd8+hc8+hcc3HJm4Qj BjXiSAuSZiRtjhRGPOo/N+n4ZhuOsyFnQ0547kV4Qqmj2yQiMcjpB86MjPbUi8VJKjtJiWPa5E49 c6KkCkqqOEk511wd0ia36qBz4VgrshZk7Y4VJSUeaZOYxGiHc4queUTXPDrXPDrXXJ3IJh3AoLMM rOiaR3TNo3PNKT3SWMXda9IJDDrPjjUja0bW6lgbsrK71yQwccHakLUiK6oKrlUVB7VJ5zJoiAKc FRlBVc2pyjno6us1iUsM0gNyoqoaqso56BEc9Jlm/I2aft+YnfqtyYqniZ1Z8ifoaW9pzTRpTCzu kzHXyixOiaewo7rdEAb+hFeUwxi9xUjcEZ0y+ngmbCyUH0BZGjqbNOrsgJDmdIY5m+FLSj/HP8y5 oiF/CU1mle67E+d0WVgVUviVgSl92RH5yeyCUrlKG2uyUICDLc5Wn1blu10sntms25G1WAQJnJCh X6AZqjpBVqFXmHPTCMtdMvTPzUPJHYR/BWu4fWcrX1pdMvSxmQiW23cuwFMCZ8VKlo78UlwTH6fm jQFmF/DhUG/Y+O1j4VtXrOqsdkF13Uyx/pgnejBgNkM1MkbgnoxyRXaxzmWaXmVN0i+dHpW6VRcn DRifTuWJXJowqtCfTuWZE6X3vyHEb8+gTI32/qqBb0zr2e+XTP/y30L3ELUUqoHavin/xiw0F4j+ /GHNDno+IahTqvVpR59Nu9RTv73blDu2dCLXLLKDFjt7uh3Pg1MruN9gmIHEOnmNLZ6wxpV8bNkm 3h2b33DincG/Livivr50Kp1fDufjc2yHvU5rsuBcXoSzp13kH4dGdUs30W5np+nt88E3l+d0RV6O sJw8ARIDD8yYnD9N2k6bywTpxr0hE6e1Z7S+deqKZIvGU6KUCYmbTcjlZZ4eOise2bLM7CQ8LhMG eGCyqxFPd4itOcU1Ul6tLumEXN+cm5CrwK+yXG2vaKm5pXJxBGXtKexdW5OF57LQUBIGE364Rhjl Do/LaAw6vXQ5ZLnCcnMbHwnXhLc5o3n+sBOmja4wv8mMac6YkMm35KzK5rSeZ3vJFi3wFY6KMeIq lZgoDY4tniU/rYFWkwm2RALLjAI2qAfQXQMKCJTTPTdMsDykJsCpYSbVmgAG/6rWnHhve/n6lAlQ bvlAyfuMbH6vW5BSCZmWF+F7BfKPQ6O6JdQvIPb5mOyy+PniRdpEFoPfGlmBRB+/XAtmVq+mUjLz hiDPeaPkIa4mRjNmuQ1eX9n54C24PIExPXzmFkoIopvFPxH5gdbno4wetFytoFjNgvknDHrgkf2N ugXT7b1q4+FT9weqFdSLdvCh+dd9WAB7NuOVkWazHHbMPdNd+12cdX0VsOkbVZqLPA/0gi/8Mpnc GcL9KNvlfvwcfn3ytJ81WDgtgAthRIm8ihWGszTS7xcJY1Su+oexz6AeW1wqZF/WJWL4uG6TKbFu CQO+gpXcDrO1XVp51f+wZiJYuOeUZ891S3hZaeabCWzg49S4be8ql0QJ83J1kighYrFkY7Vo4wPV ZTMYaKYuNwxphEvpZalIQqMheRvHx2BNch0/8r11OwnEGePDPoGf1xTJ+Turl3zzbgkU2b9qR+78 P4Q/gnefwp9+oCcoc8B55n5etySMsp5xRSy9YGFUZ7X14PTLdT1EdXtQEjQgyyzIuVmC5bHEeWy2 +EeoAr4uq+E+rhvFL9fzcTm2wz731RHM5UX4XoH849CobulupBnypUcoVfTm52aSTWSx+62RFUjW I5R+oQfoLPYd9AE680sl7yyvFmh9GEkfoHN9lAeoQJ4V+qEBBE2Y/IFbhZ91bcZlfotU5Gv/N3xJ 3xg6plkHDEafN0+S0LgM+sUO0Ug5ylOVHrQh/J5iFVy989zEw7HpnP/NSSdyCSsH9tVKztLXSH6R MEa1JBoXw9dlRZGmUiDzx6lVY/A31iJVK7ldTm7jdKI6NrOBdWgmuWbmLcaHmV52vpLb8+Qo0+E2 qxmeV2zpE0o3Viu6oyLr7diMccLzSvc7SnnRDs+rUDd4XlmT/BtHWGW72tbzSnGPbsLOVezrSfN6 d33vkfXj9+Gzkix0JIGL+8ip09K3YgVnRTt+t0woYPN9sjUgyyIGSpe05W6iOTQLtYMQ/hWs4fZ5 mMBPVjfRrGY2sNyeR3lcabX9sYjoLzTxcWreGLzgg7zcpmdI4DGtGAGNeMFz3QbEvGafEzg2xU8N zeii5wlNhNPnjObO0i+Sss3bbE29bQU9kjnXdQ2UUvdjOrfnjQxsfnumX65EOsTjpYoQQxxeMqQM txj9XcvwHg2t3B2NQPqOhvxzmEhGpHk6NOlzLc4+76JVptOgysJ+NSPg7rJRL4zjUSGb293Ob5N2 gxeFYYKB+ePY5tp07azsPn2zgeZrDO4p2ZANxegnFsePbBrnpDkP9PPkn5Ls/Fo3pfno7rw4sKm5 fvBTgLcYHNkHrqdapXmxY2ZWXPhFYYbqMxU3JX+G+lwn4WhOym9f4iaBBhqQs44hZvCNbh3knriW VoxPhD+/3yAxrkhztyQGJsYAQ6JmcwtYlFjU4gEFJWOeEXMyIr1DpdjZWhyioUN7GDgD7FcxKPKj e0aL3S96Mg3zVfsFdrSJ0PeOu4VFRUIX1g/flmzkj5brOM+xHgW8eJGja7wY/LYnxmpdPJeNnnEy nhYncbJA5SYaZ5cEymrlBmi9xMl4C4mTLaangqbp6O3sY1wKGovCcbG5P8RK/ncjHWf75e/Tvr0I w57+udD5FWySR1rl+jzyUFJrQ0snFFX9WNlYr3UViZxydpF4NqPLNRRZn9fyRsq/bJwp34TS4F/V ku9d2L6ylc4WE78dmylg4Z5vnDBaAxce6rS/SX7promPU/PGsPZdWyXSwnPLOfAmlm4sVr+gumwG pxNxgWQGzGZonurGb21qodnFG2d6VHK3rVGbefyp20laA3I+HSNvPDVyvs24CrPyHTRfpcYOAgfv m++aOHZPklb8Cbr5R44M7zW/0P5K1czTer2tVnSW1C+X7biCuVrR/RbghAPxbIam/g4Wd5Bf+lqO Ipar5gOKxeBfwSpu34tI+8pisfhmAlhu3zk9rnI8kSwlos8/QRMfp+aNYe27Hg7Vdaz0QmtIaEas 6Kx0QXXdDKSRzgwd/nk2QjUEOcpTG32qhpfxxNEv4tvJFr1Lro9inr0/GZTmMS5ebjwUN9YOfTiq zK+Q6YNy1O3M5c+XLhdYphrXhLalspZWLXGjmZT8ExfS2OHvsDxfiF80A0QK/LqW5RN8ukdixQur +KMANnVp5UfOtuT5YDJddxFGR/5xatgYDgfPM2ESqYWtHS5WdhYnhnqq62ZgKs4svj8EMJvJ703q W07iRL/JZ0PHcI3yb/LxV9lOZiwA50O5RuroA5UTey5X6XfDH+Iv8VfLusk/2g3zlweC5OPxBNkg SXFqDbQkq1+2Y2uxDPwNp5TMp5MCvoLVZPYwE2ETH8fmgWHpJCCpWH7fu9u4X1AdjzKA5Zrprhku EEQvqqU6xSbrcd+7I+247wbXFAa2NHFBN1bLXYXDvvtjbu/g78e155ykECS3fqYvBJkTwWkNq0GK WnHqA23B00WB67GyudLpyT//vbq+65Z/5Lspret3RaJ+VkQ/mTL7ZFviHwX+wxmiheodXHxuZN/z mQ/a59dE8Psi8Y9f9tX0DZIkb/B/35dF9sefpOZ/hWU9gO8+PAtVMssXoyoazsQ/fKGXHakdr9f3 u5vr/c0pytuAr7D8dH9TpfIPuZ17z+/s75/uXtjMG26++0//qByEmeEULwm/k4QQqczfeUOfhaDz lfy/FCukBxXHImZcqbLxKbWnyeCvctOUNl2c/RcvUgbUGxtUIoe2o6qeRDeXPl7cDs/L47Btp24x 0jbzXyL6oJKwQb7XxE1KhHnQ4trbj2NJgyc3tjtax2tt8cGtPfrEnVy7rQfz8WYH+YmHDqdjnaUP O3uf/pzambYr8IF7+8N9S+jTBSyQe27vE8PUxvxQEf3d6vbL1r8ECkaMnyMnm/NrlW59z/abxOtC l1+3X7/ox454e1mueJPfCi/SxLOkn6+gqQb6jYqhBn01QlO8Er9QoCTGzAkTc5G/Kz7e5ZtGTBV5 pMONyLeOOF8jWtIJb8qLk/DjjY3EIRAx4koY4V2y/f5dEvTHDbzQmu3FB+7fp9tvOBo7yo83OP5P d17WudLz92Fn9RPONJx9uyofuN8/PhWC4tVUcVbU2FWMJKw8w2mPRNMC3SQ8I6AMNeZokMptUTE+ LogabZHPJBsiHTHmCeFN6SQxpYqnBTU+Jc3y7ky+XZ5J2kR0vPb794kGj9vxfuv+cHtqe7+O6uMN jvfTnQc8O+usfeD5/PTn2c6+XZUP3O8fFk3eCmt3iaZJltHsw9qMsoapm5/K9Iv2Tm33BZ9oiD6G QmPdQh8VznrstDgPqr5y1oVErjobSQphsDGHdLwZ1SJnMlr84DZS0hmFg5fnF8ay/MvfC8tSC5ES JHJm7L4Xuo8fb79DLHaAi07I9XB0Bz5hl9ZOrl3/eINj+nTHimdgnZkPPGefeCb13Or5/lj7+ONd yZgT1NTXWR2Jm1H1RAWFv5DFH0irlT+ow4Gq2tXgz9PR+zVZlABalT6ZwlxV+uQim9IdxJTSk3Bj 6zaLavCHPGhTWcx8W/IHPhTPX/sgUNSdhv3/XT2KO37HC62tvfjA/fv0+w1Hs47y4w2O/9OfFzhb 6yx+4Pn99Ocdrsa6Sh+4/z+soUpv8ql3cz2LDibbHHL+RGnSzzyZXeMU2qJ684XCTKPxIsc5StUF 8gL5u46cpicGFbukjShfnKlo8YMayDLjY5bmp4TyGHkjWqjy2jHOBjj9pFQCl/aS/ftdfck6tsUl zJEPZbX+iXu09jHrIcDxfPrjhKNfZ+XDztYnnsF1TuU8f+j+/Xg3QjMbaRdcR7JLYHA/sksg0zu8 /HOKMuuy8U/zZWFegCWROTmW6+Fvszct80081VT/TgnBRjWhN87d4QL5m3qgYABMYivf3ZDvJoHB SwNYszYUN4cN7sEXDOZ/X1YgBODiS5I+JkY1PjftPxaNAgifliEEMHBW4+SvJgnMVc+6q2wI2EXJ LMvoSFmR8rq2IWCBshslT0gXmJvzHqUgyQ0lz3h3LMsYSPmkXCKAgROUUlEpFSfR35ZLBPDirCii hiJy9ZhkZrpRSrVEwC7KhjpqqCNXNum2AiOAgRN01FFHrsJRlDIkx/qLgF2UHXXUUUdQ9fO+pCNg gRJ0NFBHrmbQbUlHAC/OgUIaKCRfd1N6z3vOjpwmJAmZMOxQI/OuTCSAlVNoxAggJHzJ/40ikQBe nAGUJOUyBeZK2yTpNhan1ogEMHCakqS0pcB8jcibspOAXZQRlCTppAuGxRzlHj+VnQQwcJqUpFyk wHzRxbtSlgBenJoRLMZATixld1fKEsDACVLKKCVfylB7eC1keeDMKKWMUiooJbzod6UxF3TxFZRR QRn5uoB3ZTEBDJwgo4oycjX8kkjyVBQTwIuzoo4q6qg6Hd0V2gQwcIKOGnRJUuBvcbJ+jmU2Abso G/RIoaGMXAW7u8qdgAVKUFFHFblic3d1OwG7KDuKaKCIsDeSS21VOx0hdEUDNTRQQ65w220ZUAAD p2kobqAhX2PtrgwoYJVSWJbRkRIlJCfwVAQUwMBpEooBJCTvGZXzrrAogBdnAA1F9JZ9ebEkx3wq Kwpg4DQRRXSko6/Uu0qVSlHRAyf62HH52Im+yGUqinjkd2VKF5T4jGIZDfhckVZhshKlB0b1Xo1G DPSxo69Qe1P2FLCLcvnYbHSkRBndFD0FLFB2o0QfOxZUERU/e0HBUwAtruVcszGQC+sdXhVQBSDw DeNDxzo6xzpLidNTAVUAL86KykHHOsIJtJKsWjzVM8L5a6ichsrxxUjvyrECGDhBOehVR183VCon ai3WA2VH5XRUDnrV0XnVVt/1UIkVwMAJ0kG3Orry+LfVXQG8OAdIKG3IiAKSQqnHyq6AVULi0EVQ UHIe9SoVK4d/YEwbKChtoKCEHnXCayMlUo1R6sQu6OILoCD6iIbxOQXdVZ4FMHCaghJ608l506vy rHCuyrMAXpwRJJTQnU7Onc5ae/NYzRbAwGkSSgkklJKT0E01W8AuyoQKcoSooFXJ9o4RJIR8KCHn SWcJQlhtXHSrUkYFZVQQOtIJr/dNWdyFXHQFBVRQQM6PtpK4QqglcQEMnCCgCuOc5Gthy619rLIL 2EVZdZjDRkdK1M9djV0AAyfoB93o5NzoLPGHO86GAmoooI6MD2r2AnYRwi6iD518wea7ir0ANkYU EDrRaSAjhx6M0XnRaRjfQAUNVRB/jw7rP99VAAYwcJqC8vKiOxmgIKv/q5xS/xfAypk36ILycqMZ hsVKJfBwqikMYOC065OXG93JwNKiNxWFAbsoAygoRySEy6OhxXtGu0AZDnt50J0MrKh6U58YsEYI CsrLge70lVgjvKtNvKCLDx3onEBB2TnQGq1bjJoZC2DgBAVlVJBzoG9rHQN4caIHnTMqyHnQt9WO AQycoKCCCnIutIbB7jjRlc7oSueKjCihuwrKAF6MFfhQQs6V1njVHSO60hld6dxQQ3gm7+oxL+ji Q1c6N9RQcxq6q8UMYOAEDXV9jtH3jTFep7El45RKzABenB0eZLmjhpwvfVvdGcDACRoaqCHnS9/V dgbsokRXumxIiBKSiMM9o0mobMAHEirOl76pFA1Q4wMBlQACKnh7S6zB6kQ7Qru3CzrSJYB+SnD6 uSs8DWDgNP2UCPop0ennrvA0gBdnBP2UCPopzpHWsMWp7DSAgdP0UxLopzhH+raUNYAXJ3rSxTGi gO7KWAN4MSLfu/VqJaOfWr/BlsFHFQoxCsoH2e4KYi/o4kM/uhTUT0H96BDeimEfGFE/6EeXivqp qJ/bAtsAXpzoSJeK+nGO9G15bQADJ+inoX6cIy0j7lN1bcAuSvSjS0dCkM9tvW4AL0bYxY7ycY70 bbVuABsjSmighHAfZSTb5FF2ZIQ9REe6oCNdnCOtw05jnDcRMKKE0JGuG0iobiahv6ee+LcyPU4p cXPokTdRY6bFT0nWkEI3jb7A7QwCrU3npJw/S/r9/F+zFOfP+TyHI4Qvg+akhPZzKl/SocYactB4 MNPnfD1HpO1mfnaSdOwys65dxsnsETN9AZnKSDxJOJlFTTmZ4itNkO9C8PnyluF0ptB3N5S8U8Jl u862wjZkqFkDNsRX0VcJJ4ZDvgF8HENgnPNnbxNODGuczKKWVPATHOYz3CScGNQomcSstiirz4u5 TjgxKFI2oORK/oxrOOK4SzgxqFEyiVkdKDEt5i7hxLDIiUrpoBTn1N4mnBjWOLsTUQcRdSeim4QT gyIl6miAjgZ22LcJJ4Y1zoE6ks+RKg51dJNwYtBFKSRmLR1J7sR3Ek4MipSgI/nGJ+OC09Fdwolh jTOgkKSWlOJQSHcJJ4ZFThCSfDiTcc67vU04MaxxRhQSfaVTUe7roTfpJoY1xoR8pqLgPNvbZBPD AqNTUTYVheSTYq6TTQxqlNmpKIOKnHN7m2xiWOREGRWQUXFvEO+STQxrnMXJqICMnJN7m2xiWORE GVWQkXNyb5NNDGuc1cmoWn8UqhPSXbqJYZETOiT5WCDjfNLFXcKJYY2zOSl1kJLzc28TTgxrnN1p qYOWutPSXcKJYZETtTRAS8Np6SbhxKBGOZyUBkjJObt3CScGRUpQklRiJlzcnJJuEk4Muii1FrRY AQjdl2evE04MaoQB6UxDMeBD7TbhxLDAiBqSL8cpDjV0k3BiUKOMKCGdhE646CR0l3BiWOQECcmX zBiXnITuEk4Ma5wJNRTBW/afH7tNODEscqKIwJGO2YnoLuHEsMbpfOxYgBFVdJdwYlhjxEsOPrb/ 8NZtwolhgdHJCHxs/EDWXbqJAo3O+dcR/Ovo/Ou7ZBODIiUqiDtdxjn/+pRsYhjj4q3N6sCFrxOv kk0Mh3yoGnCqo3Oqb5NNDGuczqmOAxjxVeJdsolhjXEg33tdjM6jvk02MSww8kxythJ41GlDxutU EwUuuuS86QTedHLe9G2iiWGRE2STgskmOXf6NtHEsMYZUD464VRwKJ+bVBODIiVcnQTedHLfvbxN NjGscTpvOiVgdPkwN8kmhjVGPOxkCkrJKegu2cSwwOgUBN50cow3qSaKND7nSidwpZObSXKbaGJY 5EQJFZCQc6XvEk0MapTFKaiAgorPh7lJNTEscqKEwJNO/pPhN8kmBjVK50inBoSooJtkE4MaIZ7I BgJqTkB3ySaGBUYnoG6ORcJdvEk1UaDRdfQpErjPybnPt4kmhkVOPOwB+nHu822iiWGNczgBDRDQ cAK6STUxKFKCfjL4z/nwKeebZBPDLs7sHOgcgBGvzl2yiWGNMSCfKogmMjof+jbdBNHAiirKy4sm JL6qu0s4MaxxOjc6LzeacSim25QTRCMvyCknlRMhnTN9l3aCYKNNqKicVFGCxHeLd8kniEZelNXy qAnpP8B7k4KCYKN1TnUujtR9zPomDQXRxlqQ06nLuda3qSiIBlanrqp9FCFdgstNMophjbNiR5Wr U1fFR91tOgqikRfV1Zy6nKN9m5KCaON1LnduTl7O6b5NTEE08qK8+jsqoTt53aWnINp4Ow6I8nCs Lo3mJkUF0caK+zqcvpwTfpumgmhgRX2VDfVVNmS9SVQx7OIUFrMacjp93aWqIBp5QV8loL5KcGkR N+kqCDZa55KXgPIqzim/TVpBNPLCJSsR5VX8F+mvU1cQa6wRxVWS40Rx3aSvINhI8QQk1FZx7vlt CguigdVpKzttOdabJBbDGmd22so6/GacT7e5SWNBNPLCMLwUpy036/s2lQXRxuu89VKcuJy/fpvQ gmjkRXFVJy7/FfFvcTqXXVjUaujJFDf9+za5BdHG25Y3M8hCfbnZKrcpLohmXmNSqzt9wXP8NsnF sMZpTjxbqK/ucinu0lwQjbwdeIfTl5sNfpfqgmCjNV+ereFo8STcJbwgGnmH8dYN5VWdS3+b9oLo xStMagWUWHUTw2+TXxBtvAElVp1zX910ltsUGEQjL0isRpRY/Q9Pgon7Hs98DT7wKsufnEFS1+jV luR32+46BSbQF8mnv5Eviw1G+mw9fej7N6poWg41UzJn23DttllybI6z7fC7xol0sKYJAYmqncwN +RvBhTJ1NBqPFiLlrD/ZmD19RbowfNdwmI5JNZsC0cargw+1huPtyCuROx0+ay4JopF3AC+H0hXp xg1do4yaYqDCRbTxamifLY6pG3Igr0REJcNC7zMEG61G99WqQOvC811Gu5oLot0NgpG2Ai3nrijS jRy6hpg1b0WfEIg2Xk2lUas73oa8EgzXFBt9UiIaeVFl3anMhe27hu3lFbQ6Cwg22u5E1p3IXOy+ 69sFzYU67G13GutOY8NpzI0dOgS6ijq3iDTOgfqSXBZDgr6a+EsakSvq4yN68a7UGrVAYD4vpol/ px160TEZopEXFCZJLYp00fwm3qg+gIqOTRFtvAEVJokthqzIy76zPjDLKs4OaOQFhQX5NpQgXZJM E0dfH/BFHyCINl5JAl/WcLwdeXlQon5O0cgSopEXNCaJLop0Y4gmgyj1yorGABFtvMnpLKPOfNKM nga54YpGQBFstNnJLDuZucSZJkNUS5IDzuwklp3EipOYi/Y3feuvrzA1NI9o4y1OYsVJzCXPqE+j Q5RV4A7RyIsSq05iLoFGfTsdUmV9K4No461OYtVJzEX/1RfVYWXWF2eIRl6UWHMScy8B1CXXMXDW 14aINt7mJNadxFwyjQ4gdLye9Y0poo23O411pzH3RkCHURpfyPIqG8FIizIbTmbupYCO+jTKkvUl PqKNdziZDSczl1ijI1+NCWVNYEA08oLM5IvbgvTZNTr+1whW1kQTRC9e/QL4sobjRZmt3De+PbKE 8xGMtKAySZFRpJuZqVEgDTpmTTBCtPEGVJnkyRgSVXbIUcyaBoZo442oshhRZT7nph6yKbOmwSEa eUFmki+jSPeqoK7MT5Zb1nRFRBtvQplF5937oid1ZanKBwGivy206oljUis7mTnHXx1zfWeQNVMV 0cabnczU8Y/HHJzllovYsmYTGxY5UWPq9MfXofiJviLQlzzZxk2KNU7n8McKjDiqFt9DJJZtzKNQ I8TLr45+PGbhqH+v78yAzXn40Xn4UT38eEzBUf9bh8PJRjmKNU7n3Uv6jeJQSPJuUF9oJhviKBY5 G3B2kVE8puOIP6/vXZMGfwxqlN0pqIOCnE+vvry+H042tFEscqKChgxA4utQA0WdeXmJnWxUo1Cj HDj4oCQcRaGAdDqKjL+SxmoNuxi5Booum4J8Ro7mEmhuQcJdTM6JT86JT8FE5BNydEigKRDJhjKK NU7nwEsSjuLc1CkeDkimRtK3KgZFStBQiqah5Dx3HQloikqyEYxijdP57SmaiHxKjkQPNY8m2SBD oUgJGpI8HMa5DHcdAGi2T9LXoIY1zuREhIwoInH9Nckp2ZBFscbo+EBEfu6meP2aipXiesYrFhid igqoyHnpGuTWlLFkAxXFGqfz0SURR3GootfKaUs2QFEY0qGCKijIOeY6gNBUvrQ8EIUapfPKUwUB OZ9cxw6ab5hsbKJY5EQFNVCQS2xf4wY5o8nGJYo1zuYU1IER327JIH3uYdQ8NsMZW0cuUI/zv3X8 oTmq0cYhigVGpx79bNnEubdvPELQPNpoIxBGGt9A/00ycATlX5Xy2EBzfaPmlBoWOUE+eTP5ZOdw y2tSTUg2/0qRizE7Zztvph6fjqNDDc3FjjbkUCxygnokBYdxvgKKDDM0ZzxqXrdhjTOgenIERrw6 MsCQxPZo4wyFGmFEOhPQIQtHRhaazh8L3ojZOdbZOdaSdcO4i4wDnXEQdY6FIo0voYD0I4mMcgKS YYpOi4g2rFAscqKAMgjIudI6RNHZG1FnABnWOJ0jncGRPmbd8GBYZ65EG04oFjlRQgUk5EugyNBE Z9hEL3XnSAuLWhUYUUIyKNF5QBF7tOwc6YyXGxzp7BxpHY3oZKXoxg/ZOdPZOdOSWsO4i1wrnaYV 3eghQzhbK6AsyyTkK6Cwg6pzyaK92VAoUqKCwI8+FECRoY1MeQM650Nn50Nn8KGz86F1YKOz8oIb PmTnQ2fnQ2fwoX3tEx3a6GzEYK8yFGuczokuGzBeVMjSOZPBjR6yc6LLhnymnuLT2mVoo28Tgxs7 FOdFF+dFS/USxl1klOr002BvLhhpfAHVU8CD9vVPdFyj826DGzoU50IX50IXcKF9/RMZ14jXF5yf X5wDXZwDXcCB9tVPNIFYJzEHe12hWOQEBRXwoH31Ex0s6fzt4AYPxXnQxXnQBRlRQcKkk8yDvaRQ rDE6PlCQ86B1rCTdT3ADh+Ic6OIc6FJAQI6Q+zHpNIIbNRSkK04/4DsX5ztr3r1WRAj2QkKxyIn6 AQfa1z/RcVKUsU1w44biPOjiPGipXKI4N3WDuaKMQoMbN2j5E8cyrfrS0iWM84UEhSvKiCnYmwjF MqexqNWBERUkTFFGNwE9lmJOtFKsZRCQLyQoN2HkHji4cYNWP3EkaoEPjeVPdNwVZfy52VsHRhqf +dBsmYJ8HUHetSijpc0NHIq50EYiVgUXuroygjrlKcrQZnNDh7qtQaKxmNWB86IKMPHZKwbFIR9c mQoOdAUH+j8sOaXRx8e2hNLRBJS1ixBF9pakpyDLPIBzskqs6l5cpaqsvzeVWvYuWIbXP16ohV9s cMpGnz1i9KVabtNcypZX3kemD0DmbfaUuGwYuW7f3ejdNvOJIncJKIYFxnfk5OwOxT1IEjGoUTKJ WW1RHhJEbjI5DIucDTg57ss4nx1yk21hUKNkErM6UEKfcZkSYTjk68DHXSPjnMN7m7ZgWONkFrMG cOILrrvUAsMi5wDOASpyXu/t63/DGudAFUkKh+IevKI36KIUErOqUh6yQO5epRsWOatwlpW5wTh8 a3r3sluRzGcMZnXjc+9L715HGxY5O3BGk9Eh6+PmlbFBjTIuFZWVqKE4VNHda13DIucAzmQq8tVR 7l69GtQo0xJRWQkainvyetSwxpmXispK0GCcDx/fvcI0LHKiigqoCBnvXjIq0viKU1EBFfnQ8d3L QMMiJ6qIHVfG+fJ/V2/vDGd8zGDWAD4s/Xf3ps2wyIkSaiAhX/rv7rWYYY2zOQ11YHzyDsuwxojn sYN+fO2/uxdOhgVGp58B+nH1CW/eDSnS+IbTzwD9+Np/d69xDIuccNySaUE4n7Fx+9rFsItTWMwa wIkaep1fkRgM6UA+kl3BOB89vnufYVjjDCifGIHxe+8eDGdsEblMOv7zM7dvCgwLjCgdyaRg3LlW 7zGqrzhjSygcqXsiKBf1u4u/GxY5G3BmEI6vnX0XMDescWYnnAzC8bWz7wLchkVOVE+xx2w8BI5v gtGGNc6Cz9lYgfFiUukxcGxQI8SLXUFAvnT2XZjXsMDoBAR+NH6F5iomqyCjak49TdUzZ0YcPkFz Fz9FNPKigroqiJD+MzR3MU9EG293Klq+tCAvPgRxilMiGnlRSUOVREgfR76LLCLaeAeqiVIpDOcq KNxEAhG9WIlnLaukJu7wYZqb4B2CgRRVJakYinRffbkJthnWOAN6bpJRsXBOXnehMUQjL8hLsioU 6b9RcxPPQrDRRlSXpGQY0gV3rmNQCEZaeA5LcoUi4cOP/2FVeOscnsw4BO9ihZgE12dZTjsuM2Zt e1OHl+pnMJOPzaRthmS2X0L6EvZ/tl+/hPwzx2linvORtjZnJz0LAPFLIB68H0JAvzLnNkv+7hTx y/yrq1wkqFDtmd8RCaqhr6BOoc+jS/QHlg0jV/e7G73bZs8iQYYFxnfk5CCO4h5EggxqlBJOWlZb lM8iQYZFzgac/BBi3KNIkEGNUh6Dy+pA+b1IkOGQrwMfDxQY9ywSZFjjlHjSsgZwPokEGRY5B3AO UNGzSJBhjXOgiiSIo7gHkSCDLkoNJy2rKuXDSJBhkbMCZzAVPYoEKdL4AkpIYjiCehQJMixygowk iMO4R5EggxplRBVJDEdxTyJBhkVOUJEEcRj3KBJkUKNMTkTZRPQwEmRY48xORRlU9CwSZFjkRBUV UNGTSJAija84FRVQ0bNIkGGRE1XEcRzGfT8SZDjjk1jSsgbwPYkEGRY5UUINJPQsEmRY42xOQx0Y n0SCDGuMeB476OdZJMiwwOj0M0A/TyJBijS+4fQzQD/PIkGGRU44bonhEO5hJMiwi1PjScsawPmd SJDBkA7kIzEcxj2LBBnWOAPKJ0Zg/F4kyHDGFpHLpPMwEmRYYETpSCSHcd+PBCnO2BIKR2I4gnoU CTIscoKjJjEcxj2LBBnWOLMTTgbhPIsEGRY5UT3FHrMPI0GGNc6Cz9lYgfFBJMigRogXu4KAnkWC DAuMTkDgR38vEqQgo3L+s8RvKs1+fBoJQjTyooK6KugHIkGINt7uVLR86R+IBCEaeVFJQ5X0A5Eg RBvvQDVRBMdwTyJBiF6sHFHSZZXU80gQgoEUVSUxHEU+iQQZ1jgDem4Sv1m4R5EgRCMvyEsCOIp8 FAlCsNFGVJcEcAz5IBKEYKSF57BEghT5nxAJKvwpqE1PqEaCZFmiP9RX4DJj1rbXkaBSKAfoIhIU f/uSfg4UDoqR4jwcnvmB+E+p8e+P/9RUVvxn7xGO0Z9Vi1M/LKP98Gg0SOZQziTR4JhbBpRGgL6/ 4Ttu6qNAUvtLv/CDe8IhoAOJWBICEiQGQlb1XylVpenLhjVOYVlWQ04XBXJlqbVOBLLS3XtgEkue X4K8jAR1SWhuq7MFtPEK07K644XbV+uadK1/pZEWACNtB1qJCQnSR4X0KwvSm7fVNQLaeIVpWcPx 4pheP9bBamirbwQw0g6gHU5crmKMxki6Fu5eMRJAG+9AhWl4aCFxbK+fEdIqTZoHiejFq0zLqsZ7 iBLpXbDCmDJBCdHIW4UX4kSC9FNO9fNsGsv0V02KxhyIltUdLQ7VRLSr3o/O2Uc08nbglUeaIOPV cL9JsKyusQGgjVeYljUcr5vDzOrSGjqHnk9CRwcisRKqzAePdI6w3hyr9hWijTctlVGUxrFefQMS v/HhSAeQZqR0ArucibrCkYduTEJIByaxihMYzg7SopMaktQZBIY1zuLUVbDL9YGkVSNUw5LLuQM0 8jbgrU5dflqqlnXVAiSHPkxCSgemZQ3H68Z1fPSrUIgOFQGMtKiuhl2ujyvp0E4LeuTliQHaeNs7 SqE7VgwOaKlrjagdejCJLjkeXXby8sXjtTa5DHksNAtoYHXyGtjfYoxplY+XgU4+nlnc07H6WrZQ Xj7dXmvxiz+Uj53XcPIaKC+NNDHSx5r0YwkaYlwBVQAvWiVa1nC0F/M2Vs0ELY6BaOQFeclHmBTp 4k7rkyMabjt0XhJ5OjCJFR0rXjJ9waUBtzXmBbSxRuREefkYlM6AljICWosBscCJ4pKaMIrEz5od XvakQ98V8egTiksjUYpDcSWoy5UO/ZZEog4sYmUnLBeLSocXPmvGP6KNNztlZXyK+3jU+liczjNf I0hAIy90BhqREqSLSa2P+mk87tBxSVTqwCRWdaz4qU2tEyRneM00R7SxogYqPsJ9cGp9GVNjcit2 CGhgxUe4lIhRJJzZ9alRifjFQ8cV8bw2p63mtOXiVOuxJYE5ndWMYKRFeXV8hPswlUY7NPC3An4A NtqOT3GpFGNIPAda+E+CdMduqztxdSeu4cTlglRR383KKC+sNziANt7hxDVcN+vCVEHD8Bqsa4dO Zjh5DexmpXKMIH2gKqx3qiy04GklUHUgEiugvjBQFVzptBlT05cvC2ucAfUl1WMWDvUVpJ/RiF3Q QAuikRfuMSkho0gXqAoiWJ0pGQ79l0SqDkzL6o4XFBa0ppXOvDz0XxKqOjCJlbC//c8IVe3Ox2x6 01OqoSpZTu7B5S3Gre2vw1W7zpY3c4wofRkSkAr1y0/l5zAkVPWLJC79ZnElc0CCJobooIzmwvN8 ua9v5Ibo5Dk2Ot+4/M1zZwgMCOYxSDd/c3avv77+D92BP9PxfuMiuo+rb69MEcPDfL8/fuka7fuN Tiqf2t++hKphu1Pw7hsnWaKSSV9U0hfMXiuiWSGXBZcZs7b9Xaf3H9f0D53YGWftUlvi0bn9gQhp al3io2k+GZ5GSHl0xRt+fUutaaD4hcuIks7kwYbvuGnB/vgqQopAIHxHSp3pyMgnEVLDGqfU5l5W Q06XgHYXIUU08jbg1XTtBlHO70VIEW28UrtvWd3xPoiQIhhpO9DqzMcm4UiXj3YTIUW08Upx7mUN x+ty0q4jpAhG2gG0w4nrWYQU0cY7UGFrFmTjaCSw3sVHDbs4tTr3sipyuvy0u+goopG3Cu9YQU1F PoqOIphpjcis7mgxQeQuOopo5O3AqxnbjHTf47mNjiLaeKU6ybKG430QHUUw0g6gTagwX1H7NjqK aONNS2FDwpqKQ86b2KhBjTAjnRPXs8goooHVias4cbk6fbexUUQbb3HqKtjdBv9BnrvoKKKRtwFv dep6Fh1FtPFWp67q1OUr991ERxGMtKiuht2tL6Z9Gx1FtPG2d5RDd6xPoqOINla8cbuT2LPoKKKB 1UlsuL72+vOax/gooo13rP52rLimIZ9ESBGNvCCxNZ+SkM8ipAhetFqSe1nd0V597/cYIUU08sJl k7imIp9FSBFtvAElFqNjfRIhRbSxRuTEzvbw5frrCCligRO7WoltKvJJhNSwxplQXGu+peC+GyFF JHKisDL2tYdv199FSBFtvBn7WolrGvJJhBTRyIvKKk5ZzyKkiDbe4pRVsK/1dbRvY6SIRl7obCW2 qchnUVJEG291+mpOX0+ipIY1zub01bCz9WW176KkCEZa6GsluKnIR1FSBBttdwrrTmGPoqQIRloU 2HA97bMoKaKNd7iudjiBPYuSIhp5QWAS3BTksygpghetEKkF6noUIzWsMQbkQ209jJAiGlhRXBLX VOSzCCmijTeiuiQmasgnEVJEIy/0txIhVeR/QoQ0jUZFt/iUWuDEkvf4HfjBYtTa9jo6mvq4iY5y ZIm/KPcwEBopxzZX+ARehU/gDagnhsuCwa1/V5zuH9v8D8XqIiVo5dafxupkzR/o7P46g8/boG/3 bXimV4BKOvOm90MI8/123ubTvHNJtnlEvJg4oBBS98sEsS1/1zn+RzX8Q2c3bNMHnoOik0j/tJ/d 8Os8c/x3q/sZpJTTmwTTMWfhzEDk6W3Ad0rMzZExFXve/5crUCUo3BNnBPOJCCsMTAePy0lPRP39 0eh/VMM/dAXGHPxfhaLDH7/8NPhNCn2GsrKyVewxzDnh9xcjU596cTHmdvslmFf0l8gh7DLf1vz6 JUuGsX9vcHunpLxB3cNE82G1fqFYiZ8iEZcUAVv/rqv1j23+x17M0C3Ik9quajGebpubi5TIOZwB 6yPPwxcOhztHDioldpHkJOiXhsiq60TYkiJg67/nivyDmv+xK0LjpKsPuYY/7nfPvH/Cj91AiXJ+ 5pcerq/x1mdnSH+4V/uVe7Wrm+ifX/8fRpg9UgplbmRzdHJlYW0KZW5kb2JqCgo5IDAgb2JqCjE3 NjYzCmVuZG9iagoKMTEgMCBvYmoKPDwvTGVuZ3RoIDEyIDAgUi9GaWx0ZXIvRmxhdGVEZWNvZGU+ PgpzdHJlYW0KeJylVkuLIzcQvvtX9HnBHVWVXg2NoO2ZOeS2YMhh2Vv2QUgC2Uv+fuohqdvtsXeG YGjLsur11Vef2o0w/Hv4Z3DD0fEyoTyDl+ePL8NvH4a/DzDI58e3g5M/hr8OfEhXf8rxQb7b03a/ H75+UJfyYbvT5YDkxjT4EMY8XH4ffnmBwfvh8vXTDFAgzY5KnHFxqcAML46cd4vDAmF2oUwzEjjI 5fPl18Pz5fBx59wTZ+sBRzTfNEzjxM5nB+Xyx2sWSerzLnJONRtEzQYn5wvEGRxiOcLskjsVnN2k P0iWCSKeOOsj8boQHwWU/VMBkGx5uZQjzpwvigkv0ashf4mXs/yw5VL0HC8xykF16J7U07NGeFFL c3tWm5M+p+6c7fk4km7nkiWHvqGw4UzRJUoMpJloKs2eU6GasiDuWpkWSK3AFS+53cFf0aTsRr9D k3IhaL0taQbiyIIdtJpq8YZGwiDtl7/RVTgetZwCvavh5NMY9w1v6VDSbrjYHxX1XBGvCVPbv5sb Oj/SNphvwRzTSnEXktU+uaik58zvucvp2l3PnalI1qwsDtGQbuleg4+txKNStVN624ZeIn8c+1Sr YC4BrrlsEBnR2Sgba1AfuBkWTEpD5O7j8rin1iOENsZrj3x1GCUWTPN2HKLNnqVPlhS2cZo4bP9z soGodjbmtGnrJgxNKyhaUiXoZmAZfGkiTiJPwSi08Bed5HGWx5P8bwR/C51xoqaObyQ05nwzc22G mjzxAB9pbtw+dnKbaNGqOgqLEmrTZW08Y5Y2FNuToolFvJZJUaK83cnMyEkinOVs09cuPg1Wsdtr nB2z9K1domUWM7FSMsInetYCdLKe5Oagl5+rFsYbsqWqmyeh7FZEOXKHT1OpY/VqfDkp2r3rP83e aT+mdXxI91G9N22sOr4h275drRtHHTiqGo3SmC1j5ah0y4Oa0nr3WJB20cFpJYhKxlXliRehhJ+i SeFWpFovV2b0Cw4EIyuvZt+vn5qj4DFtt2AjLvWmXiX5+oJs57iE9PB+tAY9HE2Xx/fdNehQjl1T 67mQ7yCcTXZqS29fM271uwvU1ftDf02KJr9WUrQQaa5XxKsKJrTh+LkN6GqPTc6MuR7LY/nr7NZy vPzlw9ayC4kkEndF7F89HpIMUhzDvVv84UsF5xxeeb3b07O+393LYZIcAjThZT4oG/AOGxhOngnw PenA3KhvA61wiQwNDl0G245W1OZgrhMqOctOVPyE9U7unPZDwTQVUlniOYMWcDHibw4uIgRQMNZz zW/WcSO5FrTbrAEPUeG7O78LFYhtSDoqEP8HIhAhdRSsRJBLx0CASWpYVjzeUufH4T+NO804CmVu ZHN0cmVhbQplbmRvYmoKCjEyIDAgb2JqCjEwMTQKZW5kb2JqCgoxNCAwIG9iago8PC9MZW5ndGgg MTUgMCBSL0ZpbHRlci9GbGF0ZURlY29kZT4+CnN0cmVhbQp4nDWMMQvCQAyF9/yKNxd6JunZIpQD z7aDW+HAQdy0dVDBLv59LxV58OUleQk7wYfeYJScbaPGrTcuN5wKvEhgWmZiW+BJObS6h8Vh9c/f 9E5Tsb405buYSKvaNaikdop0xWYQeI80nVvtec+Rd9wEabkLktmHMnMIFbd6kM46jRrDJR2pTzTS iC+pVya6CmVuZHN0cmVhbQplbmRvYmoKCjE1IDAgb2JqCjE0MgplbmRvYmoKCjE3IDAgb2JqCjw8 L0xlbmd0aCAxOCAwIFIvRmlsdGVyL0ZsYXRlRGVjb2RlPj4Kc3RyZWFtCnicxX3LjuW6cuU8v2KP DZxs8U0CBwmcZwOeuX2BHhg9ym67YWQ1cO/Ev99iPBgrpK0qZd2CfQ6QxdBeWqSkJTIYfGh7DY// ePnrY3v8tO3JFuffkuffv/2fx//8h8f/ewmP+f/f/u1lmz88vrzsIEp9TPhj/qt/+ej/ffnXf3jZ XtM2+mN7LVvp89+WQ2iTJuX0Wh55izsVp9PI+2kpR0iHQ3riFUNp4JEshay312HUbezWIl9WOFt0 3qSNZj2Q02fT0mu3bGrZrZXNssLZovNWNmw9kNNnU8Zrs2xK2K2VzbLCE2uet7Jh64GcPptcVjb5 kVJ7rZbNsoL7LazzVjZsPZATspknh/3klU2cMM4GrXC26DzOZlkP5PTZhLafvLLZxmu2bJYVzhad t7Jh64GcPptt/tVs4iivaWVjVjhZfN7Khq0HcrpsYh/7ySubHjCbZc1/0dLzNBuxHsjps2nlNVo2 tb3qC5fACmeLzlvZsPVATshmnhz2uuWLWiXt1spmWeFs0XmczbIeyOmzyc0yScOySMMoMZ2bYXKT 6kWPI3FKRhzzK1AvK5ytlKAIyehTOtCHvmqvaW2rWkErPLH6qr2W9UBOn82WV+0VH2HUV7t8s4L7 LazzVjZsPZATspknb6ta2a0eV7WCVjhbdB5ns6wHcvpsWl3Vym7VvqoVtMLZovNWNmw9kNNnU+Oq VnarZMxmWeFs0XkrG7YeyOmzyX1VK9PaVu2F1lQ8WnreyoatB3L6bFJe1cpuxbqqlQBWOFt03sqG rQdyQjbz5G1VK7sV4qpW0Apni87jbJb1QE6fzVYhm9FXrQJGOBp80sqDrQcSujxGXDVXePS8Muj2 RCBJ6FWGKJVXsl+AufXFW81fASMcjdYXqPXF3PD9e9SkVcmjmHsSzHBJQq+Mk9QmSY8v1jy05njs JRFWSwaXHAbQpDB8vCzGtHyPvcHQisiSAZOpLMBKCoMwhpD3C2D3ktPsRoaQIB0PaXYjGcPupfHI xQuZuJdiiQsp5MuKZ0vcS0GKQ4mcPhtxL8USF1JOXlY8W+JeaoGS/JbBwmzEvVSLXcj9ZEhH90tc Z61MxJ1ERp+JOJe7tdxHymKlI6bXGYISFxO5jH65lJQWp5FOW+no0+IucgHYrTQeJBYnktLiJtJJ Kx19WhxEwogjaTxILG7jTKtjOE+ydHRpdQk5c3YdjQeI1VHkdDDilQ67zMyKyw1kFLuLxoTU4hzO Jz6WAzgfFlrhbIlzKEhxB41xZcG07BqKJe6fnLqscLbENdTiBPktgYXZ5GaZpGFZSDoc07kZJqvu B7qGSpaMWNw/OW1Z4WylBEVIRp/SgV5cw2n15f7NE9EKB2ss11CRfdVYw7uGQiyuIVvq/vHJZoWT pa6hZCPOIHK6bNQ1FEvcP8lmWeFsiWuoBdq0ogALsxHXUCxx/+TkZYWzJa6hIMUZRE6fjbiGYon7 Jycva9ahaPXlGmqBolYhYGE24hqqtWntM4mXFZ5YXaqNuFhWNugazkNtuYZiifsn2SwrnC1xDQUp ziBy+mzENRRL3D85eVnhbIlrqAXaXqNlg66hHBLXkC1xAPncZYSjoa6h5CHOIBK6PMQ1ZKPrQ6ma DHY0KHqVIUrlleyXxVzZNeRkNX8CjHA0Wl+gpq5ABddQyNKqTKq6gXzSMsLREFdQCpCWP1Odk8hH xE0UIwD/MsLJGAgbr1Z6+QX4k/oU0xAHkU9cRjgaqQAsqVuR7JevRS7/x39/+esjvm7bSI//eGyP f3yEx7+/xLi7ttt89WcqsTvwzy//tLPEsYWxn57Hfv6kSft/TDON0oGmBoqf0ctdJhHxzFQJFGjd ZoOHaUbYedP7/Vq+f/s3ivLO//fkr395KX122vNeh5fHX/7347/9Ofs8j7/867/8HP98Sz+H7e2n +HOMbz+Fn7c001t7+2k//ufbT9n/WrZfQn/7X3/5x5c//rKX4MnlpbTJxYWySdFnii9o/rWURMOS pviyfNnjGNM3f1L2RAXbC5Te9pL9sv36FrKUP+ZZ8rtX8SkVJKo79r4C90HJStxI/PM1Fz2SE1cm pfZO7cns/HArwDfhrL+2qf72FOjvGXPN81UiNKe/jq+1UoB94msdFDb8WllKnZUb97w4Dfyz71al LU4cFpc2NWmQHK1MFp0EScIYlVYIwDfZuVc0rZnmX+JKT473i1y5zRdSg38Bq7gis7U9tch9OGQT wXJlz+w+pCDpRePo309Z29mr3HopJIowHS+2Ph5qFWflJ1TPs6nmj1BtRz/PTBIRBSKdfmyqkR78 dC8bZMi/dapr5axdtR/AJvSzDpnvStz/DTGUOuuz1nOSCoDUmNp0Kej9j/r+b+Ftf/3jfMfnW7// zfNAoWSVaiz9HLmaYBz/HW9ULcTt5zDmj6G9ERjrNaiCpnfWnheBM6OC9EUtufwyk79S8jc2fvc1 59euOobneU6Sv/z7V+utnYBucsnU7MdcJb0/CEonSEdOEx7T3OIbj2phkRHxq1qzgzRTWVLz9Pcn 2RHVi1Ap8MtKdygjp8spfbgO4HFljOQz7CdN9ZClBBFo3w/Z6ZmrjFzo6d7EOh1qtnaoWMlZkeVu NEgPbFrSWY+XJPTTPUq79xOJcIZXNqo8Yp19L8tq/lL5NvA5pZMrbmxInyNHBve+ehJrb3LFCs7a mJTOwnTg+7zYNEIIlJzBEGu+92kNkPaV1ftFxtFUgfAvYHVXarbqU4uCmYdsOliu7BsHOPkOk2VE 0WXxfsreGFbZ9XImaS5tSo+svU0TKzhre0L1PJtoj1XacwLMbKoM1UYing1OSp0b0hqASn7L/A7K eZk7dch5s16eJQuk8/K0YrY/VCdS3fwH1YZ/Xle1OWR5X0+kn6hqb9eymUJGT7P7/VBhX1e2cu8S u3ElzBqGrenlstWcxY2rnkfWYmnut27tJBJzNl2sst80/iWsdCMX4nn23VwIhH8BK7uysxWfWpPq mE0Gy5W9cq+HO2VkKVHdrwazeD9lbwyr7Ho5RErBFrZ2uFjNWfUJ1fNsIE4i7i0BZjY03yHx+1io lk8p08tXSneZRqlUpnco50kdhJySzT89/voImdQfZ8nmtByZmjNf3Hg+HDLdRgrLTNoTOj1mjcKH X4LMtZj+cXrGvTeLxrzJwb2nQzQH7OyTxsWcZmBxP/yY9csR+8hS6smbZhc9xbLj6yraws7Ds3O+ eOebMQ/3xxNsf1S76JBmEH0ebo+xigbo9ujronfmTAf3XsN0uE/g+hh21YGe9Ty8O+hhFQ/gZdJU I6deXZyOB3nBJ3x+UEBw8W+vXJz9Mebzk6HjabHHIUfn0PP52dDxbHcgRGqW9+N7v7menw8dL+sR vcwoHz2iR2hPntDkaMZd6AHNBiWcrpOOtnX9O3OmR5QpSHx+RpmiuiDvyOjdcQvnh0THt1e4L+QK xVltxnh+SnQ82PXP4GGl4+UR0/kp0fEIr0+gbkgkF7ecnxK7vos9UFiArni9tg4957zAKxQqPaX5 4GI7PyU6Xu0uhFDoKU257B3103Oi4w1epZD5ST3SdsJOBnuRuAKkW5LCWV903CqhFx1zm/f+/JTo eIAKgEYr5vFZ+52fEh2P8C5tg54Sxe3y+SnR8QTv0kaeXuR5emeN8ZQ4YycPJtJkrHZ+SnS8Qk2w FXpKM5hGHtYJnx7U6i3+TNdDd3ScnxId7/A+bVS+SAPTJ4XxkDa8TxSC2o8/aILOEf2guQOLeV7n rIxout4BS8ej1QFj0EOa6s/5/JDoeLLLH52e0Xzpcjk/Izqe7U0a5NBGivu28yOi48ZNbQ3FxWjs 4gQuk2TVAiPTA5rXksf5AdHxbjdgsN7m7Svb+flECpraWzR47uK87SWeqwA6vtkNGEEOz6cJYhzb PLofMwm99E63PNCTWMy9kR5mzZYjPLhO3ZJIIbi4rrIXulcUXY92s146vbNUieYEl9nZfZivZ07J JNTZUZgvf04F8OwUzKvKu2xX5dWYfb/KbM3gozW+ov3oYm7coM97lHOx62nceG8kuCWURpMA4ka6 2uzhN+piUv2di1XRj8bt7mzR5yOxXLl9nX9nv2vxV25JZyOSKzRItdPR/SqrVUUP7iLSnOlcoQGo 3AZu+1GruF8qtXVhvq65rSq30jsc5uuVm71wD+7Ghym13KCaq9RGhUGvilVQL4Xaov34VD9ULGU+ lUQtVe7RrojCBfP4fqUd3uhCrUiY9yF3exdf2F2mxjh3eI9K4qva3wrjphAsjS3nAb4evVDzeKfg /aq26L1MFCzMY8Ar3ema5jtRNnCPqI6Yx/d+zpaxvqBrmmeVzZySF6qx5vH5TqMrSHVrmHeoBHAE qPJMQfpO62jgUu7HrNl9WbODh8NSV43cJc9c6A7MZtiXpNHRSuVeasz87Jpcp7EPyXNeZ4S7y3km uov2hILkOWdh4RPliSqBnpHpJXOe83nCkyNFk0cyn/4AFTU63kktpkbyMMkxzH2902WQLqbLmTu+ Rxvdbxpu6PjWBck1k9btLY2S636lrcA7kyTX/UrbetfpzZxHw3zv3NtLue5vKdRc9MamMJU03+n1 XlSqjcN8frlaT+cxI2Dkec36AuudSMdnfQT1buOZ5YXqI6sXk+Q5ayNoRVuWPPfrLAHyZHScNaPd ldYkz1m/goPZ+C3ac8x2jN+4TLUu6LNvdAczNefmeL90fqPne5QTdBk6v/00sJTgTexcW2RqX+DN 7Vy7TC3nOKA9qpLv9FzWPe9dct2vlLx1PT4k1/1KY8RWlHPdX9UBzgX1q2m+Sw7g0o1ER+cbevJ/ +Dj0GUaWw/PFNefC4M3RDAox69j10S3gQXcQ+qCoeUjsuowz+37cfAhqRcOsd3I9uWJ8HPppY9DB KSGotRZ4NnTQBdo2OT7FdXIh+Ti8cbsTKRcaHtlCCMA/xQhdzy3ylT7yOTwxj4JjzR0FckQ3qMsF S8exy7wVOtqnQ3x6nOSfbnid/PJTRQJvv+Fl9HLhye3iINXpltNRiyVQl6PR8dm1gPp84akrYh2s bZD8qXLIr0cflY9DXTSnMwQ6Picx2ltk+Bnhxa5hlKudzs3Jv5YQG3RqQpKrnbmc2acLBlcbuHmd 0xZPDjNNYMSuJDfRPBPRfCNF0/GBV0rdMZ5hdu6O8fGOV0qdPYrJxHp+rnS84ZVyRUhzRc6dST5e 4VojuzCBQwQnVdJxaE1D5KOZx/FP7PkRob4P3P0LvKLjXMvwwhJQJYUY5vH4IC/0xB8fEdqNQHGs ROoJ/VzPzONxM3Tla32Edq5miAGucyLng3a+kWA3DvvAdQ453p8FrPg4tGASDgsbzWU71zJ0vMB1 Jq7y5wsf4lmTdBzYqfNAYT/w1YC9PAI0BXMqW6Dj+RG2syY3DuaZJjkOGWazMc6SnIfDBpLkEGeY FfGTkscHxsM4eBqmWc/1zDQbvHwclJ3Tbc51zPbAKB6tAE1ULeVT/UKHoasWaG3QPNwhBvwwdMeI ceAhwG2+5cebQQchEh0y9cY39maPd4MOQzMiofJN6/8TNx7mxa37/cyJIql59I2D43/7t5dS9iuh SnzOwZHi8jQUMAylwwLfPG2s09aw4yxC22bR9j7OqL4IA5nGYgo8y1EwNrFyP1zIUdABqf2a9m6Z 0dG8TsexjAKEBQl5qKzwSFffWs9IWJCwIGGT4Y8gGjDKuv83KXf5Tuq+1/JVOBcaWJuxBuMkdRvn fgtpJG6nyk84Cb04AzDS3TNURs5O/kiRUcK+/9M9ZwZOHjRVozpWmF0YeyUPt8ij6vOhO9KGpBVJ B5DixcuUvxIHxQh7TCUiJ146yiiqjKSRjcjJ70Oh6YOTs+XiOF+jsUbUUpSxLsWBmpJIvOxPvjxn LciakbU51v44Tdik0deddH+jmyftSAp6wmcfUU9JdFTyLnzi3G8uckbUEzz59Oo5QU9JdLQ/940e 0/5uBs8JekqvjrU6VpytKg+9zKc/WalCc6wNWSuyDscKdVKSx152ydfnrANZQVT5Fd/ShKJK8oRK 3asRYp01H7AmFBUTLSM7VhSV3M9SMz1+qqocaUHSjKTNkaKm5HbO66dKar6tjrQjKWiqAGV2mpKr Lrv26UlR1QqcGTUFdWlxmspOU3zNpem1nyhBUgUlVZykss1J3EmlgGU2H5N1r7GyZ63IWpG1O1aU lLCVqSVinc2AYx3I2o21OkkVJylhmxdO9RQ1BMBaUFIVJVWdpIqT1P5mUpMs//ZDNVVQUhUlVZ2k ipPU2AVApPzvmbQjKUgKKStKKlP7zh3brpNYELn4kM3JqaL7kKnZ3PmEd97U4jnBg2iop/ZaHGtF 1tlwTtbpPEzW2QQ61oqsBVm7YwU9ZWrldlYpc6dW0LEOZAU9daenhnrK1CJNVi7zibWhnjrqqTs9 NdRTpuZjsrKHR3PzHWlB0oykzZGCnrK4rlWK3KnJdqwdWUECWO93JyjxRSu1oXvHbTu4Jx1FBbX+ cKLqWEdldnSq3INBvqmjhDpqoKaG01R3mhK/tFIj+pS1ImtB1u5YUVPimla5tePQlHSU1ABJxc1J ajhJiW9a5d4O8nuBdYCkhGgZ2bGipMQ7reKZD/JYHGtB1oyszbGipsQ7rfzMBrksjrQjqUkqwuVH 55pn8U6rPLEjZ0TXPMLFO9ecxuSAk73TSi3T5PTNM6GNEzQVA2oqbqgp7YBVUcEg18qxVmQtyNod 60BWej5VVDB0PQ6CgRQ0FVFT0bnn2s+p/OoP8qeBFL1z4VlGdqQgKe3oVJHWE9aCrBlZm2PtyMoO aqW2ZGfVSZKIBlbQFD5/vHhxTquIdZCTDoxw7fDkE/Z2aCTfGMUxraLVQS4lMqKeEvR2ZG254VBP 4phWWYo1yKd0rBVZC7J2x4p6Es+0yjvwhHUgKwjKuebRuebqkjd5BwY5/sCKrnlE1zw61zw611xc 8ibhiEGZONKCpBlJmyOFHo/6z00qvpmH42zI2ZAT2r0ILZQ6uk0iEoOcfuDMyGitXixOUtlJShzT Jm/qmRMlVVBSxUnKuebqkDZ5VQfdC8dakbUga3esKCnxSJvEJEY73FN0zSO65tG55tG55upENqkA Bt1lYEXXPKJrHp1rTtMjjVXcvSaVwKD77FgzsmZkrY61ISu7e00CE09YG7JWZEVVwbOq4qA2qVwG dVGAsyIjqKo5VTkHXX29JnGJQXpATlRVQ1U5Bz2Cgz6nGX9lA8CvLGX92srG0yrQLPMnqLW3ac20 wkwsrpNxrpVZPCWewo7qdkMY+AOGKIcxeouRWBBdX3p72WwsND+AZmno0tOoqwNCmssZ5mqGt5R+ jr/NhaUhv4UmS1D34sS5zAwOhRR+ZWBKbzsi31ldUCpv6caaLBTgYItnq0+r8tsuFi+D1vPIWiyC BE6YoV8gG9qigqxCQ5jz1AjpLjP0z9nD/jwI/wLWcGVnKz+1uszQx2wiWK7svFtPCTwrVmbpyC/F ZfF+yt4YYHUBXw7Vho1HHwu/umJVZ7UnVM+zKVYf80IPBsxsaEONEbgmo7kiu1hnmpZXWZb0S6em Us/q4qQB492lPJH3MYwq9LtLeeaq6v1vCFGWW14s60mNSv8sg68s69nfl0z/8t9C7xDlFKqB5tpN /o1ZaC0Q/fltrQ66vyCo01TrU0HvLbvUW7+92pI7tnQh19yRBy129vQ8XgenVnC/QTcDiXXxGlu8 YI23/bG0Lbw7Zr/hwjuDf1lWxLI+dCmdT4fz9Tm2Q6nTWiw404tw1rSL/P2QqZ7pFtrt7LQWfjZ8 Mz2XK3I6Qjp5AiQGHlgxOX+atJ1Ol9XUjWtDJk6rZHS8daqK5IzGS6KUCYmbLcjlNC8PndsjWVpW dhIe04QBHljsasTTHWJrLnGNNK9WU7og12fnFuQq8Iukq5WKUs2lypMrKKukULq2FgvPtNDQJAwm fHeZMMpdHu+5Mej20uOQdIV0cycfCdeCt7mief6wE6aNnjCPZMY0V0zI4ltyVuV0Os6rveSMFvgJ R8UYcZVtm2gaHFu8pH5aA60mC2yJBNKMAjbYPKC7DBQQaE73PDFBesgGAqeMmVQ3EDD4F7XmKn0r 5eNDFkC59IGSy4xsvtQtyL4KmdKL8LUC+fshUz0TNjsg9tlMdkl+PDhJp0gy+LORFUi0+eWNY+ZW 17TvzHwhyHPeaPIQbz1GK2Y5Dz5e2fngM3gvA2O62eYWfal9e8jNWZ8NGTWzvLFBse0N5h/etYDt H7XFwbyKZwW630B/bnsDrg6O+WE7+9e9JwElnCHOSAtgDnfsl3VbxBVwpeji4+sIwqYDsbO1z9eM wd2r4HbTeX123s/h1zsuwtzlhecS8FYbUcK1YoXhLB0e8EnCGJXbX8TYZySQLd6MZE9rihjen+fJ lLgzCgO+gJVcgdnanlp57TBi2USwsOQ0OZ93RuG00szhDMzg/ZS5ne/2RokSG+b9T6LElcWSk9Wi kw9UT7PB6DTV02FIJrxZX5Y9T6gLJUN4fA2WJe8USA67nifRO2O8WZE0Pnu+v/+Z+6N89cUKtDfA s0JJhfFb+B36Dyn88anKYyNP68R+f2+UMMpqR4tY+nzDqM5qq3H26boaaj0fhAcZSJr1O09LkB5L y8dsi2+mFfBlWQ3LuN4rn67n63JshzL3VW/M9CJ8rUD+fshUz3Tv3QwrUzNNW4xz25zkFEl2fzay AslqpukXaqTn7uNBG+k5h1XmtuWVAx0PI2kjPY9HaaQFcm/nIeqk0KLMT7xZ3Eg2eoX43dr/DW/p K93TNDcmgx7uRcMTGu/L/qRA1BuP0hzz+xu+Z0MM3k70nMXN/u9cY84TW+QRVh48UCs5S4eqfJIw RrUkGhfDl2VFkaZSIPP7KVdj8C/WIlUruSInd3I6UR2z2cA6ZJNcNvMV48tMD7tfyZU8Ocp0eM1q huaNLW3Q9GS1orsqsl6O2RgnNG9a7ij7nXZo3kLdoHmzLPk3juLKebWt5k1xt17Cztvq15Pm9e36 dgv32ffw3rYvdCWBNxCSW6d78YoVnBXt+l2aUMDm62TLQNIiBpqSaeluojlkC/sTIfwLWMOVeZjA T1Y30axsNrBcyaM0V7r9/1hE9BeyeD9lbwxe8EEG0KkNCdxvFiOgEZ/wPM8D4mqzzgkc/+JWQ2eN UXtCi+20ndH5ufSLTAvnc7amzrmCbsmcN5oNNG3vczq39kY6ZX/e0y9vjTrEQaZdJ4b4x2TIvuBi 9FfdF/ho6Fbi0Qik7mjIP0dgyIi0FogWlq7krPOe5Mp0GrhZ2C9mBCwuG/WJcbwqZHPF7TxitRuc FIYJBub3Y57r1FVYKT59RILWhAyuKdmQE8XoJxbHj2waS6V1FfTz5J+S7Dx0nNJsujsnB2Y1jw9u BfiMwaMHwHVXq7T2dszZG0/8ojCHAzLttkr+DNW5n+hk+GDDV6p1kHvi/bpivCP8+UEJiaNFWh8m cTYxBhgSmZtnQFLiXYsHFJSMeUblyYg0TkvxuZUcoqFDfhicA+wXMSi6pCWjZPdJT6ahxGq/QEGb CH2vuFtYVCR0YX33eclJ/mp5Y+kZCKOgGic5gsfJ4M89MVar4nkf6xmL46V3EosLtKVF4xksgWbO cgZ0XGJxfIbE4hbTXUHTkvd29jGeCho3nuMN7X6LlfzvRjrO9svfp30bbMOa/r7QeZg3SZNWeQ8g aZTU2tDSRUtVv5421tCxIpFT7i4Sz2w0XUOR43mlN1L+08yZ8kUoDf5FLfkAh5WVrXS2mPjlmE0B C0u+8aTUGnhzo07lTfJLd1m8n7I3hlV2zZVIC69f5zidWHqyWP0J1dNscMkS79jMgJkNrYXdeGSo FlrBvPFskkrutmVqq5s/9DyZOoGcd/vIGy+/nCMmh3cnrsAxDdfGDgIH75vfmjh2T5IO/AHV/C1H hkvNg+ZfaHv1tIbQ1YrOkg3V5TzeUl2t6H4LcMOBeGZDy4sHizvIL32lo4jlWfYBxWLwL2AVV/Yi 0n5msVh8NgEsV3aeglc5/EiWEtH3qCCL91P2xrDKrpdDe0dWGjQbEpoRKzorPaF6ng1MVZ2zgPjn mQntU8hRntro2zmcxhtHv4hvJ2f0LvOJFHNvjGbQVJLxZFTkprhxf9Kbvcr8CJm+cEfVzkx/PDRd IE2bbhPaUmWl1ubmRjMp+SferGOHv0J6Dro/yQaIFPhlpeWbgFoiseITq/irADZ1aeVHntHJa85k SfAijI78/ZSxMRwunlfbJFILWztcrOwsnnzqqZ5nA8t95tcAhgBmNvm1yR6akzjRb/Id0zFcpvyb fI1WzpNVEcB5U66RKvpAW5bdl6vUu+G3+Ev81Wb25M9Ww/wphCBz/ngRbpCJd2oNtGTlgJzH1mIZ +BsuW5mtkwK+gNVkhTITYRbvx+yBYekkIKlYvuzdndyfUB2vMoDlsukuG96EiAbDZQeMTY5j2bsj 7Vh2g+s0CbZ0coSerJZ7Coey+2tur+Dvx1VynggRZP7+nCIRZN0FT51YGVLUiqdX0Bm8JBW4biub d1M9+effq+uravkzH3JpXT90EvU7J/oNl1knW4p/FPinZ6EW2lPhyfdP9pLPOad9ft4EP3gSf3/b D9NHUZLME/i+T53szZ9M//8Cab2AbzaehXZLy096VdSdib+90WBHasfn9e3q5nl5c4oyGvAF0nfL myptMZHbufb8Rnn/uBqwmS/cnLSQfuSkhbkG4JLyetbCnPjz7ESctjAnh/N+kNvsjBXapnaOtX5j C41Ge6JtHFfmjSM37iA6A2Dy3n/zRHbWBOYW2MeiwpBllrr7BYAXJ9OIwSMzC4abc6z9L3RHDdkA A8CLM8nIhxjVON2qmLUBhuypoXtVABYoq1HynuICc2vLZVmAbX8hW1UAdlEyyzI6UsLyhSjrtWz7 i1EdZUXKbpS8zFxgbkWIbVQhe1/oRhUAXpxFBo7EGMiJu4lcbVQBYOAEIVUUkltlkeRhr20qdEcJ AC/OikJqyIiL32Vlke1RIftJAHgxNuBDEbklxbabhO5QIfUBgI0RVdRRRbg+/WoniQVdfB0lxEMd CnIbCcj9s10kZL8HAAMnXPdADbnVtEmud+0hods9AHhxDtTQQA3BZ2qvN3sALFCahCT4wrDgVqhe bvYAYOUUGjECMqKEpHBrrwfdlQHAizEAH0goBKwvkzCddmQAsDGChGSvFIHd2Y1hQRdfhLpSdjRR kJPQxU4MgAVKU5DsnyKwhHUlPem1B4O09gBcfAnUI7uDLBioJ8u2BqdNEwAMnFZTylYeAnOrBy+3 TADw4swon4KMd7ZLAPBiLMAHtWRwK/Gy7BFw3NQAsEYIlaTMfxMYrumXtfy6cG5taLCgi6+ieiqq x+88odsZSPu4tjMAMHCCfBpUktjnge0MdIcE2XgAsIuyQT0pmzcsGCpIlrHbZgbDUzakBAF1FFB3 AlqbDshOBrrpAIAXZ0cBdagmQ3ebA+iWA2sjA9lyAMDACfXkQBENJ6KLDQcAuygHiEgiBQyLGxCK n2zbDbjKPG6LL24gItkJQEHY0tq2ALrZgGwLAGDgtGpSluwLLKCIrjYFAOyiDCAiWVq/YLgs+mJL AMACpYlI1lkKLLpF4eKnrQ0BNLAD4MUZobaUFesLhtsBrMX7uh2ALN4HMHCaiGRlucASiuhy+T6A F2dCFWVkxBXRV0v3AbwYM/ChirJTkfhAtmxfFtgD2BhRRQVVVJyKLpbXA3ZRFlRRgRozFqwxLxbX AxQYrb6UVcoCq05EV4vrAbw4K4qoooiqE5H4F2txvS6DBzBwgoga1JixYY15uQgewIuzQY0ZOzKC iC4XwAN4McIL2VFEHUV0ufgdwMaIIhpYXWIZuW47rVFfyEU3sLYcKKCBArpcoQ5g4DQJpQ0klOA7 3t+/Pv3ZKvR5i/gjtRLWkNciU6Zyk+rBINA69dbHlgP9L4tK6AMU5G34eF0Ib4Pij6H9nMpbOsyo Rw76HBp3uw/fa6bz6vaLrCjvb2V++G7R0FfESqVYQucpQ3cCOHORHIcfvtBkiC4E89PNaBkO1ud9 /UTxVwmXXXzgKoBjWONkFrMGcGJ84CqAY1jkHMApTivhClYVVwEcgxqlfENXLPFbBYdxposAjkGN kknMqouyujDTRQDHoEhZgVLcVsLh1qVX4RtFGh8zmGVKaS7GdBW8MSxyoojY22VcdwGCq+CNYY1T pi0uawAnBgiuwjeGRU4U0QARua16LwM4hjXOgSqiUQRFYYTgKoBj2MXIIxGaNgn5LXAvAziGBUbU kOwmyzjcBPMqfKNI4wuoocC+rqBchOAieGNQpGxAGU1Ch91fr4I3hjXOiBKSWcaKQwldhW8Mi5wg IdntlHHJSegqgGNY40xOQhkYUUJXARzDGmNGPpCQ36H0IoBjUCB0CiqgINz47ln4RlHGVZx6CqjH +buXwRvDIifKp4J8/K6cV8EbwxpndfKpIJ/qQgRX4RvDIifKp1lzG5zHexXAMahRNmxtQwdCDBBc BXAMa4wd+UA9fu/IqwCOYYHRyWdYUxswDnYRvlGg0Q1sZsMABfmNGC+CNwZFShCQLj2duMOWiVfB G8MuTmExqwMnCugqeGNY5ISnE4MJyO9CeBW8MahRBhRQjECIAroK3hjWGCPyWRUZ3ZjgZfDGsMCI VWRMJqCI/cSL0I0CjS6hgGQPO0G5yMBF4MagSIkCyiAgNxx4GbgxrHFmJ6AMAsouvnQVuDEscqKA CgjIDQheBm4Ma5zFKagCI0YGrgI3hjVGrS4GRQlMQdUp6CpwY1hmNBa1GijIDQdeBW4MapRtqYgt U5EbD7wI3BgSGRswdmtmo9sL+DJwY1jj7Ku5ZasDJ4aXrgI3hkXODpwDRORGBC8DN4Y1zrFExNYA TpDRZejGsMg5jDNtJqTkhgQvgzeGXZzColYwISU3KHgVvjGoUQYUUgrW5Ca3C+xlAMewyAnvUIom pQQO9d+zweDXZkQdYztxf/NnrIEvukr6g6MfCYq2uk/8u52nkR0fegm0Oxt7G4f5Vn/M+VaRtvCj Tc/+pJlXBWcrNe6f8Kwb+qTjLMGXF/6KKReGkhUjTs5IEn7iU2WS36fjXj8k40/FvUgQOZ/nqYXf 93s271qVW7j9KWtjwtdCYJG+MngKgM2zeM7b9kvk+W3lbcwgWKbtJI7z3dwToYefZLk3CUWb7Xkz oBeLacHYud/1TH5c1p96KnNjn53IZrWdH4usVjo+FzvceUlS1G0sf+Wg47P7TTPIModR6bOKs8Us sz9hFUDXjoP6fxpRSzT3a57IO4bRlzNX5BEtRK55wd8+mfuqinTRyK6b9usG5ho5RLTxalRTreF4 O/JKt0u9Z41yIhp5B/ByOFGRbjy+66cLNO5H7SLwyoi8Y1KryuRlRQ7kla6nhCi1PUCw0WqEU60K tC482cUx1y83adOFYKStQMvhRUW64fmunW+N0WrLjWjj1VCnWt3xNuSVT2NIR6Kol4Fo5EWVdacy F67sGoCQL3epQ4Rgo+1OZN2JzI3Vd/0yiIbUD6XtTmPdaWw4jbnhevyOQVGvFZHGOVBf8rUuQ4K+ mkSGtLdb1MFG9OIVJrNAYMGN2zfxNfXDE0U7F4hGXlCYBB8V6cbum8TINIRQtCOEaOMNqDDZFMGQ FXnZ31YHrKxlFIBGXlCYRCEV6Ubxm0QJNSpTtFlAtPFqVFSt4Xg78nKPQ/fOLtq1RjTygsYkGqlI N5rfJFKq/nLRsAKijTc5nWXUmY9t6m2QF65oAATBRpudzLKTmRvYbxIstk/aAWd2EstOYsVJzA3s N+lzatemaJQK0cZbnMSKk5iLdjb9VJYGZTWchmjkRYlVJzE3xq99Be3eZQ38Idp4q5NYdRJz4/za r9GuaNaoJ6KRFyXWnMSak5iOa4gmskZoEW28zUmsO4m5EX/t2mlnPGs0GdHG253GutOYG/fXbqiO cWQJoiMYaVFmw8lsOJnp5yJFwlmj/Yg23uFkNpzM3BwA/TqBhk6yjksgGnlBZhLPFGTcnMz0q3wy 5JN1XAbRi3fFV9Uajhdltob3+PXIEnVEMNKCyuQLKooM2OHX8I5GubKOdSHaeAOqTD7aYkhU2eEr ilnH+RBtvBFVJl8oUWREldXDaGzWEUlEIy/ITL4mokg34Va9eo1KZh2RRbTxJpRZdN59TCizusaj ZelO9K9FTCiz6Bx/iXoq0jn+6phrYDbrODeijTc7manjH1cUdLGKWy5iyzrEb1jkRI2p0x8pjSEl /USpKC1bv0mxxukc/liBEQNK4nuIxLL1eRRqhPj41dGPKwi6CNm/18klwOY8/Og8/KgeflwRUOUT /1s7ucl6OYo1TufdS9xTcSgkGW7Q8YxkXRzFImcDzi4yiisGujhZ6/LUkwb7DGqU3Smog4KcT6++ vA7lJOvaKBY5UUFDOiCR0qggceYzP/dkvRqFGuXAzkfagBAFpJOJpP+VNGps2MWYNuQzBSXnx+vA mg60JSxick58ck58Ciai5Fx47RLoV+SSdWUUa5zOgU/BRJSc+67dARlgTDrwYFCkBA2laBpKznPX noCOgibrwSjWOJ3fnqKJKDmvXSLG+vHAZJ0MhSIlaCgl01BKPrDNHQAd/k06gmVY40xORMjowtrs +usIdbIui2KN0fGBiJyXrl6/ft8xxdXGKxYYnYoKqMh56TpIoGPzyToqijVO56OnAipyHjpdNT+i ZB0UhSEdKqiCgvzkW+lA6Ic90/JAFGqUzitPFQTk595K30FnTiTrmygWOVFBDRTk595qv0HuaLJ+ iWKNszkFdWDEETbppM8SRp3YYDhj68gF6nH+t/Y/dDJLtH6IYoHRqUc3GJg4LJ/0EJJcebQeCCON b6D/lgYoxznd2jfQKTxR58YYFjlBPnkz+WTncMs4pX661/wrRS7G7JztvJl6snO1tauh3xiO1uVQ LHKCenIw9WTnZ2s3Q6dYRZ1aZVjjDKieHIERn450MOTzytH6GQo1woh0JqDsXGvtWSQpaiz4Imbn WGfnWOdkAsruk4Xc5dbpb1Gn5inS+BIKSLczYZQTkHRT9Hva0boVikVOFFAGATlXWrsoOu0v6sRJ wxqnc6QzONLZOdL6QUmdnBitO6FY5EQJFZCQc6S1a6LfPY9e6s6RzsVJqAIjSkg6JTopM2KNlp0j nfFxgyOdnSOtvZEkRY2u/5CdM52dM50bSMitGOKeiH7sPrreQ4Zwdm5OQuBIZ+dIi4Oq02WjjWwo FClRQeBHZ+dHa9dGNn0AOudDZ+dDZ/Chs/OhtWOjs4SD6z5k50Nn50Nn8KGz86G1a6OzmYMNZSjW OJ0TXTZgxElq0rmJ8m9wvYfsnOiyIZ+ppzgnWrs2OkYYXN+hOC+6OC+6BFNPwUlq0rOJwhxs5IKR xhdQPQU86OI8aO3XRPk3uK5DcS50cS50ARe6OBda+jXi9QXn5xfnQBfnQBdwoEv0E+m4Qxvl32DD FYpFTlBQAQ+6OA9aO0tRujjBdR6K86CL86ALMqKChEmXPwQbpFCsMTo+UJDzoLWvJNVPcB2H4hzo 4hzoUkBAjpDrMak0gus1FKQrTj/gOxc/l1eYdD1LsAEJxSIn6gcc6OIcaO0nRenbBNdvKM6DLs6D LvW1AycqSLii9EKD6zeU6hRUV9Sq7hZ40OUwl5e5ovSYgo1EKJY5jUWtDoyoIGGK0rsJ6LEUc6KV YqVBQH4ur7yEkWvg4PoNpa+Iu5GoBT50ebKkPUr/c7NRB0Yan/nQbJmCnA8tPa8ovaXNdRyKudBG IlYFF7r6ubxSuChdm811Heq2OonGYlYHThSQlI/4bIhBccgHT6aCA13Bgf5Pm57VptOoE7SkUDoF axURosjekglayDIv4DxdK1Z1L55N1lp/LyYZ7VWwdK8/v8yOBzZ4ykbnT4u6hXaX01zKlte8D170 njf+dLOlDSPP7ZsnvdppfqLI1QQUwwLjK3LK554Fd2OSiEGNUj/1rFZblIcJIhczOQyLnA045QvK Wz7NDrmYbWFQo9SvJ6vVgRLqjKdTIgyHfB34uGpknHN4L6ctGNY49Qtgag3gxAGuq6kFhkXOAZwD VOS83svhf8Ma50AVyRQOxd0YojfoohQSs6pSHmaBXA2lGxY5q3CWNXODcThqejXYrUjmMwazuvG5 8dKr4WjDImcHzmgyOsz6uBgyNqhRxqWisiZqKA5VdDWsa1jkHMCZTEV+CdvV0KtBjTItEZU1QUNx d4ZHDWuceamorAkajPPh46shTMMiJ6qogIqQ8WqQUZHGV5yKCqjIh46vBgMNi5yoInZcGefc36ej d4YzPmYwawAfeC6XI22GRU6UUAMJOdf3cljMsMbZnIY6MN4ZwzKsMeJ97KAf5/teDjgZFhidfgbo B8t4NTakSOMbTj8D9ON3brgaxjEscsJ1y0wLwvkZG5fDLoZdnMJi1gBO1NDjPERiMKQD+cRg8vGz NC7HMwxrnAHlEyMwfmvswXDGFpHLpHNYwXY1UmBYYETpyEwKxuEVP4/qK87YEgpH1p4JykX9ruLv hkXOBpwZhONXsF0FzA1rnNkJJ4Nw/Aq2qwC3YZET1VOsmY2HwPFFMNqwxlmwnY0VGF14+3ng2KBG iA+7goD8ArarMK9hgdEJCPzoiDfySUxWQUbVnHqaqmd/coela5fxU0QjLyqoq4II6RawXcY8EW28 3alo+dKCfLKU9hSnRDTyopKGKomQPo58FVlEtPEOVFPaHCuuqb2KBCJ6sRLPSqukJi65nSGugncI BlJUlUzFUCQ+sKtgm2GNM6DnJjMqFs7J6yo0hmjkBXnJrApFul0iruJZCDbaiOqSKRmGdMGd5zEo BCMttMMyuUKR6UfsofS5GM+M2dCuN1LECjEJW7xEdQWmGbPOfb4Mb46sR2HysZnEi45+4R3Of9l+ fQtZVibhLuj3AkA8CMSd90MI6Ne1e/pPIcwd48OzDdXnvS4pfV8kqIa+gjqFPtAm0R9IG0ae7jdP erXT7kWCDAuMr8jJQRzF3YgEGdQoJZy0rLYo70WCDIucDTi5EWLcrUiQQY1SmsFldaD8ViTIcMjX gU+/9dRvR4IMa5wST1rWAM47kSDDIucAzgEquhcJMqxxDlSRBHEUdyMSZNBFqeGkZVWlvBkJMixy VuAMpqJbkSBFGl9ACUkMR1C3IkGGRU6QUYgmo3uRIIMaZUQVSQxHcXciQYZFTlBRSKaie5Eggxpl ciLKJqKbkSDDGmd2KsqgonuRIMMiJ6qogIruRIIUaXzFqaiAiu5FggyLnKgi+T5c6LciQYYzPokl LWsA351IkGGREyXUQEL3IkGGNc7mNNSB8U4kyLDGiPexg37uRYIMC4xOPwP0cycSpEjjG04/A/Rz LxJkWOSE65YYDuFuRoIMuzg1nrSsAZzfiAQZDOlAPhLDYdy9SJBhjTOgfGIExm9FggxnbBG5TDo3 I0GGBUaUjkRyGPftSJDijC2hcCSGI6hbkSDDIic4ahLDYdy9SJBhjTM74WQQzr1IkGGRE9VTrJm9 GQkyrHEWbGdjBcYbkSCDGiE+7AoCuhcJMiwwOgGBH/2tSJCCjMr5zxK/qbT68W4kCNHIiwrqqqBP RIIQbbzdqWj50p+IBCEaeVFJ+h3X/olIEKKNd6CaKIJjuDuRIEQvVo4oaVoldT8ShGAgRVVJDEeR dyJBhjXOgJ6bxG8W7lYkCNHIC/KSAI4ib0WCEGy0EdUlARxD3ogEIRhpoR2WSJAi/wsiQaWVFQnS aIi2mBYJoroC04xZ5z6PBM3vyLWnkaD45/z+GoWDYqQ4T5KPw9+O/8zPWf/d8Z+ayor/1HiK/og3 OdbOzFoPj0adZA7lTBINjrk0oDQC9O0TX/FUHwWSfdOkdehYEg4BHUjEkhCQIDEQsrbClO3JdPqy YY1TWJbVkNNFgdauoqK+VdkCGnkb8Er7JcinkaAuE5rbqmwBbbzCtKzueOH1bbjjLX6cGsBI24FW YkKC9FEht4GwLuIHXo4LHZiWNRwv9unXjszZdv1ztN3RDqAdTlxuxxiNkXSp0euKkQDaeAcqTMND C4l9e3GK1i5NOg8S0YtXmZZVjfcQJdK3YIUxZYESopG3Ci/EiQTpl5zq5voay/RPTTaNORAtqzta 7KqJaNd+P7pmH9HI24FXmjRBxmfd/SbBsrr6BoA2XmFa1nC8bg0zq0v30DnUfBI6OhCJlVBlPnik a4T15Vh7XyHaeNNSGUVpHKvrWbMG3DbbjnQAaUZKJ7CnK1FXOPJQjUkI6cAkVnECw9VBUtmskKSu IDCscRanroJVrg8kaRhyhSWXcwdo5G3AW526/LJUqWvWBiSHOkxCSgemZQ3H6/p1fPVroxDtKgIY aVFdDatcH1fSrp1u6JGXJwZo422vKIXuWDE4IDWN1gz5UINJdMnxaNrJyy1V1S6URhUtNAtoYHXy Gljful2zpZrRyGI+3lks6Vh1LVsoLz/dXve5Fn8oHyuv4eQ1UF4aaWKkjzXpHt8aYlwBVQAvWiVa 1nC0T9ZtrD0TdHMMRCMvyEv2v1akizutvd013HaovCTydGASKzpWfGQ6wKUBt9XnBbSxRuREefkY lK6Alm0EdC8GxAInikv2hFHkk88YaM2YDnVXxKtPKC6NRCkOxZVgX650qLckEnVgESs7YblYVDoM +KwV/4g23uyUlbEV9/Go9eUOXWe+epCARl6oDDQiJUgXk1rfLdF43KHikqjUgUms6liffXJX7vBa aY5oY0UNVGzCfXBqfa5GY3IrdghoYMUmXLaIUSR+s0a3l5SIXzxUXBHva3Paak5bLk61mi0JzOmq ZgQjLcqrYxPuw1Qa7dDA3wr4AdhoO7bislOMIfEe6MZ/EqQ7Vlvdias7cQ0nLhekijo2K728sEZw AG28w4lruGrWhamChuE1WNcOlcxw8hpYzcrOMYL0gaqwxlRZaMHTSqDqQCRWQH1hoCq4rdNmTE0H XxbWOAPqS3aPWTjUV5B6RiN2QQMtiEZeeMdkCxlFukBVEMHqSslwqL8kUnVgWlZ3vKCwoHta6crL Q/0loaoDk1gJ69v/ilDV7nzMrDe9pRqqknRyDZe3GLfOfx6u2nW2vJljRGnuU00BqVDffio/hyGh ql9k4tKfuHG1OiBBJ4Zop4y36K78xQ3Z1Fu2LSKjr+9p14MhMCD4O7YX/3EF+Nw249sjU8TwsN7v 97futrGWW/vnW6gatjsF775ykyUqmXSgUvdHl3SFuSyYZsw697tu74/L+lM3dsZZ+9MN3C/u7Sci pKl1iY+m2TLcjZBy74pP/PKSWtNA8QPTiJLK5MaJr3hqwfr4WYQUgUD4ipS60pGRdyKkhjVO2Zt7 WQ053QS0qwgpopG3Aa9O124Q5fxWhBTRxit79y2rO94bEVIEI20HWl352CQc6eajXURIEW28sjn3 sobjdXPSnkdIEYy0A2iHE9e9CCmijXegwtYqyMbRSGC9io8adnHq7tzLqsjp5qddRUcRjbxVeMcK airyVnQUwUxrRGZ1R4sTRK6io4hG3g68OmObkdHvu3kRHUW08cruJMsajvdGdBTBSDuANqHC/I7a l9FRRBtvWgobEtZUHHJexEYNaoQZ6Zy47kVGEQ2sTlzFict/JvsqNopo4y1OXQWrW7G+FR1FNPI2 4K1OXfeio4g23urUVZ26/M59F9FRBCMtqqthdes3076MjiLaeNsryqE71jvRUUQbK7643UnsXnQU 0cDqJDZcXet20r6MjyLaeMeqb8eKaxryToQU0cgLElvrKQl5L0KK4EWrW3IvqzvaZx+APEZIEY28 8NgkrqnIexFSRBtvQInF6FjvREgRbawRObGy9btoX0RIEQucWNVKbFORdyKkhjXOhOJa6y0F980I KSKRE4WVsa71e2dfRkgRbbwZ61qJaxryToQU0ciLyipOWfcipIg23uKUVbCu9ftoX8ZIEY28UNlK bFOR96KkiDbe6vTVnL7uREkNa5zN6athZeu31b6KkiIYaaGuleCmIm9FSRFstN0prDuF3YqSIhhp UWDD1bT3oqSINt7hqtrhBHYvSopo5AWBSXBTkPeipAhetEKkFqjrVozUsMYYkA+1dTNCimhgRXFJ XFOR9yKkiDbeiOqSmKgh70RIEY28UN9KhFSR/wUR0jQabbrFt9QCJzZ5j8fADxaj1rnPo6Opj4vo KEeW+JuKNwOhkebY5gofgazwEcgB+4lhWjB49vd90e+HZv+pWF2kCVq59buxOjnyG93dX2fweRv0 9coN7/QKUEll3vR9CGGOb+dttuadt2SbV8TJxAGFkLpPE8TO/K57/KMy/tw3E7fpA89O0Umkf+x3 N/w67xz/3ep+B+NXPmM55iqcGYg8jQZ8Y4u5zX+5MlB4hYPCPfGMYL4RYYWB6eIxnfRG1O+PRv+o jD/1BMbs/D//lujbT4NHUvCToje/KToy1ak/4Kuil29Kyhvse5hoPazuXyhW4lYkYkoRcPZ3Pa0f m/3nBmboFbz+cO7ptbn88GujQbl64rk54HB4c+SiUmIXSW6CfmmIrLpuhKUUAWf/PU/kB2X/yS/y zn7S3OLgeCfXt1/D516gRHN+5pcenj/jz3wq1rxk/y9VdlOBgSf1JvqaQRhLoWKkPL2+0CH5eNck uaovbMzT+bwZ4WG6mXp/cD4c9wkUYZ3pPUXnzH+J6P2Fkp1RnCXDqSBQ2vfHd9WwcLWO1/Lii1sl +sBCrmLrxby/2EV+4KXD7Vh36d3u3oe/p3an7Qm8Y2mnX/pNQW76Hsq3tevjyeDtkwpibxj7W5B6 mKtk1pBU0zq8O//mt9BVYSA2Ol/S9ZbwYpyjDfKuxjo7l/IixjrU2PvPD5rVQIlAj30mI30U+IWS M8bOp1DUh6kouWMpE4kGRYq3MYj2peRTOTkJ31/YSBzNEoOyp1O5SFbu75Kgv27ghdysFO9Yvg9X brgau8r3F7j+D3df1r3S+/dud/UD7jTcfXsq71juT4sxpln5xmCb/3YVIwkrz8rwlmhaoJeE19OV ocbHo9CezIk/hD3Hm+hiJMl3kg2RjhjzhvCpdJOYUsVDmemdbJsaT+7ky9M7SaeIjle5v080eN2O 92vvhyuplX5d1fsLXO+Huw94d9Zde8f7+eHvs919eyrvWO5Pi2aO+ZN2l2gaN6h51mFt9l/D1M3e wYq/z0ot/npLQ5mWoRURUJtJvnZKzouaH3HXxPS938WQL0qJMRt8Pi3NLhKTUfKd80gaIuZbMBfW 8Vnz30KtYg6PrD37mSlh85x9x2V8/645Q+sCF52Q6+VoAT6gSKuQq+jvL3BNH+5a8Q6sO/OO9+wD 76TeW73f76uMn69KRnnQ99N8RfL7W0irOrmjglLnhcj3yWqdESD5sFmtXY1aqSJOluSagY1I+Bcx GNSpZqZk0+S8LzOz9ZpFNeYtanyqJDO/lmQkxQe+70VOpaJB+b+rRnHX73ght1WKdyzfhy83XM26 yvcXuP4Pf1/gbq27+I7398Pfd3ga6ym9Y/k/raG6NW1FXc2yVyBxiqjRpMHdnxk3PZld4xRkHHNG UWkSnKNkGrMN3TuskiAvkJJZBoLZmAMVfFLmuVtVk++UgX5wKSZOxhj5JEpUrvn3ZGFPJ84sJ7jM J0Ll+666ZF3b4hLmyJeycv/AEq0yZr0EuJ4Pf51w9euuvNvd+sA7uO6p3Od3Ld/nq5FOwdN8rEh0 4miaEshTAbsDnGjFc0q7MOinWMh7FgDMvvv/4TUxlQplbmRzdHJlYW0KZW5kb2JqCgoxOCAwIG9i agoxNjI5MAplbmRvYmoKCjIwIDAgb2JqCjw8L0xlbmd0aCAyMSAwIFIvRmlsdGVyL0ZsYXRlRGVj b2RlPj4Kc3RyZWFtCnicnVVNi9wwDL3nV/i8MKkkf8QBY5hkZg69LQR6KL11t6W0he6lf7+S7EzM 7Ga7LYGMI1vye89PHujR/O5+GTAH4OFA8vZO3k8P5sOd+dmhkefpSwcyYX50vEhH32W5kd/1XaJf u8c7LSkP501LRxb6wTjv+2iWz+bdBY1zZnn8mBAzDglsDomOMGRMeAELDo5AGX0Cn8dEFgFj/rS8 785Ld39T3FlGa2PobaltDbpeqifAvHx7KWUQgnaIvV/hWFI4cIYIkcEggMvoEjESARUp5AMmRQgT Y+IFOj5lBEm7FPAnmmjSIaEk0IyXmgSjEJOi/KshrVzHPtfq5NFJJiLBHuVhFPzOC81KmfqglGmH MivIK6xtOFNUzg5k65ERRKF2uOIQdA5xiwwI+uE2KXyZcXyC1wk5upg0whMyb2O2ViIs6FFliSWP NIF3inWiqEilnO4wo9WJphz4GmshlFJHcjTvyVY1ANePNxrokTYcajUp7AJNVwDsDV2CVNg0eomT x40jjckNrV22hZMUCGto2zmWxWXdrtmxp5aCu1IoBo0wFiEyiedmGFjenWLeEjvoRT301LgSVWxB pZfPI3fiwb4CsajMZlkbcqt6LlLaM4NDBye4sJJD+ovRyY7/aHNiYnizOYbWZ2oxoVO6Vs98YlQN QLkLZAlfPjN7ej10ici09kcbXc0Z2fz1bOdNt+bUg8BXKbVvVhTXXlJ59/rpLdrj+LzLn7VgYe7F tHttDYFP22cKNc/Wtni9Of/L7tfhzT2061z5Q2l4vqkNDnpf75QMfDWOL0gnKFav3Zs/7DuJywpl bmRzdHJlYW0KZW5kb2JqCgoyMSAwIG9iago2MTIKZW5kb2JqCgoyMyAwIG9iago8PC9MZW5ndGgg MjQgMCBSL0ZpbHRlci9GbGF0ZURlY29kZS9MZW5ndGgxIDExMzQwPj4Kc3RyZWFtCnic7Vp7eFTV tV97nzOPhDyGACGQkDnDkBGYhEAAgQSTyQvUyCsEzFCQDJmBDCaZODMBsRZHW3xEFPUqVdqK2oeo VU8SpBPUQqW1LcrjWmqrbREfrY+vCPb71HrVzP3tPSchsejXfvfP65ysvX97rbXXWnvtdV4ziYY7 A5RGMVLI09zm6xjNGCeiF4lYVvOmqJb2Rv4pYJC5YX3HhrZ77y+9nMg6i8j0/obWLesr7j35KFHG JMx5oiXg8/8smpWO8QmML2wBo77/BgtRZhrGk1raoldvZB7gzOkYp7WGmn2tVAaYOR+Ntc13dcdq 9RUV4xqMtXZfW+DGUQ9XYOwnGmntCEWihEAhahVNRzjQcc2qk3WA24lG/Bw8hkN84IOZxZjT/++P aT+NA403PUzjVBflECXeBr0j+v5g4h0hFz1/D8pxg4j20OMsSI/TAXqOncWsJ6mP9tJvaCzV0Pfp WrqbbiIzrQLnFqrHYQL/bjYusZeK6UFs0YN0BLqX01baT9ksJ/EuXUfblN9h1jZKp4lUSUspRLex yxKdtJpeU79Nc+gyaqcOFks0Jm5P3JX4Ef2Y+pTfJD6nETSemnEcSbxv+mPiz1SEGffQffQauyvl KfLASwyaP6Aw7VLWqCyxIfE/iMBBmxGDSovoCDvI3bAeoLdZDrtWqYaVHyb0xC+hlUdrqIV20X42 my3kDtPqxKLEEcqGj6th9T7qoX044vQsvcrSTGcTP0qcpXFUSJdgPXvpKDuo9H9+fX+FSDSyNIXm QRKin9Ov6Thzsl/wkCnNVGLymK5JnKDRNINWINqHMfNv7GO+Fcd1yvPqgkQVZSAvd4ps06/odTae FbMlbCWfwkP8fiVMVnicgcNPQeT7Xlg/ydxsH0/jx5Qfqo+pn5on9J9KZGBHXPQ9+gH9gqVjpRqL sBvYy+xNXs3X8u/xN5S71UfUlyw+rPoKaqPb6DH6mGWxuWwZ+wZrYdeym9id7D52hB1n7/BK3sCv 5GeUFuUq5Vm1CsdyNaJ+23Sj6VbzO/2N/b/s/+/+jxMliRtpGerhekR/D92PlfXRMXoFx2v0BjOx ESwDh8YcbAX7Jo6t7Db2ENvDHmF74eU4e4O9y/7BPmSfcsJh5rncwSficPIw38zv5t/nx3Ac53/n nyhjlYmKW5mtzFe8SghR3aTcgeMp5XV1vHpMTSDPJaadpt2mPabHTM+ZzprTLDdYyfriZz/8fOrn J/up/+b+nf09/XsTr9MY7OF4ZMFO8xG9D8dG7PdOVNyT9DuWhtyNZ1NZObsMmVnLNrKr2NXI5HfY LvZjGfsT7Blk6Q/sDGJO53ky5ml8Nq/iS3BcwQP8Kn4Hv4vv5S/z/1EsygglUxmjTFUWKmuUgBJV tig7FV15UfmL8obykfIZjoSaqtrViapLdasL1bVqp3q/+rb6tmm16QXTX82p5jbzjea4+QPLhZZy y1LLMssayw7LPssJaxOq8xA9RT8bes6zU8r1Sq3yFN3OZ6rj+FF+FPW8lvzKIo5K5XvYzfxbbC+f ZLraXMbL2GI6q7qQ6+f5bv4RL1MWsTq2nDbyGUlr5tEqrvA0Xz1Ep9VnsLajsHy1OY1t5WfMadTD iM+Dz18p01W38gK9qrzGLOqD9Cc1lY1lp/nDylJUwbNquamRHMr36QnlKvYteorXEqV+at2OOl7M HsV1oYGVsH8qCVL4YlTRHOVN+jZdyf9Ip3Ee30zfZX51A91OM9m19Db9BGfFFFO7eap5DPstD6pd fBTbS1x9BKubxyYxxTSavsPWKLvMZ/gr1EnH1FQ6qfwU0R/jTyiL1LOmetaCM+BbdCNdlbietpga 1ZfYBlLYSipQT+Hqdq1SojrQX4erympc0/bh7N6P60ClsgicHFTOZaiLFbhC7MJxL64TKiooiHP8 clzFjtJecwOP0wZTBsNVh0h9ob+eViV+QvclNlB74i4qwvXgpsS1sLiH/ko7aA/b1v9N6qB8nDkn 2WWmBfyYaUGiiHfxV/hyvnP4/iLbBSyH3sPxBAblpqepS/0DLaeKxPbE71Hdk3GFvY/W0aX0Flb5 PjxcrBykmf2LeXdigdKB9b5GyxIPJ+wslVoSrbSEnqEfW0zks7ixxzp7Cev9JgV4fSKqBPqDyMMO ZMGDbHXi+nOLp3pFQ6Wnovyi+WWl8+bOmT1rZsmM6cXTigrdU6dMvsBVMMk50aHZ8yfk5Y4flzM2 e8zoUVkjbZkZ6WkjUlOsFrNJVTijwlrngiZNdzXpqst58cVFYuz0geEbwmjSNbAWDNfRtSappg3X 9EBz/Rc0PUlNz6Ams2nzaX5RoVbr1PQjNU4tzlYtawS+rcbp1fTTEi+S+A6J04EdDkzQanNaajSd NWm1+oJNLV21TTUw1z0itdpZHUgtKqTu1BGAI4D0sc6Obja2nEnAx9aWdnOypiMofbyzplYf56wR EehKQa3Pry9d1lhbk+tweIsKdVbd7Fynk7NKz3RLFaqWbnRztW6RbrSgWA3dqnUXHuzaHrfRuiZ3 mt/p961u1BWfV/gY6YbfGn3sNW/lnBvCeFZ1401DpblKV21OUBPDrq6bNP2BZY1DpQ7Rer2wgbm8 YEFT1wK43o4k1i3X4I1v8zbqbBtcamIlYlXJ9QWctYLTtFHTU5xVzpaujU3YmvFdOtVvcfSMH+/p S5yi8bVaV0Oj06FX5Dq9vpq87tHUVb+ld5xHGzdcUlTYbRuZTGx3RqYB0tKHgsCgTCKpLlBd/WBm mYjIeQkKQteaNUTS6MSa5oomMJe6mudCDR8vwyzdjx0J6inVTV22UsEX83VTgc2pdX1IqADn6b8P 5/gMjrnA9iEJKOpksNQgH8C6261PnSpKxFKNPUWM5XI8u6hwU5w7nR02DR3SR0uRW5+3tBjpdzjE Bt8a99A6DPTYssbkWKN1uT3kKXZ7dd4kJAcHJGNWCElsQDI4vcmJSt4rH5HH6FbX4F+mLXtUbUup zrK/QhxIyuuWO+uWrWrUaruajNzWNQwbJeVzB2UG0kdVNyq53EA8V5FSFOXqQWUxaEzT1QL8mWVR ++MWK6pScpi2QLc1XZxsvakOx785KZ44K2bJ7tw0I0y91D18XDZsPCy8tC4FAeNWWdewqqsrdZgM pZZ0eInRoeKpodGhVeu0AmdmAf7iiYNzBXlzdQ9SVi0UUH9JljEcpphrYC8+ojqLChfgQtfVtcCp Lehq6vLFE7F1Ts3m7Orjz/HnujpqmwYKJ57Yf2uuvmC7F7lqYaU4KThVdTvZzcu6Pezm5asa+2x4 abq5obGHM17dVOUVieTVDY1Dd0+eEt4iVANn8gHXJF6+8DbnGOkYWYCG4ab3maYc/Mxjok9JUw+K 9682dpy34DlgBNn7cENd7slIMb+o0XQ833WmXf5wjtv20ZrTVHx6xvRRsy6cWYK7g9k50dV2T0vw nnuCLffwo8G77w4Cw1biM3ZYDfFvwGu+J5PNJj7epMHFOLX3mhz3Yttba2x/o+JFMKXMdoxR1Qg7 fOed8vYowuU7dj0UuXNt5vwPrblWedd86M0Lpg7cQROfGSsiSjHeGeU8S3n/Yqo+d6P9wuuUYgaL H6E2NSIs4F5bhju+0FLwzpl82eRkwxPDStz1W9QJcCG4pfznlHx1JTxrkoEZpcqRImdlUNTACp7P bzCwCp1TBjbh+eA9A5spg1kNbMH7S7aBrTSdHTdwCnXxAV/p/FHeNbie2aZZBmZkMgUNzMliajew QsWmTgOr0PmRgU2UaXrEwGboxw1soTWmAwa2Uo45z8ApVGse8JXOVpi3wDJTFfhKsxyVWGTIZnlV YrPkvyexRfI/ktgqsNUscYrIoTXbwMihNWZg5NDaZWDk0HqvgZFD60cGRg5TyMDIYUq+gZHDlEID I4cpfzMwcpg64As5TH1A4lQRZ/qlEo8QsaWvlDhN8tdLnCFxWGKbiC19q8SjgLPSt0s8WurcL/EY aedRibMlf7/E4+Tc5yXOlTonJJ4gdd6S2C7xWYknSf3PJJ4qcEaqxEUS5whslfFnuCSWvjJKBE5L 8ssllvWccSk9QhqV4ASeThcCNeB9OYB+Ed5z20FR2oLnVsGpxigMLFof+EGpMQ2SSmrFoVE9eBsw P0oROQqgD0B7E1o/NCuBg5jbKmUb8LTZilH4X3yVDtHUvqBbijNP2IwY/jWaDcvT8fas4bl4EbjN kIYgD9F6WJwyxNaXzZyG9V89RG8RsjA0gqBcjw8UlWv3w06bjOZK8ISf/zxvwmq7tJictwKjIEYi Uxqe9aNSN2B4bge3WFrQpO0WuRINaw0hM+0yrqDUnvYfR/Kveg2DqEZqbpaxbsB4Cda6XuZYSIvk 7oTwJpJcy2IpaQFH7FWECsFbKj2FpSQoc7gcbadcUXI3NOzAPJqLPfDK1Wgyt1vQd8r6SeYouQfr ZaxRyQuh9Ut+h/S3ZTBTGjhhGVPUyFG7zGVy7JOWOqT3NpnzgayvkzYGdqTVWGf7YBTJGQNxhIfo dsia8yPiZukjmY/NMm6RkfOvITkWus3w1ikz4pdn1BczIWa0SjQZ+lPQiwpcZ8R9ftvt/4e1n7Pu H9z7sKyvgb0cqJ7zrWBobQ+Pq2zIHomVJNcSlf4G6lLYT67VD85mufKQPOu+qhJ8w3Y9YJwpXzxf RFaj0OuUM0W0mwarOWlHaLZC46tqaNojWsn06RdqDS0BbVGoPRTd0hHQqkPhjlDYFw2G2qdpla2t Wn1wQ0s0otUHIoHwpoB/WmU46GutD2zobPWFB2aVSqZmcEtXBsIRzNdmT5s+Q5u8KNgcDkVC66NT pNZQ4bSSqyVvUUPSQDCCF+lo2OcPtPnCV2qh9V8amxZs16KQrWgPRgN+bXnUFw1gcru/OBTWQpCE teZQZ3s0HAxEpn2ZkUFeg2hqwr7NwfYN2pL164PNAa1Iqw+tg5fFweaWUKsvUqgt9cFcc9CnLfd1 tvuxDG3GvLkl3lCn1ubbonVGAogIK1gfao9q0ZDmD0Y6WiFAUFpHOAhmMyQB9L6I1hEItwWjIvR1 W+RCWuGzXZiAQNgIS25HOOTvbI6K1W5uQSBDPKAPtje3dvqxJ9pAEKH21i3a5OAULdC2DraHaLd/ pXep7herDwciYpUiPeccJLNt2CqTK5ochJdooE3kMhyEV39oc3tryOcfngRfcunYjsF9CXVGOzqj mj+wSaQZOi2B1o7hGZqGa3BInts+edbgrGbpqNqNqNt35VV+QJa8w4gzUZxxfmWX0q08qxwA9Sn7 lZ9+/VTw9VPB108FXz8VfP1U8O8/FQy79p7DPlkz55O9PkyvFXaGXpXldflLbLZCZ8vQsZqvzlDr 1IXqRWjnDfPQDrtfZmUx2k0yi8lzvYXp7EGF5N6Ka13YuI74vsLCOazQV376qEGZ3OvKsR9/RplC p0BcmdLjnmDvUy5QJvSU2T1xxdmbNaYks7JIEV8EFctWQxsCPQk6AFJprZIPvg3tdaAY6EnQAdBx kBmv//lSqoFCoN2gU0KiTFDyejS7rfICZRzmijfvTGUsnQElQArZ0RaDloDWgnaAdoPMUk9wQqDr QAdAZ6XEo4ztuWsmYh/bc6vseje2lsihLzlcvUYOey/3JvtFy5J9zSVJtdKk2oxZSfa0qmR/QWGy zyooiYk+Nb3kYGW2ko1Filf6DrSM/5IyGSM7PaCMIR3EFbPB8ShZvZNcJbsPKCoxhSsMm2xPHFRY T/rIkspUnuBnKIvs/H1+Oinhp3szRpbsrryUv0FPgg6AFP4Gjtf563QdPyVyjrYCtBt0AHQMdAZk 5qdwvIbjJD9JmfwvVAyqAK0F7QYdAJ0BWfhf0Nr4n8V3RrIVuALE+Z/R2vifsKw/oc3krwK9yl9F aL/rmTOvpE8Cd7EB7AUGGJtrgKzskjh/qeeTKagoF3YaFfW0MpHKaaYysadghj2u5PTMD9rj/M1e zW1/oHI6P0E6iCOSE/B8gjTQUlATqANkBnoZ6GWKge4APQDSQagytDaQxg+DXgS9TNNBHtBSkJUf 74GbOD/W46qyV2bzo/zXNBYZP8J/I/sX+fOyf4H/Sva/RZ+P/jB/viffTpUjICfMsaG3oS+G3MR/ 0Tspy56oHMnFt3F2tMWgCtAS0FrQDpCZH+ATe/z2LBh5mg5bCZo99K7sf0IPWcmz0e5xVaMANdG4 Si8CQrNb2+3iHtfO+zAUjev2u4BE4/rOdiDRuK65Hkg0rtZNQKJx+TcCica1ai2QaFxLGoDQxPn9 P5t0gX3OkiuZVpnJNyNLm5GlzcjSZlL5ZnHQJ6qI7Xs9U6ciY7s87ilT7bH9LPYMi9Wz2EMsFmCx rSx2PYvNZ7ErWMzNYnksls9iHhZ7ms1FKmLMs3fYcJ4nh8UOs9jjLBZhMReLFbDYJBbT2BxPnDt6 Lpkpu1rZ9VaKkw79ReW4+mRyBzLqQM07cE04gPYYKCFHHihpE5PK4/JFP7F3akVyPK20JFR5MT+E iYewDYfoNZCKDTqEMjoEI4dgIBNtBWgt6CDoDCgBMkN7IgLfIdtMtMWgCtBa0HWgMyCzDOcMiFPI CPFJGVixEfQSMeKHcIh/tHBwh2eCLc/mtl2s7MhjmflsSX4in8+hbPGFctZI68g4S9/3cfo/P06n lMoUfjvfQROwEXcY/Y6eTybY4+zeHtfT9sox7LuUr6Lq2DxysQL0cykix7Mpzyr6WZTHH0Nf0pO3 EtMye1yF9v0sQ8zaZ/8k7y37u3lxDvhO3tP2P2hxlfXYfw/OY/vsJ/Jusf+2OG4F5xlXnKHbr0nV vry59scPS9XrIdjVY98qun32b+UttF+ZJwWBpOCKCEaeTHu9a5X9YtiryVtn90Rgc5+9Iu8K+/yk 1mwxZ599OkJwJ+FUBDslTzp15kuDK+bEWYun0LLT0mhZYrnQUmIptDgsdssES65ltDXLarNmWNOs qVar1WxVrdxK1tHxxCmPW3zfPtpsE51ZFa0qsY2Llid/cODMyulS0kcpdbxueRWr0w82U906Tf9o uTPOUpet0k3OKqZn1VFdQ5U+110XtyTq9TnuOt2y9BuN3Yzd7gVX5zfHGTU0xllCsLblit9/+4ix kdtuyxX95G23eb2Uk72pIqciq3zkvAU152majNZ97pMzDE/Qd9Ytb9QfneDVSwRITPDW6f8lfiDu Y/9gZ2tr+tgHovM29inl7B+19YKvlNd4vXVxtlLqkcY+gB4q5gOpZ8WNWeiRZs1P6u1K6hVgPvQm iQ56KSlUIPUKUlKknsqEXndkUm1N96RJUmesRhGpExmrDdU5XACdggKpkx2jw1LncHZM6OjlUiUv Dyr5eVKFjac8qZLHxkuVledUig2VWwZVbpGeFHZOJy+pk35qQCf9FHTc/+4nUOV2s94yb/Nq8eN6 k7M2AGrSb93UkqPH1mlad7PX+NXd1bSuuUX0voDudQZq9GZnjdZdtvo84tVCXOas6abVtQ2N3as9 gZqeMk9ZrdNX4+1duHTWnGG+bhn0NWvpeYwtFcZmCV8L55xHPEeIFwpfc4SvOcLXQs9C6YtkjS9t 7LZSlbd6dbLv5SNSUa9NuQ5vVbato1wWb5kjZ2vufjyt7KERbq+e5qzS00FCVFRZVClEOKeEKEP8 B4Uhytla5sjdz/YYIhvYI51V5I52RjoppzZYk/yL4ANWtFMkPNm6I1/2gaxW9/hqxP+j1ulTl9fp FctWNXZbLOA2iSXppQO8ESNq44mDSeY0MEsFU1EGFQVvvuClpBiK/7r/nUZfLc6CGH+6l3nyWZQi XkXPr2vguBQ0GD9V78ezlLg9RLxYYIS5WWTAhhG2203JMYk1D1C000BGLqJGn5yJKZGBlAx+RLLc gxmLwiD9L9nWOZIKZW5kc3RyZWFtCmVuZG9iagoKMjQgMCBvYmoKNTg5MAplbmRvYmoKCjI1IDAg b2JqCjw8L1R5cGUvRm9udERlc2NyaXB0b3IvRm9udE5hbWUvREFBQUFBK0FyaWFsTVQKL0ZsYWdz IDQKL0ZvbnRCQm94Wy02NjQgLTMyNCAyMDAwIDEwMDZdL0l0YWxpY0FuZ2xlIDAKL0FzY2VudCA5 MDUKL0Rlc2NlbnQgLTIxMQovQ2FwSGVpZ2h0IDEwMDUKL1N0ZW1WIDgwCi9Gb250RmlsZTIgMjMg MCBSPj4KZW5kb2JqCgoyNiAwIG9iago8PC9MZW5ndGggMjI4L0ZpbHRlci9GbGF0ZURlY29kZT4+ CnN0cmVhbQp4nF2QwWrDMAyG734KHdtDceJdg2F0FHLYOpruARxbyQyLbRTnkLev4rYb7GCjH/2f +CV5bN/a4LP8pGg7zDD44AjnuJBF6HH0QdQKnLf5ocpvJ5OEZLZb54xTG4bYNEJeuDdnWmH36mKP eyHP5JB8GGH3dexYd0tKPzhhyFAJrcHhwHPeTfowE8pCHVrHbZ/XAyN/huuaEFTR9T2KjQ7nZCyS CSOKpqo0NKeTFhjcv566E/1gvw2xs2anqpTSXKtS1y+Fezi2CduKz2RgFyJOVe5Q4mxBfMDfU6WY Nqq8Gx6Fb8AKZW5kc3RyZWFtCmVuZG9iagoKMjcgMCBvYmoKPDwvVHlwZS9Gb250L1N1YnR5cGUv VHJ1ZVR5cGUvQmFzZUZvbnQvREFBQUFBK0FyaWFsTVQKL0ZpcnN0Q2hhciAwCi9MYXN0Q2hhciAy Ci9XaWR0aHNbNzUwIDM1MCA1NTYgXQovRm9udERlc2NyaXB0b3IgMjUgMCBSCi9Ub1VuaWNvZGUg MjYgMCBSCj4+CmVuZG9iagoKMjggMCBvYmoKPDwvTGVuZ3RoIDI5IDAgUi9GaWx0ZXIvRmxhdGVE ZWNvZGUvTGVuZ3RoMSAzODg0OD4+CnN0cmVhbQp4nOy9B3hUx7k/PHPO9nq2N0l7VqvdlbSSVr2C tKDeAEkIJECAkAQIUEGFbpu4G9ux45bYcVxScI0tBBhhJzHJJT04ToLtNDv43iR2nJDYqbaxpO+d M2ckUex77/P8v+f7/s+TFbO/35QzZ+add2bemTNnGRsZ70MGdBDxKN4z0D18/DPfO4oQ+hFC2Nqz a0xUn7hHC/wcQkrd5uEtA1+N/2MdQqo/IaT1btmxd3Nrd89XEBI0CNW/urWvu/ejb//Nj1AnpEeF WyEgMv2wGqE1bvCnbB0Y23NBPMiBfxHkmbtjqKe7+/s//xJCa4ch/sGB7j3DF1LeDiK0zgB+cbB7 oG+4au1h8Gcg5OobHhodQ1BQhEbvIfHDI33DT75j+w34j0H2egjD8Ec+cD1WET/HK5QqtUar0xuM JrNgsdrsDqfL60tITPKLgeRgSig1LT2akZkVy87JzcsvKCwqLiktW7S4vCK+ZGllVXVNbV19YxP6 v/ijfgihmbsvClqBtqFRaPOD6EZ0O7obvYh+hTah64Ddjx5Bh9ETaAJ9E30fvfZ/shwze5UDyMCf QCpkQ2j2w9nzM4fBTSlNC0LuBp9NIc6HzAqzf74k7M8zd88KM1MqK9JJ1xq5n0Lo3/D07IdcBfHP FhI/dxNws3TFe+qHZp6deewSGbSgNWgtWoe60EbUDfXvRVtRP0hmO9qBBtCg5BuEuC3wvRl8GyBV D6QifD7VEBoGN4LG0DjaBX/DwEdlH4nbKfnH0W7424P2on1oPzqArpK/d0shByBmn+TfA+5qdA20 zKfQtRJjSEOuQ9ejG6DVbkI3o1s+0XfLHDuEbkW3QTt/Gt3xsfz2i3x3wt9n0F2gD/ege9F96HOg F59HD14S+lkp/AH0EHoYdIbE3QshD0uMxH4NfQcdR8+gZ9Fzkix7QGpUIkwumyUZDoMMDkANr1tQ Yiq/3XPSuhrqTup2SK7pHgi/dsEVu2Q5kpTXQUqaC20HkstVl0jiTqgD5fM1or57pfrPhy6UyieF Mnk8uEAyn5d8hF0a+nH8PvQF6IGPwjeRKmFfBE7ZwxJfGP7QXNpHJP+X0JfRV6AtHpMYQxpyGPhj 6HHo20+ip9DT8DfPFzKKz6CvSi03gY6gSXQUHYOWfA6dQFNS+CfFXSn8qBw+ORdyEj2PXgAN+QY6 BSPNt+CPhXwdwl6UQ09LYdT/LfQf4CepqO876LswQv0A/RDmqh+jb4PvJen7e+B7Gf0U/Qy9ho3A foL+AN/T4FC8tnfD+q51a9d0drSvbGttWbF8WXNTY0N9XW1NdVXl0iXxivLFi8pKS4qLCgtiWZkZ qeFQSjDZ77ZbBLNRr9Nq1CqlgucwyqgO1mwUJ8IbJxThYF1dJvEHuyGge0HAxgkRgmouTjMhbpSS iRenjEPKzZekjNOU8bmUWBAXoUWZGWJ1UJw4UxUUp/Calg7gt1cFO8WJ8xJvlrgiLHmM4AkE4Aqx 2r21SpzAG8XqiZpdWw9Vb6yC/I7odZXByj5dZgY6otMD1QObSA0OH8Gp5VgiXGp16REOaYzkthN8 qLq7d2JFS0d1lS8Q6JTCUKWU14SqckIt5SX2kzKjW8UjGacO3TYloE0bo4beYG/3uo4JvhsuOsRX Hzp004QlOpEWrJpI2/dbN1S5byIjWFU9EQ1CZo2tczfAE8qQEBQP/QNB4YPn/3RxSLccogoJ/0CE kirOiQniGUdQNigh1C8QIGW5dSqONoFn4mBLB/WLaJNvEsVj0c4JbiOJOcViHO0k5iCLmbt8YzBA mqp6o/xv11b3xMFNYmYGSF/6F4J/EC9O8OGNm3q2EuzuOxSsqqJyW9kxEa8CEu+W61p9JDsG6bs3 QiX6iRhaOiZiweEJe3ApTQABImmD/rYO6RL5sgl75QTa2CNfNRGrriLlEqsPbayiBSR5BVs6TqK8 2XNH8kXf0TyUjzpJOSacldAo4epDHb2bJ/wbfb2gn5vFDl9gIt4J4usMdvR1klYKChNp5+B2AemO 0lVQt0tSs8Sk5uqQRuzgfHwnaS0IEGvgK7h0EUQI0FySl7To0kViB/YhlgzuIqcg7KJ8wMOHKutI FE8urazzBToD9PMJRfLJZVKGJjQL8hIgYK5M9D4fWzSamhQoTazuq1pQwIsyVcoFlHO7cjk5Igv5 xnCFhjRnHYviQ9BzIYyDbKQg0opucQKtEDuCfcHOIOhQfEUHqRuRtdS+jW3BxpY1HVJry1qy8iIf jS+mvgkUgGjm4SpBB2uiPtaskr9W8s956y6JrmfR4iFNsLHtEMk8KGeIROhBUGlVuL771mJrPnTN GhjdgjXdQVEQaw51T80e3HToSDx+aLh649ZSkkewvvdQsK1jkU8qa2vHVb595FZW1IgbVy7NzICx Z+mRIL655Ugc39y2puOkAHb/zSs7JjnMVW5c2nkkBeI6ToowuEuhHAklgcQjEg/JqRU8Gim972Qc oYNSrEIKkPw9UxhJYRoWhlHPFEfDBBbGQZiChsWlMPKBRnJvBRHDcFst9pLmOdC59dDGTtK5kBOa Ev7hCRwsRxNcsPwI5lSGCV2wb+mEPriUhFeQ8AoariLhalAM7MQgHDImHdoYhHEKFKoD+TBVRZ5k KU7Nzq7sCJzxne8MgKqtA7emY0IbhbFfGWqAdLXEbYTg2omDPd2kHKi9g1yrDtX3dILasgwhSf2E FnLQyjlAihrpGqKOcFEPtA00oHT9QfBMHOyc6IySm3b0d0rqLEygumApNDvNUxkmN4p1HrIGc6W+ CV1BF7qJgBbKhto6aIgPvHCzTioktQFK3hOEqJ6NIkhbgXraQNXpWKrz0ZA+GBIV4T7J6XxyJCLV 4kN6o25CmwUZwj/C9VmkSypD6s5OWnjJd5OcAO4tTOihROEFopQvAOlAVD0pC/y7CYpKkn6TZNMy hVqDe2BkIYWWclJD9IQxVN8Ngz+9Xg8hwWJ2sYaMEXo5j9M0VE1qbgC586GVU7OPBfcGFnwyM4Jk ciCKiXwnQbFR56FLAybWRjMzNJeGGqXgQ4c0xitfQOWlMc4hCRSrYdaAhEpYoY3yP4UVFY/UqAQ1 o2Vo7deQEbciJyrFx487qqo0mepv4EroCCJeiTQI48q4WcEZT3i9FcETBarbeUv9FM48VqG+neNQ xfQb0y/Fpt84by2Jncex1998403hvZcsJbG8N8++mZONLQGL5OwmTq22q4LJWVxBJFyYl5dbzhXk h4PJJk4Kyy8sKufzcpM43s5Cyjnix/xPP1rDL59WcVcHK1blKZO8ZrtRpeQS3NbMRSGhbW1oUVai mlereKVGnVq0NLlxR3XyL9WWRIcz0arRWBOdjkSLevpXStOHf1WaLlQqdly4h1eVratI4T+n03AK lWoqye1JLwvUrzLbBIXeJlicGrXVYkitWjd9oyOB5JHgcNC8ppvJ+DB7XvG2MgDr1Ag6EPfa3QJu tgtm+LIb4ctqgC+3Hr6muLy4VkTZ0r5L0tTs20chJeC7R80S/umoUcI/HzVI+PZRuCrpBS4P6ZAH p02a24JTOHpEuQpVnK8A4b4ZJZ+zXRRBuJeISp2fxRGBOuxJHJGv4u2Gu9+4565Xbq1quOeNe+44 e3v18cjazw0Pf25DWnjNZ0d2PrA+lbvvCx8d2bD68D8fuf/DZzes+srfnhj8+q3LVt72wpaRU7c2 r7zjazBGJs9+yO/nv4Yy0H+dRMHZ9+MuvQE3+TXkO+THSZQkYSepGdTAIaNdrpFNRquMFoiPFwEp suCwBUcEnKrEyakQsDgZpyTjAKEVAZwSwKIUKuIUEUfMeFcAB6ZmX45rLY66gGjDzeB7O64FoQZE M/URSQZI/ga4MJBaH9B76/VNqKLCWlKBY5LcULQrJ7urqytK/+EuwqUP+KNRSVOxiSeCjeB54bps riKbrKL7McdzM2cURm9qUlKqx6SYeUmhxBqb35UYtGkVMwr+AqezBXyuJIuaf1ih1RnUHz2hN2l4 hcak41cbrFpeY1Bz8KWd9hoM3O+1Bg3PafQItKtr9jx/P/8DFAXr7Zl4eqywonCokLeJoCo2Ukub LZAhQC0ziIplCEbyZRZwU8YU/uB4VfTLUS4K9T8OKaP5iqnZc0THFLKOSX69hG8fIxcppjhdPBDI +O5BxZ0K7pQCv6zACkVC7NfhBvc7G03DJs6kfSehGXr52S5QQejjXTtHmCbmvh6lqgjBRB27MPTv gKx7RE4Ou0m1UD8dkUKpw6v5+yOe6cmkmuGWeG99zKDWq3iOV+sLV+2MDz02Urpo5yM92+7dmHmY 37t78bryZJjvI4HGPauyHF6H2uSxGm1mg97jtpXvm9o3dvJT1VWjn++wXXtPVlNfEdkbvH/2Q+4x kGAe6jk2XIDD5qnZf5Lam2X9A3xXqj0hRDzmKfx+3IrioFEoboEvImzk1U3hUFwbbQibHWK9gypR CemHp6HOUsVLpErTTihVzMJ6pIVV2SGpjIp7jFNpNRpXYorDk11QGtRYRY9btKlV1gSXM1FQh5aU liQaAymJBgWP+U3OJItWq9XYs5qKpic0eo1CAV/89Rq9lue1es11hVURM6/R6bQmH9GZdaAzFVKN 42giLpqX+pfGlvJ6rSvfAPXOJ0qST9QlX9KU/Cn8r7gJRSJmhA2IDFyolEgCkpYSERll1FOURFU6 xWnidovr2yhfyOfKTuVjlI/z87OWpE9hX9z8cjJOTlYkvpPVsPjXhmYFilXI2nLeIunM+i6mNKej 67tKYgJVIJDf+q6QXQVaEg4XFKjmtSWvQB7L5BCFJEw1Hd2cebmFRXyFkODz+k1ln2mpHW3JLB97 vP+AM2dZyeLu+hwDdCyF2rd01eb87ptXhr98e1XvUn/niiVDi90Gg0plMKypqAnVbF7SNNwQqslf UeBLDCZqBI/Zk+gNJtoy2q9eedqVWZFW07a0iuoTLlfuRA4UP1HhWu4acvFI1hwkqxaS5YWYaiFQ qWM6oUZJ1EauOihL6DL9wOVMFdQ20e0RrRoPa29lscZAmEFDtHoNtLHIfx8VoFuOJMB9TpHJBMkd nNz/GLl/RJ5lIvIsE5F7fkQuIOA75ILIFKePG2MmbPK85Y/rjHX+lCnMHbM18H/MgbyPaY11OTCi qI5opc4fPS994Zg895AekJuTHbq4g6scdIIPJqscC4YBXuSUas+ixo5Y9319BUt23t8ZbakqcGtV nNVojixqL919TSDetahkVUXUoNap+S9aPBajJ5Roje8/On7Di/vKBG+y22RzWyP+QGrgxDOrr+uI pkSDGlui3Nv5R6B1ctHeYxX5ON02P+vQprDJbWST28xGersrSa+DMD3pGnrSNfRS19CTOB2KkyZN SvcIIIETmQ0pNZ4mqSEriFLjWIz2fmFuDFjQrFLtrzAOWAoLJeQfgfaW2tmdVZ9dfqDq0uavvbN+ zf6mwJwWcObm9VUpHe3Tt16uF9O/a6xfvPmWbjIG3ABa2qKMgZYG0G0nKoLLg0NB3in3Z6csA8lv k/AcUROnrCZOWWjOF7idKAE5qKQc8lUOOdbBROoAMT2n88fhSv8ULj/mEeol+bx6Pirr+lmqJ1fW eRtRkzCYfHm5zsv135ZRVholzjM/8KlphdU4uzQ9rQQcafnA7F+4AcVXUSladywNWYKZcmUy5YbO lIufKfeJTLkamaSRDS5j5vlgXaLxvKsuZworjqibJSvrDBng82jpc8+cpkquoOWXCk0HJza602lO FQxayKhEQ7kBjSCmZblqeuOJV5utSo1RcxWr3FtkZLKa3yqqdaUk2DVKrVKxNjFZMGlVocbRZZxJ TLF5LepX1ZBKoTUAsXhtKeKMrmuDVqdVmtyk3vfASPAg2GO56K64H1ReHyFaHCH6HNGQPi6pckSQ ujv+4Dk6p/llqfhlqQC+LzUmIUQsfta6ftl+g6b9IK61ZdZH9EpPPYwPyqOmZqmdz0szITNIo2dp f7i4K7DRfL4HSENCYdFcAP+g2procCVaVM33NROdV9tFN2iCxhWryy7fX622+0ExrNo5PdjdvmzR lls2cclsHpz++/INlaGOdm6chRD5PI0QfwBWOBlo1xHvFQbEd+kASMUwPzBylrhWaxNtItIi7xTW xI0Hw/hUGL8cxuGwyjOFDZPGFhCo4YhqJapYaA3FQBoxeU6Th0UuGABlsYCKB6ghBAGWSyh/QKEz aqbvdqWlubjNoCVKoiozKjypMZLmB76MwxqjTlFr9Vk1/oji2ohfY/XZrT6LZmabVkiwWb2CeiZH Y5GsgODsh4qrlXaUjMLoCydRCkwJBmjO4JRMwqSB9dIoJxMdsZS9hIUE8m2Uvg3SdzwVh0h0Bggn JRgO/d2gN7iTE4M6I3YqDMggGLhngy8Gfxzkg4agwZrYam1XtpMxUhJGV5fFVQJzf4klTzifa8mD xUpUtrJRNBpyOql2RPgAT0ztcLiwCFPj2qUO8gHFuAYLIb8/BNb00PTvt/E6WzAhMWTGGjypMHoi SWK616TYj3+Dv7XY6TMpeOgwuGzm+1oQnNLkcyomibXNa8z626f3E9k8DbPEd0EnElAa2nMkRSVr hUrWCpWsFSpZK1SyVqiIVrgsiWSJl0gspUTBYMRNiSLEJU5xuZPIEprCuqNgU8BSTX/U0WIgMpAX w2cXTBNELyRTMbhgbpC0QyEt24IWMnzw343v/uqeu7W2gMcTsGvSvdiR3tw/0JR2vGx1V8bDn1+2 pSaFv7v7wcFFM1kwHCqVZGB8MjVZ7apYt3f18m35pukPUmt7SI2JvfAG2IRklfp4PKEiDadacRqs uIw4bMBhDQ6rcTqP0zh8hZXpuSuuTMliISmmw7oFS17x4iXv85wOodlTJ8yoeRi6IvQaPGluANlw R5TN8jK2Sx4uYnML2S72+e9WtPwbpaNfHRn6ymBhyejTo4BFz/jKty2v768K+Cq2La/bViXi3w2e vLFx6dXHRgAbAA/UX7upJH/Dtc0N13aX5K+/lowQoA3vQ0/JQ5vilhwyUGaT0scIC+hkcehkcehk 3dDJuqGTxaEjumFwRFoDOsHXKsxr/9zACE0O33REDIcj+ApWgaz0MIWoMXY6+ffV9mRfMMOpnknR 2AMed7Jdo7Enu4ky4B+oBFfA64Wp0midacMvWdQJZNBQCTrupum9aiNRB6Na8UumGNPf5CpgDlEo IcDodU3PTj/gtYFm3A+a8QpYTOloMbrhOJinukChXLdCua6Fch0L2ZxQSOZMnyNKWjpK2jxKBBYl ehAlehAluqFFDl1hQUChzIa54rlwg69GaCoBKjc96ReuEtYtovNrgLlmjzgulxAVEFMEtcUpzbP8 K3k9d3ZF62tqIjAiOuwJVhWbYlMb6+pSN926OvUZR/6quFger45UHags7yjy4LfGX7i+xhIuTRsE kSgUIKOFtlRacVBYdt3EePW1vYut6UtzZ+5vW72oZz/oS8vsee4l6Ev16KW4IdZY0bi88ZrGZxuV S2RxLZHFt0RWmSXEPLfJfkFGPUH867g/JTcl1+Ajc7aPzNk+QSBfevIF1/qex/8kXSiuIwszQxzC DZBdPAz5VRieNXCGrNeLdH+0rLBstAxb+CJLkcW56FdLfMq0BufbVNCgg+ct0igsnBekDjc3RUuD 87xtJgtewfob3ZXLUs0b9POLW+iDKu6lvPXXLsteXZ3t1ClUerU+WrGqOL0q1xeJr2hviUfSWve3 ptSVpjnUPM+rdSptcmF9LD2e5kiNt7a3xSPYVL0D1tMujz3Fb4O5yyf6rMHCUDg/1Z8cLV+1qKC7 PsNgdQgGs1OweAS10+O0BbMTIgWpYnL6opXySK7Qw0heiKrQZ+JJQpaliBg9RUR2RdLoXES0sojo Z9EUl3ciLQ7etAqL3BYWuY0scttZ5LazzO1OcZbJhCww/TXPDcdxPO5aDCP78UCLS+7hdF1bMtfJ F4hWnvqliS2Lv2yod7qSeCJRdRLvsjmdOD8cCYflkV+hV9lTkrwBu16x25FZvrJslE0CaS5sy1ni BfMwEly6rkTMz0y1j5k0M9NVKzwVeZ95vKpnqd8rabFWMOCc/NUVwelfzE0Oz0T8St5YvGqocsmW 5aV2U3TRspyZ/0pJ5G9o6nepVTNNgbIVoOGh2Q+5G5V70CLUO+kUyLJSWk76ZCH52Djgk6Xlk6Xl AxtxMjs9RLblrIIFDAnd+cJab/h8dp3YJNRJ8solg370dN57dOWYd3p+51LSq4VGNJPXpaa2grsR xjGV2pGU5gvli6bvg8GntJq/r7FRk/EaQSCd+Jpg3UBDcGmKQcMrzTaXSanVa915LaWbqB390R/n lhQOaml3rb9pVZrRbLD5EJ0XVFHQrUXo6biwsXy4nDNmZ7tiMV2W2+2VpwWvrD9eWTReWSJeWSJe Mi0kpeQYDDqihzqilzrS2XWks+uIWuqe5yxSF/cQGacUtujdLmPMnZOl8qe2+NuZIVUBo6UlD2YT pmFgRwlzzFKyOJaXRyyrBSus4NzG5cW6JxlWOI/MNpKwVVENGNeugE3DzeTxekei3ZFk13MztVhj p6uxDN9WMTvFrcW7lfhGvdcf9gyYfTaDd06xtly4hyzWFdDLFTsu3D8Xfjg9xeBN9X20mj8MK2i9 1pbomJ1FZlidvqncx4UROSOoQh4zYjsrsJJxINtJ8JMdk3q2YzLfl/5HmyXzm2MwRiD+BP6pch/k 60S+uFEXd1yj/I2SUx4UnKCSb3jPdOGY9wyoIZlcwBZVq5kRqsZJMMsmuxOCJk5zXGO0+OxOr55X 365sB4tSoTZatH/QGdS8SifoT7DdB5gZctE9cbOVDOI28iVvREgDjk1WmPn9h7n9CHmjgkytSXoy DdD9CKIydHtC2pmA+BPylgTZkIjrMhvSPSn1bE/CKu1JynsSzNoksypitscnbkw4/ruNCbZF2fTZ 5v9mY2K+DXa3L1+8+ZaNbDXWCLaWF9WdRA7ahxyySBxTH7fRwFmOIa251UGef9DlFo6dYebigvUV sxKk9VQjrJm006ddaXMm08tkQdBo99m0sHp6hunohUe1lgQEWinZyJJWCuQIJfLgFYjMMffP3MO/ Aq1KLKSDxEIKFF7BCqQj4rw5SLaVHNQ4kswkN4mmdtIVraN6YfnHWkdXMo4+bnP5isbRXetTq5bE UxY0ot3hs6rTmppbMjcdIsZRnmQc1USq9lWWdxZ58R92fe26WiE5PzhTzmwixR+gLWEZpdfuTS9P czRd/+x49ad6F9nSKnNmHmjrWNR7gLRwJ4jsNf4EiqKyk0iAJaOTzMJhaS6OQXy+tkrLaUMWWAQc 9dSZI9JioBFJE0MuWCdnuoidPNeoZMQvwgF5ApDWzNKkiQnlXlNpTJrpVx0+8igF3z5zjWAjrQ7z pwVMOgibGceHYf2sqrH5wEAOJJucTo/AbQuEYNWsVpmcFtHkdnmF6fvUgjTmr539s2KPUkQV6Ivx xIQEs5uM2m7SBd2kM7pJP3STPu0mA7sRvRjBYiQe2RjhI2Z5SjDL6myWH+mZ5YXT/AMHLvdYLB/n u2GteCw5uSRW/gLWISXS4bTJkjb7FM44EpMe8k2flTbM5zdaz3Z1nWbmReySeZPtp1wkKgvbrZLn TcUesm9mKF5/3ZrtT+6qqN73RN+i/QUzZy0WhRbWzJ/XO606a+m6Tb059/3pS6u6njh/Z8O1fdVe nWK9LdGmCWeFlx36xtCBU9dXJSbivckpIFSNRkiwzti84cRkt6Hr6XfveeDDiW5vMM2bDD1KehIh 9Sil3KO8Uo+CUUBxB8ysVuRHu+OJFQFsI3K2ETnbyGNTG+kyNtJZbC9wuXCN93843b4Ac6kWubFh 0tQC1kj4iHLlxY9L5xaZ+NJtmQULb8Udq77y7uGZP5O9GBx6/O0vtBzPH3ryxmePHHhypIR74PEL X2ml+y+rv/T2/f3Hr2/4yFJ+8JtEfz47+y88iM4hPXIdQSqw1J8j+wdaHsbmM2D2fJMMVwu2Owdj 5YuyiBuojWVVg5PtDtwBY6QDrZCeMDwrPWF4W97hl+r/sU8ciC2hgxW3g5gZrZLlgNlkcOUnDh1s Zal1BFxkmJyf1OfWjvLsrHwZRsEV6J24j85tpH3CAnneG3GT7+FWXHP5fHbZvPfO3L57UpKT7Cwk 5VJrSLKLJJNIoGMp/uDECrJduaL88g06mu1lG3kv4Pdh+hWwarKxIYXMj8YlDeU1mcX1mU1z2/bS g7t5nSiRLSpLSSx60Xz5yTPmfzOFuuS1vPJlOpPaNPaMqqyS0WpipIKtpXZmVGaVjF0+sd5RX9xZ lS1ktjTWpqzeVe+fn2KDJZdMsZeHXDTpemNLUnOq0m0w+TZBbwzM/oX/izIGvTEs9cYI9xCSd875 Xyu+ipagxZOxJQJ5VhVNSoqayYavgS+ILqkToufLCupgZFIcDTVr6Z7vGRiwcSz39Tet5AFeLhmO cp2OeXEEFuyBOz5p05x/PMkJMnEREc3EFmyFf/y+OX/C5/3os/PW+/yOuDUxYPnY7XPZFiTPRfTI gOxgZXI7j0HvNBA78IxkZuKF3RO3sKcPM88qzsgPG2aOQD4bya67cgAkWYJujfsryrDeV0L0t4Ro cgmx3krIaFZC1LnkBfwBSDpGN9Jicl+Oyaobk5+2xmTVjhHzQGcL1OhLIj6FKR2Mgkl3Qz4R/tyG OzEOKi4Z03KvuOEuidnpmt9mD4cXbqEV8Q+qLWAMJFg1tfev7bltdWrups9sWH5dXG33E1FrD1de VVXRUeRx5K9aElgMBoKHbZbsbl7VfN2RTWMvXF9bXcnp2Y7TdHXb6kWbDsSrru1bbE2vzCEa1j37 rsKgTAJZbTqagMqisgSisgSi8kwZlWfKqCyJ6De4PGSC0TyGAiiMMyZtbYoXcDoqQNk464h2FXkW Kc2STBLCq6ev9Bjy0lmSikJh4JQae3zD/vqrf3hHc9t9P7mmeNuaGp9GyROlMuUu37l81e29RQU9 d65tHm3JN8PShj8huK0me1rEt/LL733h0Y+eXecQ030mm9dqT7BpI7FI9Y3fPLD/69csCcfCKksS 4lHBzN38Lfz3UDlahjagl+MOa2Yt0YtaDdSzVhRsuKk2r4I8cwF/hSwPwHPPkagK9XKgcaPZipuW +xTmbD5PrZYf1vnINpCRPMjKU/t86rxMBVmnx/PJLNBBbtEhCnBZR3oorgcMmbPVfHHDLw1tbzsc G4v5PyyqSxeX/qK4Ye0vxOXyA4wKam68Sk3PaN4ZskB3lcTgA6MjBApnovAvyr5obwk6nVTE4YgK hO10yfsarA2KwqQ7Sd9UFwO5ZLNj7uF9OWfLD0ciJl728bfYzJ8KJuR2HVxW1OOzupYU/rFyuDUr f/vhnQP3b8oQAjliTiw35E/JX/epprRaPxYslpmZvq7s2pirb21OXczVtqHlD2KaW3v9rsa+ch8/ FvSnrI4t29OWkei0ZiUFszgdF1jcWVY+3J4TinfmB8qL8zyepozFG8OhrqXN+1ZmajWBmffWbRGL 61M7N/uL6qbXl1ZwGk9mWqpjSWVidjnR6RtmHsN/U96Kgig57uDJfMgTqfOSocs7/PobUEUMx/Kk iReroK5W19yDjixekgydIvBfNnRtWKvEpkSP1Wsz8IWtxQn+ktY8rBUSnK4EgVNu+v5M56uvzaz5 ocGiV3IqjXLzT37++s6dv/7FT7coVCqy/CQl2gclegtKFEB5J5GV7jFaZZUieJyUzEoOCzynl060 0BJGc6Ps+NjcwrfQWpDPRcLyWSenFb+VUNxSyBtA072JRqxct379egUnJLgcCRYNt2Wc8+x8/ec/ 2azUqDglmN4/wI+99ip+7PtaAfqMSqU4M7Nc3mdfCVZfBK07orbJY8B/vx7mLMeNiSgpUT2FDUdt No9qCqceTW7xLHi0EjttKbn4ucpFj9cuecTC8SsVOqN6JoxPqY06hcTjGrvoJVZQmourkUJP26Bq M3WwLnCAmaud/t3clvoz5LmbdNogGVaGB/ivoRTUehL5YK1TQLYZfTjNh92SceTGYVOhiYtosZds 3ZZ6sacYsMyD/fUena1e16hYjhplo6QCxu8oPYJGjqIFeKosRTbytCCcL5sVOM8mdSOnXc3l7VHl 5HpFC6c6oBX4mRc1QkpSUrJdq8SYf19lSRYTUiyqmeOCRWmwm3CJwqrj1zncJiWvMRuns7hXbXol mTOhJrUwn/UoA6gevXUSLZGP+SzY0J7fuKYb22/TDWwuIx7NjdvsuCmX2Gl0L1ta6dIdbbqZTXYI iTh8z3M5ZKfrqE+yWE8d9chop/ic2YKbkCHrBRxBRbAcCsf1FrEIF8WJcWkh452OMLrJDdpwnOxz tzmncJp8KnJuqzsalTe7zxKlWLDZHZ2zhf+3m918T+XuR7uWDK0uc+nJatiUt2JnQ3FXZUpua//g 1ta8sv7PrIyubl5kUyk4nmyGx6q6SgtX5Htz27YNbmvLw9vXfron1ykmu0N+Z6JVnZwaTCpakVe0 rCwnr3zlzuUt16zKNHv8Nr3FbbPCbJIQTEzMXhoqXLYoN29x205oo2p8jMviFiMzMh1Dav15BSK6 L4/CAWq0SKviLKtlZr0VPviLsPZV4g8iSf5wOEll8UIuT86eV+yF1UUU3R9P3JiJRbLyEsnQJZIm E4n9IpLWEqc4IW5BcQc5Bzd3GM4p99sFZ0XevexMCTk78jwnkJXIUYe0O3IqroUsdOFWodU3317S IUzZiInON838g79LjguULwhR7K0+ODW+feLqKtp1bZqMtvH6xvGWqLSgCdi0+I1dJw8uLd/73G4+ yFYyH/11zY2dmRkd167mXQtXN4tnP1SdA6ksQluOhhfhXJiR45XSSdZcHJOOtKbGcEha64RCOFla 7qQlY7dISGYOzszGmSk4M4iLWtNbg9l6fuEjcLDVKuhjTSydKpX+QnN7szxjMPIWLtibXcDIrqTy OoWQkJbkjyaYFDPvcR/yJm+aGMhIMPMzT6qwJSz6U2xqDgcxtvNaeygpIWDXSk90E3mVLZiYFBSw MmyyEFvYYuJ/8lGMccVTLq9JwWtM+gunFaV6M7GqzfoL31GU6YArTV6XvCpVrAAJxdDUsYocHDTI imCQFcEgK4JBVgSDrAgGslGS4ErRuwW2mSltay7Y0QQ9cy3UM/LkgG5xuqaoBhF8DuJc6a2wqMuI m08Z8MsGbLj4pEGsa+d5SaUW9Prc+QfuC1a/gSs8fA3Mq9YKUCG3V7Rrpo9e+tyVa5bUbeE6mSuf /tbla+bpD2GiXqBhySC/rSC/FDQYT0whD6lSU7CXYNiLU13kSXyGB2e4sYedspQImbPdLISQuJUE edwedzjkb3UrrXSVD9OHxYrpEEf6EAJlkyaSKFU0xeXqRZSKO6EweSKJzoDbYlDzM50abE1NTghY tQo8inE/r7GnJPlTjLwmiRyhwAolWXRNSocsNEbdhRcVFSScHLIg+zobYT58UN4pvTHug/7t10dJ a0d1BvaQWFraR8mKKB3lUXs/T1ahPHlvIE9WrTxZhfKkFZEjVK9fHPUrhCyyIvI2FJMVkdCsXP5x K6KL9ksLLLSXzR3gLvj4FZGWLNj9dnVaQ119hGyP5vZ8ZkNqTXVtOjn8b0+wqC9bFc0cY7uk+Exa SdDMVkaWUFnaANs2nfkHXRrRR8qwNMIIz7zF65TfQA6kOSIoUQyW0S5aUHYCRv24wmhPdHgCVoWK 61IYbUkOWJcplO8ZzeQJhM2o2m80a3m1wW6E/LLBxvqn1Euz44lpMZyWhYkR4sIRJ05FOK01qLck tloWHE+A6YOetpg/hoPx3CmcBUozpzOY/61RaU1LFlMcesXMuZnXlQZHSlIgbFYacffMswa1AFNb 2KlTYSe2K3W25ER/xKIwzEyUO71mJREEx09Pw4KcV5q9Tq6Nq3D6zApebdTMJODfaoxqSZ+mv016 zUPcg/xq5c1QH3fclJTqj8RcarOg0umDehAWtDkUH4qkUkUiNidZuRfZ1CryILWoKELs2EKXiw+T tbyaLyp0ggGuVvP1Js7lSjS8ksCLWVkin3DWkORyYdN775kwrF0MZ1n4K4ZEl4szvcc/pgpGUq3a z898aBZgAFN9XmtNjQRV27epg5GIVfsAVgrwmbnwAISHg+ptsO4jO5wYbN0kmGeL0ffiXj8Z/vxk +POTDuF3k0N3onSujsuKp3od5Lm0g5yhdDj0GSRxBklMj/STxBkkccbzZDd09tRxYsCHF/QT1n/e lfvPP+V+xfqPEDc+oj+l5/TeyN9zctQwiGonhZb8Kaw/op47v0YeY3dd9hwbBpSSBUaT4rKzbOok Hs/vndoWbKNif/Hynp31M89IG6nhsXvAAIouSS9YV506M+0tXtMwebqytdCzLFS7veWlD8s6KsN4 dPGW1vJ0B91ezVi5rzlrZW2xVVfQOsjhWFNBwkxXsGz59OulHYv8M8UJRa2gI+kzb+BRdA75kG5S 70pAwtkz9KkQfZBHLOi514dGVSaX5Ral0eaxWVw6rLhB707xelJc+jv8+VmZnpfUOuiqMGJj20Gf KKhUgki0sGD2Q+X1YOdXozdOgql8Kr4YpF1MlvBpxbiIYAj6WQCHRRz243ASDifiSAJOVeA0HpeW 4bJSXJaJF5G3uWEyE+QVPMG4DtpLECEHwSwHE4wbyIMDEmxeUi+lIy97VAjLhSHhGkEhxK3OOiGv PlRfemcGziBxGcTGEmzOui0ZuzO4agh1NUk7da+Q91m6TldUnIG5gL71Mj9F0EmCfujrLqq5t134 iHqBZXKFF18WUOX1CuXMv3ijKzXJn+4x8F/nuGd5oxeslQj4Zj5QKsgOX0KyVcP/guO+y2mtfo/b b9Vwr3H4VU5rC3jdieTlGLXdPP9qDHe7Vjs9Ov+ijNmu1urVHAwU2mmvVsv9XgtDBWmsaTfzcRod OwV+QLkT5aN744aKQpyWg3PiVtycMzX7sjSR5siH2nLIjGOQUHoVJucFLoKSkUF+mvfxNg1+P+51 ZmYi0mGpqeJM1itT6xNqLAt2maWXQ0pignTWIfccM2/nDJH/0SGwAxpbstcXdJtVM9df+sAbr9RY PWCYJDu0RvPM83jQqJcOfxAZ4b/OGC8/JP7RT/EunVELctNrDW5h5vmZkMVB5u4QyOw55V5kQ+2o 8kg1eoEbRDrkxx9MtreIZCvYUZid0VLXfH5RrZhxvtCsLKwLN3mkqp49I5Cnd+ShbN6bua+/d/bN ly5/HY2/kvF16XbwwncGHIFcp4N/TuuMJCVGXDqdK5KYFHFqreyQx8yOS497pNT2VzkzUnw6WILp DBqLN5RQXcqpvR7FjxLCJIdwQkLIo9V6Qhdy5jePLz8G0t11w6o0Bfn9IMFtERPUGvXm4R6fdNq6 D9bJ98G4Xo6+GTdGinCkUDo8yeO4tH7lmjfGcZE8LBdJB2vJESTyVCYV+mcqhKYSlUk1Lc8dyr0m l8+98mHS57k8GObfPgoZkoXxcUnLbFPkKCV5h8HmLiRWsCGj9O8ieddGmdHiJkM5WwXHumDxSx5n YOFV2RA+3XX2rETJJgM1DuUWUl2yAibCD140xMPqUn5hgb+v5uCRHYt2rCw0q5QcTOVqXXptf13l cEtWpOXAqsUd4QS3P5FbrDHrlHbrTGKwPnvo8FAJfmTrF4dKLR63yWDxWi0+i8aT6BWrtjSUb6jw G7whzhwQtbDoTUmduVfJFXQfAkmPIMQ/qAyApO+JG9MKcTQJpyWSsZVIOu6ionYS6TklwTmJ4JxT XOZzeSH4QyVyK5Q8z12D9LCMIOsNMpLqyRLEUlwiiiWwFM16Ls+pymoTSqZwKj1bkCtvI8SkDkxk dmbuQDcVHhkz5QfU8hYnDIOXvKKkymOzo1racnpQqTVrpwtMDrOa15kNF1b3l1gTClbkSy8oqfVq BafUuMs6t5etv70ry1l749AZLg9WYcoGshegFpKcdjBSjFi37q49m6LR5tLk5NRkjTXJYXYKJkdK 0F2wbl91+f47nh15VWuVzn8TS/BnYAmmoyx0IV5KJqhMHMnAKRGcEsahBBz24aAXp3hwyI1DMDs7 cdiBw3YcFnDYjFOUOEWBoz6cTcRtJeKGJa7TDcQpCvI7I3T9f+4EWfcnZGXBRPVRPBFSCKRZiKkE X2R+I9a/QLZIBfJ8NoIUs2eJaitgOGYvIsZ15E1ERXYs4gP7Xh/XKaIBQdAFWnV0dQOtkXc+N5es ccj0Jb+TQY4+nWHLO7llLvngi4+SzTUZG2Yt5EFPEAf4n9mtd7HzTtPvGASjklPp1PinSltSRlIg J0m4y+KYeZSbWYsfw8OB8My7bNjAgkpIctuSPC4jb9WQw2Yao/aj7wS5P0yXkjGjB1riWRgz/Oi1 k8gCKzidJYCbLIIgHzO8+Pjh2/Lxw/cliYxJL8UKU+wqQaCGg3SVIF8lRevJTsS4QJpPJR8wDbAx KYAXvCL2c+ksn2OKdokFb/O8LR/COXccrnEoLVM486i3RS/ts+bS12XP0C4Qld+RZa/Kzr8lKy0a Fh5B4J/llVrVTJbS7ErxJoctnAq/M323zabUmbTcX00OvUpx2pro85guvGSAZYwKFjSKhtQUG2i9 ypqA8OwHM3cr0KwbGZH5OFLr/qAgTysu3/ZyKpBg+WixxWq18P8hWGZeDYpJweRkUXpDe4nyJhhJ IqgMfTqeqAtYU4l+phL9TCWnDVLJaYNUsjpNneJy4zokJmQnHEzgE3LlMSRXNrBz5QdVufKDqlxZ eLlTXN4xa0BnzJzCacdcbSFFETmqZCTD8dkzxLYumX9fe+4EB6xdlMwQUF38aqxkVSvxJefdoRY6 g8reOXZ9ec59PexN7lt/dkedLa08vX6wLtWumXn60pe6R1x+iypQsWZRUsaqw/965IEPyJvdf/1C yz3XD2cuqkw224LcucGv3bqs7fbnt468eFvzyju+TrTWDFr7GmhtMuo74Y6Tsy8WokTHyKmD//EL E2Tis8D8RVRbZSUb94nyOxFEoSSzKPrtqHA6Ordtf1lHpVr0mvQCzD1sNwaY/IIMf730eoy0W3/h obltmE0aS4LNRl/TR9KJ1r8onlbuQX2o80i8QXrlqy8YRPl9fYaajjwE9s1RmEaWTWFF3LGhOV6X V1da6sw8n1DbgAznnXUq6bFrbq60TCKHs6ipc/Y0ePPoA3Dy6wYff2D1ot1kR+CTXxmbf1yOt112 pLV2oCFUkwSLZ16lUWrs5ARsHiwA7qOmoVXzeyIZq/mtwjpnKMGhhkQKrZCUFnPW9sYT+ayPOfRq NhpsPtvCM7If/9aZdeEraiDbZ2Cmfljpgnnmd/GUlCSckohTEnDQh1Ok6YXuP6RJr5tYydQhzShG OqMgMmujNFmf0mR9SpP1KU3WpzT5pbQ0spdoSnKTi9x68q23yPMQoDSnWOSBbkH4KXlEfTeuhSse sWCLDVSx4miwNU2Ywmo260+fkc7G02HudDSPaac0s0QXzCwXTf15AYtapaLro6LQ/JwCOvuwSmdU T69TG/QqFWgvNn1I5Mur9FqcrjBY3VZoLtU7GpNWWUVOv6sFr83qtWj5n9+rUxiTXBa3YFC9yCsU WKHWqy7cobV4iSZvAQv0LhjNOrDyJFoDkkwgklyDczRQ6xwypuVI9lAOEU/OFFcQ1y1rCy9b5rbh 5jiZKGBybw6TJWMcQsNx3uTTCOxtA+lKnygd5aYi94FFdRxp9PJbxcQgNcmiNcmtZSLtaQMZm8qI bVsWl476lWFJ9HITxHUksMxSZnEWStN7fVvG30RRWU8e3ywwXM+XCHNPcGBeidEnA/KIKT2KJgd1 rCUCe4c6Om/FFl5kxM6/PD5n1l7BOHMk8fxd5WNPbl+ys6PUrFHxJqO2oG2oamlvVXK0bW/zfrDB 1Cq9SbtzaX99xJvfUlDa3ZSrA4ON51QaW2n7UHzNzWszxfI1ZZVDKzLxSOcdm4sciX6TyZ7oSEkQ Q2JyeXtuUUc8GZrXYfOY1cnxzqLU+kJ/MDWoNPucZpfFZAP7LWvleO3i/pYSPacuWLEdZq1WaOfD 0M7ZaCn6etyWloXTlThN2lFID+OwDlcRsYqk8aug8Y2s3RP35eCSnPqc/hw+Cste8ohOi0wmEQ1D plLLSWuKc8fImqKMtDJcWkZ6h7SxPF6GC8tqyjaX8SlluGyKi8ZNsRAOxf8qiurCv6e3wTpDc0S9 am7LiOwYST+ZAStd6VV+8OQuXGVI6wzFpU/aii6a8BQXv9lVyB+2Z7fsf2I42rIkww5GgV6jT13c mtd9a0cGV3DPxh13d0Zyt315pOWqdfGI5dnkpRsrlqwrS/AUr1naeBv3/MqnH751a5lesFr9XqfX pDRbzY1XH17nzy7bfFvbqs/vqklrHjj0aM3BZ3dkx5b3FpRtqgplkvO/20HiX1eKnAo1SKcUZT/K R3Xo/pOogS07GnB0vAJvrsCVFTi/AqdU4IoprjJuNyQkGPYV4G0FuLEAlxbgaAEugIjnhhEWEUZy ryHd6ARpg2wDNkzNfhjXgcdQOpudrQxPYTRp66yawo4jyg0LRBztOkuMrTel4ccqSxnkvED/FfLP Ilz03FJ9yYlg9t7c1/N3HN7ZcmDd4pBgzVq++/BgqCmeYYI1CFbrtfpwYXNe143tabx3SfOqnP47 O8PPuArXLA01VFd4AxXrK+LryxPxl9of3luf2rDj0JfXtz350K1bFmnNVr3RbDNZvYLGZDE1HXxi nTnJbS7pu2Vj6YalKUaX3/qpZ/ozs1v6iGxv5nfzWTAf+1ARMh5TJTtza1FF3hkyi150dkQ+IqG+ Qqj0BPyw3hV0u5OdepXRJdykNFg9VsGpw8oZ1xUiHHqFovZqL4y6KqvoTcrLyvSc0ejU5H1Rzcz5 j4kgpY3yu7mfzJVWH3HlzZV2bpsxHM6f32dUXnH3kfsJKczNCiOM/VAY/nqdK+hxBZ36mQcWREDx FVIMKb0y4ofSuM/ASltBfiAGW67xihaVyiJ6Py5COtNHSo3uva3wlQ3mRf9AHo3087cv/PHAj9hP 4YJdPao6p/wGUC1JK30w+eXg6V8u+MVcfMkP+jYpTAgpm1AP/wWUzF9AXdwP0P38K2gdfhXdj38M lsAmtEaRD/xVdAP3DAoAv4d/Gz2teBQFFcmAxRCfCghp+ABqUXSgp7mdKKR6DT0N15hJPtwqhKT4 Ljk94WmoU2lAa8m9FH+Ca5Tos5wCPa3qRfcrxuE+4OBact1GRRXq5k+jAm4AyjCA9vFfRU/zD0N5 v4pquVFUrUxGT6qT0GLFR5D/IZRMruFdCCsiKJv/CD2kWANlmUXpyilUwH8J8l+EQlDvPnAj4LLB 9fAfzX6g3AT4e2RWnJfinwG3BVwruO3E8Z3oZhSV5PYo9uCXuZ3cO/z9/K/5Xyu2Kv6k7FO+qdqp Oq/er/675lot1j6g26jX6z9tKDPUGEYMjxmLjTca7zKlmW42x8xl5sfNPxAcwj8tnGWP5UFrnnXC 9gXb7+2F9ocdCserjg+c17scruPuiPuU+23PCs/T3k3eu33bE5QJn0tsTjyfNOlv8Y+IRnEskBz4 RfKZ4I3BT6csSrk75Y9SCzejzUiJOpACNEFAMbCSUdCdogc/iS3grkX0J7gR6pW+eUkzTJKPl3TN RPZrJM5Lv3BFuQLCJ2SuRG70DZmrIPxVmavRh+i3MtegdPyGzLVI5DQy13GPcKky16NViu/J3ABt lCRzI/c5ZZ3MTWiH+pzMBWTSNMncAnzNnE7narbKHCO15gGZc8AflTmP3JrHZa6A8FMyVyKD5rsy V0H4z2SuRldpfilzDXJoV8tciwTtDpnr8ArtAZnrUVT3o7lfL3fo/iVzmIf1BplDfzN0yFxA6jlu Ab5B5lbgm6G0WKElbWQYkDm0keEnMoc2MvxG5tBGhvMyhzYyJsoc2sgYlTm0kbFS5tBGxhaZQxuZ GmUObWQalzm0kelemUMbWQpkDm1kuVXm0EaW4zKHNrI3yxzayH5S5tBG9u/K3Ar8FfQEElEuaFU2 KgDWLP2q9AgaQqPgNqMxCKuUfo2b/iZ3N4T0AxuE9YiIlqAd8CeiVgjbgrZC3Kjk6wPsg9S74LsX UlbCdTsgzSYI65fit6BxCOkG/+V3LJXuufCK0gVlvDhmlXSfUblMJFWWlO7iVBf7RHCknH2AY1Bq koMIKURAUrJ+6Veu+yRfL4SOSfXuBd+AVOLtEDY0d82VYzf/r2QpSr+LLsqlEVE7+PqlMpD7twHr lnyj0j0HITQml2BoQQ16wDcOsWNSLUnqrP9FGZrg2h6UCiGjKA1S9UolqZWuJXe5sgyJBAYgvlcq A6nDqFTCUYn1SWmJLDZD6ADwHWgv+HbLkhel33HfBHyHdLcRuQa9kjy2SLkMybmOSRKelwCtL7kn 1QCij/VS/TZDSLf0O+UjssxGpJAdUqnH5Hr0QEyGlPOAFLJDyrEb5ELD2V0GJE0lUhqWSzkIIQPS XWmepJ5jC0pA7jgs1YX9jj2VMC07udMQSECUfsF9iySFfuk328lv4Y8t0AXWp6jM6F1EqeyDcr2o nm2SUs6XeGGNiNT2SNfRWm8Hf9Zl/Ssi5TYg5bBXksO43HsXyptpJ7n7bkmq3XM9pV9ubXpH0tYi 5DE8Vxtaxi1yGtJf98m5j0EtaAvtmmulbklHSG8auKheTIN7oCTd0v175PtnSZIagzuWQt+ISf83 wG4IHbhM/7MkvRmANGNQV9JCW6SchiGHvRBKctwstRdpyYtzZeGbpf/dYESSJ8uvU9JdKsW9Uu1H JWmNSe08KuklvVqU5EZ0pE+qYb90D9rXN0nXMklXw0jQBKMsvXZkQQzVr16pz87rzG75fwXY+jH3 pX6StgfkPC712t65NuiV4oelcXnvArkPSzUdlCVP8+qTvokmXVpvEk81NhWuSpPG2QGoV9+cDl1e qsHLcv6fy2g+dzZqiHK/p+Ngz0X97/K6z4+8F5erbIEESE1oXegoxMbOkbkRrVfq04NS3+7+2JpS OXdfJNM+eRy/dDQnUiWaNy5d2Sv1D1Kbvrl8SModUh/7pBb6P9Uv5vtETCoN6QN0ZMyS2moYxplP mOefEHOzswvE5v6ekaHRoc1jYuXQyPDQSPdY/9Bglrhkxw6xtX/L1rFRsbVvtG9kV19vVmX3jv5N I/2tfVvGd3SPzF1YKsoRpVKOsmdV38go5CTmZmVny0EyiP2jYl//2Na+EbFbHOnb0j861jfS1yuO jXT39g10j2wXh0jMAu/mK5dS7B8UIRuxfbB/DK5vG+se6xsVuwd7Y5DBkHSDnqHxwbGR/r7RrCvm 0DTek9o9mib29om1I0NDYwtK2C0ODPX2jQyKo92DoyJIoH+zuLl7oH/HXnE3FF4cHd80tqNPHIEb 9PYPbhkVoTxQkQGpAHDfkUEQQJZYPyZu7useGx+Bko30de8Q+8fgHj2jGeLoQDfIuKd7GDi5ZGB8 x1j/MGQ5OD7QNwIpR/vGpAxGxeGRISgxKTDkvmPH0G5xKzSN2D8w3N0zJkmBtBSUDC4Rd/QPwr1A Zpv6t0gZ0xuN9e0Zg4v7t/dlsfaKjIoD3YN7xZ5xaF5abiLOwb7d4kg3aZR+qDZc2D0gjg+T20CO WyBktH8fJB8bggrtIlXqFnd3jwzQexEB92ztHoGC9Y1kbR0bGy6NxXbv3p01wOSf1TM0EBvbOzy0 ZaR7eOveWM/Y5qHBsVE5KeGbu6Fw20m6zqFxKOJecXy0D4oGrUKixW6QSN/IQP8YafVNe6VCV7c3 LYHYEckD8uodp5LZvbW/Z+uCawH7B3t2jPcShRsSe/tHh3fADUjZh0f6IUEPpOobHMsS2b2HBkGw qf1pYt/AJnLRfFaDLPEVSyQlJ6oBYhoFHeyh7Td3d0l55bzKpAKk9sNdQIWIdo4QResd2j24Y6h7 4U2hzN20pNAQc2o+ND42PD4Garyrv6ePpNnat2P4kgpd3Htl378XG/9ebPx7sfHvxQb692Lj34uN fy82/r3Y+H9psXFJ3Nz++sd+JrW8OMVdf0zrxg1ArmPkWkY+xchBRq5h5GpGrmLkACP7GdnHyF5G 9jCym5FdjIwzMsbIKCM7GRlmZIiRQUYGGNnByHZGtjHSz8hWRrYwspmRPkZ6GelhZBMj3YxsZGQD I+sZ6WJkHSNrGVnDSCcjHYysZmQVI+2MrGSkjZFWRloYWcHIckaWMdLMSBMjjYw0MFLPSB0jtYzU MFLNSBUjlYwsZWQJI3FGKhgpZ2QxI4sYKWOklJESRooZKWKkkJECRvIZyWMkl5EcRrIZiTGSxUgm IxmMRBlJZySNkVRGIoyEGQkxksJIkJFkRgKMiIz4GUliJJGRBEZ8jHgZ8TDiZsTFiJMRByN2RmyM WBmxMCIwYmbExIiREQMjekZ0jGgZ0TCiZkTFiJIRBSM8IxwjmBEkEzzLyAwj04x8xMgFRj5k5ANG 3mfkX4z8k5F/MPJ3Rv7GyF8ZeY+Rdxn5CyN/ZuQ8I39i5I+MvMPIHxh5m5G3GPk9I79j5LeM/Bcj /8nIm4ycY+Q3jLzByOuM/JqRXzHyS0Z+wcjPGXmNkVcZeYWRs4z8jJGfMvITRl5m5MeMvMTIGUZ+ xMgPGfkBI99n5HuMfJeR7zDybUZOM/IfjHyLkW8ycoqRFxn5BiNfZ+RrjLzAyPOMnGRkipETjDzH yHFGjjFylJFJRo4wMsHIs4w8w8hXGXmakacYeZKRJxh5nJHHGDnMyFcY+TIjX2Lki4w8ysgjjDzM yEOMfIGRBxn5PCMPMHI/I59j5LOM3MfIvYzcw8jdjNzFyGcYuZOROxj5NCO3M3IbI7cycoiRWxi5 mZGbGLmRkRsYYWYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYPZmYP ZmYPZmYPHmGE2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y 2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+Y2T+YmT2YmT2YmT2Y WTuYWTuYWTuYWTuYWTuYWTuYWTuYWTuYWTu48ighYDVPJpX7wWaeTHIAXEt9n5pMKgU4SH3XULh6 MskAcBX1HaCwn8I+CnsnE5cA7JlMrATYTWEXhXEaN0Z9oxRGaODOycSlAMMUhigM0iQDFHZQ2D6Z UA2wjUI/ha0UtlDYPJlQBdBHfb0UeihsotBNYSOFDRTW0+u6qG8dhbUU1lDopNBBYTWFVRTaKayk 0EahlUILhRUUllNYRqGZQhOFRgoNk756gHoKdZO+BoBaCjWTvkaA6klfE0AVhUoKS2ncEnpdnEIF va6cwmIKi2jKMgql9PISCsUUiigUUiigmeVTyKO55FLIoZBNM4tRyKLXZVLIoBClkE4hjUIqhQjN OkwhRPNMoRCkkEyzDlAQ6XV+CkkUEikkUPBR8E56lwF4KLgnvcsBXBScNNBBwU4DbRSsFCw0TqBg poEmCkYKBhqnp6CjoKVxGgpqCqpJzwoA5aSnBUBBgaeBHPVhCkgCPEthRkqCp6nvIwoXKHxI4z6g vvcp/IvCPyn8Y9K9EuDvk+42gL9R318pvEfhXRr3F+r7M4XzFP5E4/5I4R0a+AcKb1N4i8LvaZLf Ud9vqe+/qO8/KbxJ4RyN+w2FN2jg6xR+TeFXFH5Jk/yC+n5O4bVJ12qAVyddqwBeoXCWBv6Mwk8p /ITCyzTJjym8RAPPUPgRhR9S+AFN8n0K36OB36XwHQrfpnCawn/QlN+ivm9SOEXhRRr3DQpfp4Ff o/AChecpnKQwRVOeoL7nKByncIzC0UlnBcDkpHMtwBEKExSepfAMha9SeJrCUxSenHTCeI2foLk8 TuExGneYwlcofJnClyh8kcKjFB6h8DDN7CGayxcoPEjjPk/hAQr3U/gcveCz1HcfhXsp3EPj7qa5 3EXhMzTuTgp3UPg0hdsp3EZT3kp9hyjcQuFmCjdRuHHS0Q1ww6RjE8D1FK6bdGwGuJbCpyYd7QAH Jx0wGONrJh2FAFdTuIpefoBet5/CvklHL8BeevkeCrsp7KIwTmGMwijNeoRevpPC8KSjB2CIZjZI Uw5Q2EFhO4VtFPrpdVspbKEl20wv76PQS1P2UNhEoZvCRgobKKynle6iJVtHYS2t9BqadSe9UQeF 1bS4q+iN2mkuKym0UWil0DJpjwOsmLSTOyyftBP1XjZpvw6gedKeCdBEkzRSaJi0g12A66mvjkIt DayZtF8NUD1pvwmgatJ+DUDlpP0gwNJJaw3AEgpxChUUyietML/jxdS3aNLSCVBGoXTSQlSjhELx pKUWoGjS0gFQOGlZA1BA4/Ip5E1aMgByacqcSQupWPakhfTNGIUsenkmvUMGhSjNLJ1CGs0slUKE QphCaNJCpJRCIUjzTKZ5BmhmIs3FTyGJXpdIIYGCj4KXgmdS6AJwTwrrAVyTwgYAJwUHBTsFGwUr vcBCLxBooJmCiYKRgoGm1NOUOhqopaChoKagoimVNKWCBvIUOAqYAorPmjf5iZsx9/inzb3+j4Bf APchuA8g7H0I+xe4f4L7B7i/Q/jfwP0V4t4D/7vg/gLuz+DOQ/ifwP0R4t4B/x/AvQ3uLXC/N23x /8601f9bcP8F7j/BvQlh5wB/A+4NcK+D/9eAvwL3S3C/APdz43b/a8Yc/6uArxh3+M8aw/6fgfsp 8J8Yo/6Xwf0Y3EsQfwbCfmQc8P8Q+A+Afx/494zb/N819vu/Y9zq/7Zxi/80XPsfkN+3wH0TXHz2 FHy/CO4b4L5u2On/mmHE/4Jh1P+8Ycx/EtwUuBMQ/hy44xB3DOKOQtgkuCPgJsA9q9/rf0a/z/9V /QH/0/qr/E/pr/Y/Ce4JcI+DewzcYXBf0Wf6vwz4JXBfhGseBXxEv93/MPCHgH8B3IPAPw95PQB5 3Q95fQ7CPgvuPnD3grsH3N3g7oLrPgP53alb5r9Dt9z/ad0W/+26r/hv0z3mv4EP+a/ni/3X4WL/ te0H2z/11MH2a9qvar/6qava9Vdh/VW+qxqv2n/VU1f96qp4s0p3oH1f+/6n9rXvbd/dvuep3e27 nhpvV4zbx8fG+b+P46fGcdU4zh7HHBoXxsVx3jDWPtI++tRIOxpZMXJwZGJEUTYxcm6EQyOY/Nje 0RFfUg15PezAiFGo2dk+1D781FD74OaB9m1QrP7iLe1bn9rSvrm4t73vqd72nuJN7d3FG9s3FHe1 r3+qq31d8Zr2tU+tae8s7mhfDelXFa9sb39qZXtbcUt761Mt7cuLl7Uvg/Dm4sb2pqca2xuK69rr n6prry2uaa+GKqMEIUFM4AVSgGUJUBLkw0uzfXHfOd+7PgXyTfhO+Xir2ev3cmlmD65c7sFDnms8 d3h4s/vHbi7uTsuoMbt+7PqN6y8uhS3uSsuqQU7BKTp56aV1Z/PKGgkrqijmFEh1bXYGwzVmBzY7 /A6u2u/AyHLO8q6Fd7wo/FjgzGZsNs+aubgZkptNfhNHvmZNfNyUU1RjNvqNHPmaNfLOuBFCSI4R w4qVNWa9X8+1V+iX67m4vqKyJq7PzK5BPBYxRlgA4DXktWjs8NfwX8PkQL4SYXznkZVt0WjjlAa1 Nk5oVqydwDdPhNrId7xlzYTq5gnUvmZtxxGMP915BHOVKyfsjS1rqP+G229HiUsbJxLbOib5Rx5J XNrZOHGQ8Hhc4rOEI0jSGV0/Oj4ajY6th6/1o2NR6R/48DjxRUkg+Tc6Bn7yNy75UfQTPzQZwIZR +IzJYWOffNH/3z/4/+sC/N//OYJATTuWzHLXo17uOnDXgvsUuIPgrgF3NbirwB0Atx/cPnB7we0B txvcLnDj4MbAjYLbCW4Y3BC4QXAD4HaA2w5uG7h+cFvBbQG3GVwfuF5wPeA2gesGtxHcBnDrwXWB WwduLbg14DrBdYBbDW4VuHZwK8G1gWsF1wJuBbjl4JaBawbXBK4RXAO4enB14GrB1YCrBlcFrhLc UnBLwMXB/T/V1EWSEEYAheF5OEGChsE1QIKT4FZzDDZkMyfgBGzQ4Anu7gR3d3cNBPfg7vDdgqKr vl70qjfv70NvetGTHnSnG13pQmc68Tu/0ZEOtKcdbWlDa1rxK7/QkhY052ea0ZQmNKYRDWlAfepR lzrUppBa/ERNalCdalSlCj9SmUpUpAI/UJ5ylKUMpSlV9MVdkhKEgoLieMtnPvGRD7znHW95w2te 8ZIXPOcZT3nCYx7xkP95wH3ucZc73OYWN7nBda5xlSv8x2Uu8S8XucB5znGWM5zmFCc5wXGOcZQj HOYQBznAfvaxlz3sZhc72cF2trGVLWxmExvZwHrWsZY1rGYVK1nBPyxnGUtZwmIWsZAFzGcec5nD bGYxkxlMZxpTmcJkJjGRCYxnHH/zF2MZw2hGMZIRDOdPhjGUIQxmEAMLiosGxP5j/7H/2H/sP/Yf +4/9x/5j/7H/2H/sP/Yf+4/9x/5j/+mPBkQDogHRgGhANCAaEA2IBkQDogHRgGhANCAaEA2IBkQD ogHRgGhANCAaEA2IBkQDogHRgGhANCAaEA2I/cf+Y/+x/dh+bD+2H9uP7cf2Y/ux/dj+t+7wd376 fusPfOen8I9+BV8BIU9ztwplbmRzdHJlYW0KZW5kb2JqCgoyOSAwIG9iagoxOTk5NAplbmRvYmoK CjMwIDAgb2JqCjw8L1R5cGUvRm9udERlc2NyaXB0b3IvRm9udE5hbWUvQkFBQUFBK0NhbGlicmkK L0ZsYWdzIDQKL0ZvbnRCQm94Wy00NzYgLTE5MyAxMjEzIDk1M10vSXRhbGljQW5nbGUgMAovQXNj ZW50IDk1MgovRGVzY2VudCAtMjY4Ci9DYXBIZWlnaHQgOTUyCi9TdGVtViA4MAovRm9udEZpbGUy IDI4IDAgUj4+CmVuZG9iagoKMzEgMCBvYmoKPDwvTGVuZ3RoIDUyOS9GaWx0ZXIvRmxhdGVEZWNv ZGU+PgpzdHJlYW0KeJxd1M2OmzAUBeA9T8FyuhiB7zVhIkWRMvmRsuiPmvYBCDgpUgOIMIu8fTn3 uI3UxUQHY64/X8Zk2+Pu2LVT9m3s61OY0kvbNWO49x9jHdJzuLZd4iRt2nqKV/Zb36ohyeZnT4/7 FG7H7tKvVkn2fb53n8ZH+rJp+nP4lGRfxyaMbXdNX35uT/P16WMYfodb6KY0T9brtAmXuc7navhS 3UJmT70em/l2Oz1e50eeE348hpCKXTtS6r4J96Gqw1h115Cs8nydrg6HdRK65r97ZXzkfKl/VeM8 1c1T89zPk1e5WC4XyGp5USB7jgtywXGHvGDeIpeWJUd+47giL5nfkDfMS+R31vTIW2Zba0eP5T2z rXWwXKC+y1kHZhf978j0L+B09JcwOPoLG49+7NHR73fI9O8tR/8emX6PPTr6D1Yz+s0T/Rvk6C+R 6d/D7+j32K/QL6gv0Y+eSPTDIPrsj9C/sEy/h0Fi/7GWxP6jz0K/WM3ox36F/tLG6S+sDv1i69Jf oP9Cf2H1ox/1lf4CNZX+EntU+gXvRekvbD79gv5o9ONdKP2C3ir9JdbS2H+8U10+35fSr7Ap/Wp1 6Ff0R+lXqxP95on/P3D66Md87571Pf1qc+hX9MfTr+ibp19tPvySO/h9ybyzQxdPF44fvg9/j3Va f4zjfKTtI2JnGae47cK/78zQD3jK/v4A0dYRtgplbmRzdHJlYW0KZW5kb2JqCgozMiAwIG9iago8 PC9UeXBlL0ZvbnQvU3VidHlwZS9UcnVlVHlwZS9CYXNlRm9udC9CQUFBQUErQ2FsaWJyaQovRmly c3RDaGFyIDAKL0xhc3RDaGFyIDcxCi9XaWR0aHNbNTA2IDY2MiA0NTEgNDk3IDM0OCA0NzkgMjI5 IDIyNiA0MjIgNTI1IDIyOSAzMzQgNTI1IDQ4OCA1NzggNTE2CjUyNyA1MTkgNTI1IDM5MSA1NDIg MzA1IDg1NCAyMjkgNTI1IDQyMCA1MjcgNTI1IDM5NSA2MzAgNDc5IDYxNQoyNTIgMjUxIDc5OCA1 NTcgNTI1IDUzMyA0NzAgMzAzIDMwMyA0NTkgNDUyIDQ1OSAzMDYgNjQxIDg4OSA2MjMKNTY3IDUy NSAzODYgNTE5IDQ5OCA1NDMgMjQ5IDcxNCA0NTQgNTI5IDUwNiA1MDYgNTA2IDUwNiA0OTggNjcy CjQ4NyA2MTUgNTA2IDUwNiA1MDYgNTA2IDQxOCA0MTggXQovRm9udERlc2NyaXB0b3IgMzAgMCBS Ci9Ub1VuaWNvZGUgMzEgMCBSCj4+CmVuZG9iagoKMzMgMCBvYmoKPDwvTGVuZ3RoIDM0IDAgUi9G aWx0ZXIvRmxhdGVEZWNvZGUvTGVuZ3RoMSAxODUwMD4+CnN0cmVhbQp4nO18CVhbx7X/3Hu170JI AgToChkMCBAGs3pBZrNZvGCQA14RQoBsQFgSduzEiZs2dR5Z2sRJmrTN0jVNaVpZ2XCbNmnqNK+L m7RN3TVLX9e0pUmXrK3hf2bOFQbipOn3ve97//e9CEa/M2fOOXPmnDMzVxZJPDoZIjpynAjEFxwL THgzbWmEkO8SwqUFD8XFJ5/6mA7oFwiRG4Ymhse6m57lCFG8TIjaMjx6ZOjnL/OthJhkhBjGR0KB weknArmE5MbBRvUIMF5U1Kmg/wXorxgZi196k5Asgf7T0K8djQQDWVlp2YQ4QYbkjgUunfgFf1AN /ULoi+OBsZBx121XQn8TIfafTkRiYFeAoZoVdHwiGprY/aOXX4J+M/hwHngc/NAX9VlB+7wgkyuU KrVGq9OT/5Mv5Z2EzJ1cwtpG9pMY5Pw4+SC5npwkj5KfkwHyfqBuJ3eTz5J7SYJ8nXyL/Pi/04+5 I/IxohMeJgpiIWT+zfnZuc9Cm5EbFnFOQs8iEy9w5k3zf17G+/PcyXnT3IwijWiYrp7/AXD/xp2f f5NvoP35atrnTwBtZBp/Ud4596W5e5a40046SQ/xkx3kEtJHtpIt0LaRLrKZ7CH7SIAEySAJkSEy TEZIGOJ1gIySMTIObYhEyAQ5SKIQwziZJIeAjksc7F9KjpCj5JiEl5HLgT4C70cZdQW5EiL/vgW8 agEvcN5Prob2AXj/IDlBriH/AUjfl/KW9qbIteQ6yOcN5EML9IcuyqX0h8kt0G4kN0HWbwb6Nsj9 R8nHyMcZ9yS5lXyE9e4in4TxW5fI0rEL8neQO0HqbvIJkPwUVM89y2Sp5F3kEfJVqKlvkq9BtT0K 1OPkNNCPk+fJC+TX5Hfk9+RFzsNVcxvJX8nfyVMQ/SGIOo35BHsPw/vwQsQPQ2xTkb0CIrY0Doek MYznVSxOqbHDIHkCsnHVIp0plqeULSqdsrU4XnRNdEUXeLjCkwucC+teqoVyi2O2NIIfZZylo8sj u5j+xNuOfIp8Btqn4Z3mYXkvRX0OdjhtnyfT5AtA4fuFfoq6j3yRfAnOglPkfvIgeYg8TGYW+g9A 78J4knFSMhfnf5l8hVXBo+Qxlv9vkDOM9yhQp6XRR6WRLzP6cfIknELfgdvoLHkCaudJ1r5Dvgf1 8X3yAzi1fkGekyroHKsgN+chT5PvywrIT+QGTi48Rh7nt5BLof9j/nbIBPG17tu7Z/eunX29/p7u 7V3btm7Z3NnR3rZpY2tLc1PjBl/D+nVr19TX1dZUV3nLSksKC/JXuPOcGelmk1Gv1ahVSoVcJvAc KWlxt/aLiYL+hKzAvWlTKe27A8AILGL0J0RgtS6VSYj9TExcKukDyaFlkj6U9C1IciZxLVlbWiK2 uMXE2Wa3OMPt7OoF+vpmd5+YmGX0ZkbLClhHDx2XCzTEloyRZjHB9YstidZDI1Mt/c1g75RW0+Ru CmlKS8gpjRZILVCJQvfEKa5wPccIvrCl/hRPVHo6bULIbwkMJrZ19bY0O1yuPsYjTcxWQtGUUDJb Ypj6TK4VT5U8NnXdjIkM9Ht0g+7BwO7ehBAApSmhZWrqRMLsSRS5mxNFR3+dAUsOJUrczS0JjxuM dWxfmIBLyPNNbnHqFQLOu2f/tJQTkDiKfNMrhJJ0iQthgvEUTcA38BDW53JRX66d8ZEB6CSOd/Vi XyQDjiTxeT19Cb6fjjyWGrH66cjx1MiCer/bRVPV0i/9HhrJSBwfEEtLIPrsNx9+YVxMCAX9A8ER ioHQlLu5GePW05vwNQPhC0hrbTlV7gX5QD8sIkzD0NWb8LonEunuRhQAhkhzEO7uZSqSWiK9KUH6 g5JWwtvSTP0SW6b6m9FBasvd1XuaVM6/cGq16Li/kqwmfdSPhK0JklLQMtU7OJRw9jsGoT6HxF6H K+Hrg/D1uXtDfTRLblOi6AWYzsVmZFqwtmXSKWG6cmW+SuzlHUIfzRYwxFZ4czeuhQETpIt1aUYb 14q9nIOkxGAWSYJSS+xAR8hv2kSHBKratMnh6nPh6x1cckg+yfMTqkW2TMBY8AnneVvXUJo6VCS2 hJoXObjEqFxyULJ2cT95GgtpYtBQ0XRuSg0J+bBzgceDGcaiWcwQE2Sb2OsOufvcUEO+bb10bTTW LL8d3e6Orp29LNtSlfQs6eF4LfYSxAXDqQ7fBDXY6nGk0sr6G1l/obtp2XBbalicUrk7uqeocbdk kIiwg2DRioK2wLW1aatha7bC6eZuDbhFk9g6FZiZPz4wdcrnm5po6R+ppzbcbYNT7u7etQ7m6/be Y46jdKo00sF19DSWlsDZ03jKzV3TdcrHXdO9s/e0CR79r+npTfIc39Tf2HdqBYz1nhbhbGdcnnIp k3ZE2qGWtkNHxeQdp32EHGejMsZg/eAMRxhPleJxJDjDI8+U4vHAkyHPx3j0BUnKGIEQw3HbIg7S 9FzeNzLV30c3F7FBKuGXS3Du9STBu9ef4niFLqFxhxoTWncj5TdQfgPyFZSvhMLgbBwEh55JU/1u OKegoHqJg8NSFKhJcWZ+vqfXddYx2+eCUtsNbWdvQu2Bs1+e3w5yG2nrB/bGxPFggPpB/L1UV5nf FuyDsk0ZBJG2hBosqCULINHKdGg5glIQcgMJZPrHoZM43pfo89BJe8N9rJxNCbLJXQ9pR5vyAjqR t28qzV3B9iZsBU3+CQpq8I109yLHAV2YrA+DpNSB50E3DAX7RYi2jAS7odTxLNU4kBOCI1FWEGJN 45AGCV2WkK/VaxLqMjAIv5TWltEtKc9X9vWh86x3QhKAuU0JLXhUsCiUkgJEB4baqC/wewJcpaJf p2a6Zsh296VwslCnmSUlDCf0+W0BOPxRXwscd21KWUXPCK1k4wxylXTlOoi7kN8zM3+P+4hr0au0 xE0vB1qYxHEaCpv0TS1nJHZ5SktUy7l6xp6aUukvroDxUukXEJjw0R4+ncWEn8GnKYEoyRr2qWjr g6W2Uptq7QYNN0vaiJIbhOoXueuIinDcoC9NxufXKIQuh9480cV1NSv5HtLw7HPP7nnu2bOAZznv s7PnZk3nz82m1dV5vavKObPLzFq6gVcqFQp3XhlfU1NdXVlZsZ6vWl3Gu/MM0AqqVq/na9YLlRW5 PBNFScYFYcoVfvbPXcLW8wr+MmfL+JYVvNNhSNfJOVHutKvWbS2zGF1VhYU+r1OpUfBylUJVVN+c 17y3PmvuQUGpVWpEmy3LIJcpdSq1mGnJNMjmWuWGN/8qN/yjSTb6j5uFVauHt1fLb9OoeJlC8YjD nr+m1ZXpES1Gi0lnkFtsaQqlJU1bsK79/LUqe5ZdqdEodSaNOiPDplJrFDrT+Vo4Ohrm35T/Tvg2 2UXu+TJ/JX+c9Hpggzf5e311zaWlZXV2W55rS94usktLNJxF0NTmabvbzYVv+Nraa8tcdg2xlWrz dm1prjNUrm+r7MzulHcWejIa4JVmr2vI2nz+TAPn9XrPwMOsubLCnFZ3BuCbQFWuKnf4XP+urVXl ffkGgaXEUl1dtbqAJkMpvAMLM6RUItoRrOkKpXyY01hdGRlOs4KPyJWr1lUc6G83b+ORaVJwEaW8 Ym3lfmSmuzIynWY5/yCntuSkWx1GGZdnbNk9Vrd2R02WkN66e7S2aWdthkqZqdKqZDJ440tyVmf5 OtcdOLljbvwCM7c+c13bUqZwNeRYEFQ65RfyOn3FrrqtXndnQ3FBY+9qd/MKAuUMWeJeknuJhRSR /Uvz9ECRMz2XzPADPq3GmZub7iySrcg0znAbH5L7VrRlYgSznts8a4Ygct5zz8yaK1nwH/4XshBq TgqiVOXLQsq9KDdnF+fmFKTxckWaA6h8Cz/32oVY3c8BmQkBlnOy72YX2DUae0F2dn6mWp2Z/49V C6v/AKweKJ2SrrRrflZWJx8jJaSBnFhWkRpXaUMeDKjzamb4XF+WNa9EWNkKTLWKKAzlb2S317/u M2wWpHVsnm2YpYUD+5rzzp6p81ZWHDthOHOGFl7mu1alFZc6BFgAZLDFZayWFgJRQGtOZk230SOi RigVVhRbs0xyXtQ374muWXPJmnyr0lbesf/aXk/n+tXWIYiLCHFJk/Nz57xjzVX+5lWiqbGtJtLf npZpNsiUWvXnXB0NRc7VLW3tzvoP33jtgUaLqyyTm1PqlXI5vJ0faNm0avv+g+NlgeG1+2+6BHby Ju4Bvo5fR6zE/iDRpv1KVZgKn8c7y3nPwsmWbpC5XezkklW6Kmx8HadQac2auT16HRxAOqOW+5RB y2cr0u32dF1amtxiz7Bo08D2wfmX+Cdk95EWMrUsJ0XVJZ4aT6NKvUG9oUbt8ZTX2GvspLxxU82G taqSX6k9rupNxtd9rs00sDSuENLZirN1dXAsnKUHblodrUfTsRNnzphOYHIs70KbViddjltQKKTA V1bYrPQgxjJlB7FFOqcrK2tS5wBI8U/wCo3WoP5tSKbwlDsKc2wqlRrWr1SJxV577fZaBy+XC6Fj Wp1CZ9Ff4eG06eyckHOe3xo1wkm11WaDqGmsq82VXrVGrTXqnbkZSqVBq8io3FytyxFFA/em3mLI F23nlDq1TKbWKc/ZoLo3z88Kd8BpW0bWLq/uh4orahQyop7hDT6126zLFdLT3d4ZXu+zErfiazU1 xblms67i+8Xtuud9uQsBmTXXeeFUpRl+Bm4wr73ONGummxfC+C60MIy82/32UWS1rpSOVSoh3OG7 5pmb9yvlwYhvqKNcrVbLVHqVbl3PYEXfB/tKMqt3HP74QM9kR96929o3DG6uMQ+Fr/e7+d/AHVjs Wu8Y3G+xWfQ6TXZOllpnt+gKuy/v2XDLTR8cWl/c2FVT2VDaGarNKl2LdxP/V4hWJRldfhIUms05 adkkJ1s3w2X4TL7S9rRsc2HOSoU9r82OmxjvDIiL9wytMHbmnSa6fyGeqqs8vOff9h6xgWdwZulV Kos9Nz2vb8dG89alJ590S7jsDe1dK83uXLtCIXxcZs8VHWlKjXLNyA3dc5G33gCfLuqsy5Mr1QoF wQgICYhAPblsWQQKVmVlOQpkBoEYuXTBaMi3vuGrbs93GGRZxlUFKtHTJnaqpetzIRJn6DUMV+/C /Wv711osIIp/44oVEirlQV5tFfHkh1t09bqqyL6Ot71Fm/ceXEfvTv4c3Hbnv3fhqnTW29Z1vONV WdTUC1HqhF21Q/gWqVq+p+7PziZmelXkFK5+rdAp5+SaV73t4quFJNOUyWuEzPQ3ffnSpjj/DIHg eGYbgADAG4PGKPvfVU3tKSwh2ZJ9pXBLx29FrkzYoTRYdIbcVR31vmBbea5+Z9+GPRuKTSq1TK3P WLt196pP3GWt2BK9NVDYvqEqRylsSStw2XJW5Fb5R8eHC4b3i0Wi0aBzuXMzV+RYPv3JdTeenDrg 09tcWWlQO/N7504KB+Yz4D7IeAjug+dVWy9cCA0XvxCEA/QaMGj+uU6v4SHccNg9rtfNvaKwWG0W rcUAaE/XWuCenn997mbhaajMYrinb1pWm3pXVbVOX6WvytDbMwiNf7bHrquucsmU5W8UtNv1GaIs zdGWtrXudZ9y4UxiBTfLHlKemfVKlwJc12aOlWrWuzewaA+vXFS3i4tUKT3HK5U2mh2Z8DTez6Xt 61fb/LwanlvscD9z53l2lUOZGprgKu8c9mXfB4/tK6sn+jvMLu9oc9UlTatcRv4D9TfefO3ohjTR kzW3TYXPMyrZH0wGAep12tWxoXj1JYe3ejatzl47evKSz7S2VPTsj07g847wS4hjAXyauX5Z9apz VLkz3JceKFhZsEY1w33xYWIsgGfwglU0pHYLUa9ZmVOgEFxtxW9ktVcvf35pYKGcfQZK89lZGk04 9Nm9YHsXim956pFuApu9JhXI1ENPLo9PPSWyFcXpWSYwy556toXX263ejv3X9fVdWWGRFRSmOyCM P4THnWoImdOodVZ7Fj/wfF6E/Vy7O76u9oabr4s0bWrYBcGDff6nlpbKngPR8RJ3S5173ehNvfRc bJ9/jSsmLxA7cT4C1RglOmLh/Ell59ISh62YLxVC6iLjiuUaY0ZJXvaKdJVca7QXFeS405W3GwoK 8zJ0KovDbFpZ4LLrNZYs+t0b/Y6WJ+TVK17bt8+49hWSqWJfyX3lj5d/N/X1HOyzmPx38MTKEzWV JZKe8s7zP7vwJR58Fv3okm/1DDIDIQr4KMqdIw2yHNLFx8gm/ovkoPADspn/GuM1CD2kU/jn/F5Z zvzrTEZO2kH3c9xl3Mv89fxXhH3Cq7LDco38FvkZxYhSUNZIPhvIbvpZmXlvIl7ST4j6RW4LkbHR Kv4qgl9PEzLI3gWmZ2A9gWnpSFSiBbKeXCHRMpJNHpFoOckgP5FoBfBflmgleZPTSbSKFPMKiVYT kS+XaA1/N79NorVkhyylqyPF8laJ1vO3ySck2kBGVSqJNhGDKi7RZqJTXSV9k05IheoGieaIUvWY RPNEoXpSogXiVJ2VaBnY+ZNEy8HOmxKtIAa1QqKV5Jg6XaJVxKq+UaLVxKT+rERruG3qhyRaSzxa xcI3+1btKonWc53aVomGvOtukmgTUS7QZqLQ3S7RacSguwu85WRqmiPdfRINOdKbJRpypHdKNORI 75FoyJE+LNGQI/3VEg050t8h0ZAj/WmJhhwZ7pdoyJHhWYmGHBnOSzTkyJzShRyZ/ybRkKO0HImG HKU/KNGQI6tLoiFHVq9Ep5Fs6xpyLxFJBSmHnyqgNpMwCUK1RUgM2hCJA68JqCiZYO8B4ISBGoen ZpFsIKPwI5LtwKPfeMdBi/ZCgCGQPgTvgyDZBHqjIDMAvDBpAznsBy8yXz2bcbG8uESjfpG/7yS3 g3kQk7ylOmVMa6lO6RKd5fbCbD0hwDisjtoTQUIEDMF6w+x75RDrDQI3zuJDv/8fA4ySA8CLLOhc fHTo34o59Wic2aLeiMQPvTDzgc7fzVYSZ9Gnc44D1yt5EFm0giD0JmE0zlZJpcv+DR86QTdICoET I0UgNcg82ch06SwXjyGNwBiMDzIf6BpizMMYo0JMlsZiCLj4rf0R6B2WIk9lJsFiHPghFn9cwSCL xzCzEpGsxlmEL0QA10vnxHook+qEzhVi65pkGYxJeQ2wig4vqooYKWGWxxhnlFkMQFyQn5plDOyM sihNSF6OA2eMzYo26TrjizygM06wtWCMUxFG3+lMEYiAyP6SZJhFIcz+roH+lUl8US2k9h7GDGcR me/j0rqwzgaY5AWPF6+IRu1SpoerPgD9srfsxJXM2hizcITFYVLa5YvjnapOOvthFtXAwk4JS9nG GWmuRbAxsbAa9HFYkqG796hkPQ6rwAwdWshSgNUI3U1jS9aVquAgeBJg8wel+ctYpOjf3dTD3vCC Nv0pYzW3tP7LWN2MgUwc1kozNMwsTYCFI8ClFodYvmgml1pN8Wk1Y+QOLNjrY7WLUTzCVh9j0Yqz PMdYXaK2yOJGayTEVhhmc+BeH2C6qUi3wEnQCacx6kYXjWB9DbI9e6FmDrO5gqymLjYv9qlsEOI8 yXbt4EIOBtk4rXJcQSruE2yl41Lk0VaIvdNKWr5uOo4VWwhaReycHYN1hRZq6K1ejb/F8ruP0QXr qVNDlPY9noPBJfvvrWu/cPIu9WvNogjQleBa8BRKnZ3RhRNtkO3pcba3A2+7UoxzYElMQ9I5vvw0 p1GllTfJNAfZ/qCrCS3YoZKjbI+9U4b+u/bFhT3hlf4WLiCdjGUsVxNwzrzt88C9YkV5eZW4ORyM RmKRobjYFIlORKKBeDgyXiZuGB0Vt4eHR+IxcXsoFooeCg2WNQVGwwPRcFscMLigVy9KfBEH6pnd pbwdoWgMzIoVZeXl0kgpjqTkwjExFI6PhKJiQIyGhsOxeCgaGhTj0cBgaCwQPSBG6Mii7tDFPRfD 4yKYEf3j4Tjod8cD8VBMDIwPesFAhE0QjEyOx6PhUKzsohY6J4OFgViROBgSN0YjkfgiDwPiWGQw FB0XY4HxmAhRCQ+JQ4Gx8OgR8TA4L8YmB+KjITEKEwyGx4djIvgDCxljDsC80XGIQxnERBwKBeKT UfAsGgqMimEWiliJGBsLQNyDgQmgqcrY5Gg8PAEmxyfHQlGQjIXizEBMnIhGwGPqMFgfHY0cFkcg XWJ4bCIQjLMo0OyBZ6AijobHYS6I2UB4mBnGieKhS+OgHD4QKkslcWVMHAuMHxGDk5By9JuGczx0 WIwGaFLCsGxQDIyJkxN0GrA4DJxY+CiIxyOwoEN0SQHxcCA6hnPRAAdHAlFwLBQtG4nHJ+q93sOH D5eNpeJfFoyMeeNHJiLD0cDEyBFvMD4UGY/HJFFKDwXAuQNUri8yCS4eESdjIXANskKHxQBEJBQd C8dp1geOMKdb/J0bYDTKOhCvwUmMzOGRcHBkkS5geDw4OjlICy4iDoZjE6MwAfV9IhoGgSBIhcbj ZWJq7sg4BLYwXCSGxgao0gVT4ynhi3rExGlpQJhiUINBzN/C7Kx4JVtrmAOFYZgFSohWZ5QW2mDk 8PhoJLB4UvA5gJ5CIhbKPDIZn5iMQxkfCgdDVGYkNDqxbEFLdvTSLfveh5X3Pqy892HlvQ8r731Y ee/DynsfVt77sPI/+2HlHW/0hX/Tf9vXKbUwwx9N5q53zvBHEC5N5moBDiMcSubWA0wixFEklsxd AxBN5q4FOIgwgRBJ5q4DGEcYQ4VRhAPJnA0A+xHCyZxGgJFkThPAMMIQQghhECGICgOoEEDox7F9 CHuT2S0AexB2I+xC2InQh9CLcAnCDgQ/Qg/CdoQuhG0IWxG2JLObATZjrxOhA6EdoQ1hE8JGhFaE FoTmpKMNoCnpaAdoRNiA4Es6OgAaENYnHZ0A6xDWIqxBqEfoRqhDm7UINWisGqEKYTXarESoQL1V COUIXoQyhFI0VoLqHtQrxrEihEKElShZgJCPCisQ3KiXh5IuBBHBiZCLkJPM2gKQjeBIZm0FyELI RMjAMTuCDZlWhHQEC46lIZiRacKeEcGATD2CDkGLoEFQJzO3AaiSmV0ASgQFghxBhiIC9ngEDoEw 4OYR5hDOMwXun9j7B8KbCG8gvI7wGsKryYxugFcQ/p7M6AH4G8JfEf6C8DKKvITwZ2TOIvwJ4Y8I f0CRFxF+j/A7HPstwm8Qfo3wKxT5L4RfIvMFhOcRnkN4NmnfAfALhJ8n7ZcA/Azhp8j8CcKPkXkO 4UcIzyD8EEV+gL3vY+9phKeQ+T2EswjfRfgOwrdR8lsI/4nMJxG+ifAEwpmkDc4l7htJWwPA4whf T9p2ATyG8CjC1xC+ivAIwlcQvox6pxFmkPkwwkMIDyI8gHA/QhLhFOol0JcvYe+LCPehyBcQphE+ j3AvwudQ7x5U+CwyP4PwaYRPIXwS4RMIdyPchXBn0joAcAfCx5PWIMDHktZBgI8mrSGA25PWIYDb ED6CcCvCLQg3I5xEuClpDQDciDY/jDY/hDZvQLgeTV+HCtciTKHkf6DINUmrH+AEGvsgGrsa4QMo +X60chWqvw/hOMKVCFcgHEO4HOEyhKNJK5zJ3BGc4VI0fRjhEM4wib7EEWI4XxTVDyJMIEQQxhHG EEYRDuBS9uN8YYSRpLUaYBhhKJl+FUAomU5rdzCZfiVAMJlO9QaQGUim+wD6kbkPmXuT6VcA7Emm vx9gdzL9aoBdSQtcwtzOpCUXoA+hN2nRAFyCsCNpgWue8yctcL9zPQjdCNuTFrjmua6kBS52bhvC 1mQa9XpLMq0VYDNCJzI7ENqR2YawCWFjMg3uTa4VRVqQ2YzQlDRvBGhMmumm3JA09wL4kuY+gIak eSfAeoR1STOt1rUIaxDqEeqSZg9AbdJcAlCTNNcBVCNUJc10otU4USVCRdJMI7gKoTxppoH0IpSh L6UIJeiSB10qRihClwoRVqITBQj5CCsQ3KiQh5IudElEJ5w4Xy5CDkpmIzhQPQshEyEDJe0INnTQ ipCOflpwojQEM+qZEIwIBgQ9iuiwp02a9gBokqa9AOqkaR+ACkGJoECQo6QMJQVk8ggcAvHNA86D 3BzgeWj/hPYPaG8C7w1QfB3o16C9Cu0VaH83Djj/Bu2vxqDzL8ZB58vQXoL2Z2izwP8TtD/C2B+g /yK030P7HbTfAv830H4N9K8A/wvaL0HuBeg/D+05aM9C+wW0n0P7mWHY+VPDiPMn0H4M7Ry0HwHv GcAfQvsBtO9D/2nAp6B9D9pZaN+F9h1o34b2LWj/qT/gfFI/6vymvtj5BOAZfYnzG8B7HOiv68ec vvnH9Pudj+rDzq/pR5xfhZFH9KucX4H2ZWindQedM7qo82FdzPmQLu58ENoD0O6HfhLwFMgkoH0J 2heh3QftC9CmoX0e2r3aK5yf0x513qM94vws4Ge0lzs/rT3m/BTwPwntE9DuhnYXtDuh3QHt49A+ Bu2j2lLn7dBu09zj/IjmM85bAW+BdjO0k9Bu0ow4b9Rc5fyw5mPOD2nucN6guct5PfCvg3a1kO/8 gFDrfD9X67zKf9z/vunj/iv9x/xXTB/za49x2mOOYx3HLjs2feznx3ybFZrL/Uf9l00f9R/xH/Zf On3Yf2h60i+bTJ+MTwp/n+SmJ7nmSa58kuPJpGlSnBR0cX/UH5uO+kl0W/R4NBGVrUlEX4jyJMpp ZuYfuz/qyG0F9F0e1ZtaD/oj/onpiH98aMy/H9wK1w77R6aH/UO1g/7Q9KA/WDvgD9T2+/fV7vHv nd7j3127079reqe/r7bXfwnI76jt8fune/zdtV3+7dNd/q21W/xbgL+5tsPfOd3hb6/d5G+b3uTf WNvqb4Elk2xTtpgtmKgDW7LBE+LgGssdPscLjpcdMuJIOB5zCGnGLGcWX2TM5Jq2ZnKRzCszP5Qp GDOeyuB9GUUlrUb7U/bn7S/ZZRafvaisldhMNtEmWOnabJt7Whk2NCOuqmJr3WxzF7QarZzR6rTy LU4rR8wvmF82C9ZHTU+ZeKORMxrnjbzPCOJGg9PA07d5g+AzrKppNeqdep6+zesFm08PHGpxpW5b T6tR69Ty/gbtVi3v0zY0tfq0peWtROBEjiOcCUBQgewDnNXZKjzC/u8kcsJxHz7V0+3xdMwo57d3 JNTbdiW4axL53fTd17UzobgmQfw7d/We4rgb+ugfLfYk0ul/ZMn6V19/Pclp7EjkdPcmhbvvzmns 60gcp7TPx+h5ShMQ6fPsjU3GPJ74XnjbG4t72C/0uEna81Am/Y3FoU9/JlmfeN7xhWIA+2Lwiku8 +Dsr/a99cf/TDvx//srYt5cQ8v8AgaRVNgplbmRzdHJlYW0KZW5kb2JqCgozNCAwIG9iago4NDc3 CmVuZG9iagoKMzUgMCBvYmoKPDwvVHlwZS9Gb250RGVzY3JpcHRvci9Gb250TmFtZS9FQUFBQUEr Q2FsaWJyaS1JdGFsaWMKL0ZsYWdzIDY4Ci9Gb250QkJveFstNDc2IC0xOTMgMTIxMyA5NTNdL0l0 YWxpY0FuZ2xlIC0zMAovQXNjZW50IDk1MgovRGVzY2VudCAtMjY4Ci9DYXBIZWlnaHQgOTUyCi9T dGVtViA4MAovRm9udEZpbGUyIDMzIDAgUj4+CmVuZG9iagoKMzYgMCBvYmoKPDwvTGVuZ3RoIDI4 NC9GaWx0ZXIvRmxhdGVEZWNvZGU+PgpzdHJlYW0KeJxdkctugzAQRff+Ci/TRcQrJI2EkFIIEos+ VNIPMPZALRVjGbPg7+ux01bqAnRm5l4zXEdVW7dK2ujNzLwDSwephIFlXg0H2sMoFUlSKiS398q/ +cQ0iZy32xYLU6uGuShI9O5mizUb3V3E3MMDiV6NACPVSHcfVefqbtX6CyZQlsakLKmAwZ3zzPQL myDyrn0r3Fjabe8sf4LbpoGmvk7CKnwWsGjGwTA1AiniuKRF05QElPg3S7Jg6Qf+yYyTJk4ax8e6 dJwGPiNngQ/IB8+p1+SeT75/DJoc+RT6KfJj6F+Rz4Ez5IvnvEF+CvoYuQqaBLkO37r65e9b4m9g zj/xUL4a46Lxl+EzwTSkgt/70rNGl3++AfOLidAKZW5kc3RyZWFtCmVuZG9iagoKMzcgMCBvYmoK PDwvVHlwZS9Gb250L1N1YnR5cGUvVHJ1ZVR5cGUvQmFzZUZvbnQvRUFBQUFBK0NhbGlicmktSXRh bGljCi9GaXJzdENoYXIgMAovTGFzdENoYXIgMTMKL1dpZHRoc1s1MDYgNzkxIDIyOSA1MTQgMzA2 IDMzNCA0NzcgMzQyIDUxNCA0MTYgNDk4IDUxNCA1MTQgMjUyIF0KL0ZvbnREZXNjcmlwdG9yIDM1 IDAgUgovVG9Vbmljb2RlIDM2IDAgUgo+PgplbmRvYmoKCjM4IDAgb2JqCjw8L0xlbmd0aCAzOSAw IFIvRmlsdGVyL0ZsYXRlRGVjb2RlL0xlbmd0aDEgMjI3MjA+PgpzdHJlYW0KeJztnHlAXNW9+M+9 d/Z9gFlggLnDsA/MsBPIwiQsYQlmIaOQlQEmAcPmLNk0GrVuaNRWa6tWY1urtVEzTKKOxmqqaWyr 0damrrVN27Su1GjVViPwvud+7xAS43t9ff293x9P4Dufc77ne7bv95xzzyVoOBgJEA3ZSTji7R3y jw77lvYQQp4jhEnq3RzmN5WcWArpY4RIP9owunGoo/4NhhD5UUKUyRsHt2342rXzw4QY3iIk9fn+ gL/v+UPX8oQUxqGNqn5QfDK5Ww75E5DP7h8Kb71NM9pJiCsF8oODI73+y7qaoX3XTsjPGfJvHT3E jGsgT+vzw/6hgOvk8S7Iv0KI5bujIyHoi4Oi+oto+WgwMOp67ykf5L9DiCkLdAx80y9og5HRPMtJ pDK5QqlSa7Q6vcGYlJxiMlusqWnk/8qX/E5Cpm6apbgUvr9D9pCHyGPkJ+QX5Nfkb4yKdJMryJPk T+Qd8iE5CW6UMyYmnSn4941j6nLpENFyB4mMWAiZ/mz67an7pt+GdaWbpbkJchZJ7inNdNL0xJm6 qZum4lPPy9TEINQ1sM+C9gQzMf0ZW0fz01U0z15F00KNE/I7p/ZO7T5tOKMkSCJkK9lGtpMLyQ5y MbmEXE6uJFeRq8k14ItLIH0tuY7sIteTG8iN5OvkG+QmcjP5JrmFfIt8m9xKbiO3gx/vIHeS3WIZ zd8J37cIpbTke+Qech+5H/h9cjf5AbmX/BDyPwLv308eBB1qMP8AaO4i3wXtPaClVlS3F76jZJzE yD6yH2KG+UQuTg6Sh8kjwEchmgfI4+TH5AmI40GI7FOCjmoS+S+3xM+nySHyU3KYPEN+Rn4OK+NZ OAeOkOfJC/9SyU9nNDT3S/Ir8iKstaPkN+Ql8jJ5lbxOfkd+T46RP8Kqe+8L5a+AxWtg84Zo9Qew +jN5GywnwBLt0Oa3QulbQgtHoe4xcpxRkI8Zlpwk05Ci0btFiNCtQhxp9Gh07hb8TOOxF/I0QvfO xOYB8PEDEE+ao+nbxGg8CLbj4MGE/87utefF6KC/Hwcb6gtackT0xTNiJGg7T8zUfVYoiwn1nppp 9ZRHcYa/meWd387y4Z/JXwTPoPew9JT3qMVxsKFepm2c7ts/Ql30Pq1L9bPr0LLXIP82nA7vgacp 3xUi8S55cyb9plg+Qf5K3icfC58nyAdwnvyNfAT5T0BzAnJf1J6p+Tt8/4N8Sj6DCH5OJmflJs8o mSRTEGPCMAzLcGTqVOqUVhAJI2VkcKYpGCWjYjSMltExesYAmtNL1DMlxi+UaM5SphQ0SUwykwLn pYWxMmmMDc7NDCaTsTMOJmtWWepMCQ8lTiabyRHLzELN1Jm6drCwzLItYEqYLfDpYtyMB9KlTAVT yVQzNaAphnwZ5GuhrETgIm/T+nVr16xe1dXpW9mxYvmypee0L2lrbWle3NTYUL9oobduwfx5c2tr 5lRXVXrcxUX5uTnZziy7NcVo0GvVKqVCLpNKOJYhRY3Opm4+mtsdleQ6m5uLad7pB4V/lqI7yoOq 6XSbKN8tmPGnW3rBcsMZll609M5YMgZ+HplXXMQ3OvnokQYnH2dWLe+E9K4GZxcfnRDS7UJakitk tJBxOKAG32jtb+CjTDffGG3a3D/W2N0A7Y2rVfXO+oCquIiMq9SQVEMqmu8cHWfyFzBCgs1vrB1n iUJLu41yOY3+vuiy5Z2NDTaHo0vQkXqhraisPioX2uIH6JjJtfx40cGx6+IG0tPt0vQ5+/xrOqOc HyqNcY1jY1dFja5ogbMhWrD9uBWmHIgWORsaoy4nNNa2YqYDJirNMTj5sY8JDN458d7pGr+okeUY PiY0Sac44yYoT6QJjA1GCPNzOOhYro17SQ9kojuXd2KeJz22GPF6XF1RtpuWHEyUmHy0ZGeiZKZ6 t9NBQ9XYLf5s7rdGd/bwxUXgfeEnB36gnI9yud09vf2U/sCYs6EB/bayM+ptgITXL861cbzEA/b+ bpjEAHXD8s6oxzkaTXEuQgNQ8DQGAx2dQhWxWjSlPkq6e8VaUU9jAx0X3zjW3YADpG05l3c+Ssqn j41X8LZ95aSCdNFxRM31EJTcxrHOvg1Re7etD9bnBr7T5oh6u8B9Xc7OQBeNktMQLTgG3TmEHoVa MLczrBPGdObyHAXfydq4LhotUPBN8OFcNA8KDBAuIUsjumge38nYSMIMehEtaOq0diDD5dQ30yKO Vq1vtjm6HPj1nwzJJo5JmhNVzGrLAIqZMWE/Xzo0tKYDKuAbAw2zBnhao1JxgGJrZx8nS30hdgw1 FDSczYkiLgd2LuhYaEZQ0Sha+ShZxnc6A84uJ6wh77JOOjfqayG+bR3OtuWrOoVoi6tk5Wk5LJ+D uShxQHEiw9bDGmxy2RJhFfKLhfxMtvmM4pZEMT+mcLZ1jNHGnWKDhIcdBJOW5bb4r52TVAFbswlO N2eT38kb+KYxf3x6Z8/YuNc7NtrY3V9L23C29I05Ozrn2YSxrujcYdtOu0oibUzbykXFRXD2LBp3 MlcvH/cyV3es6nwU7q381Ss7YyzD1ncv6hrPhrLOR3lCvIKWpVqqpBmeZmhLKyCjEOxtj3oJ2SmU SgSFkO+NM0TQKRI6hvTGWdQZEjoWdBLUeQUd/YIgWfvBxXDcNvJ9NDwXdfWPdXfRzUXMEEr4YaKM cwGJss4F4wwr00RVzsCiqNq5iOrrqL4O9TKql8PCgOceOIeeSWPdTjinYEF1EhuDS5GjTfLx6emV nY4jtokuByy1NSCrOqNKF5z90pxWsFtMpRvUi6M7e/10HMTXSevKc1p6u2DZJhoEk5aoElpQii2A RZNQhy5HqNQLsYEACvV3Qia6syva5aKddg50CcvZECXNzloIO7YpzaUdebrGkpxlwt6EraDKuYpC CWMjHZ2osUEWOutCJ8k1MPJeJxT1dvPgbQnp7YCljmepyoaaAByJktyAICqbWEjotLgctVYVVbqh QfihabWbbklpjryrCwcv5K4SDaBvQ1QNI8qd5UqxAngHilroWODnKhgqNf0JbWZ5nKxwboWThQ5a aEkOxVFtTosfDn+srwaNc06isoKeEWqxjUOoldOZa8DvXM7K+PS9zm2OWV/FRU76cKALk9gehYVN usbOVERXu4qLFGdqtYJ6bEyhPXsF9JdCO0NQwmsmvIGGuNfhjZEjclJD2sk5ZOXjRMvcAa+Vtcyz +xsaFMXyJyDLEp55lijg+niHN1nCam22Omel7DpuubGlTn4du5LUTf7ujcPwcSSpxnOE8bwx8dKE YfKwscYzcXSitIQxOoyCpOhYuVwmc2a52cq83Kry8rIFbGVFrjNLxwq6iqrqBVx5WSbLpSQ0C1ia Z7jXP1/KNU5ms9sccztKpYwrx2JPVig4e6Y2p5zXt7U7q/LTpBKFjJMq5HlVi5y+La1Zz6useekZ eVYVMCMdOPmUVPfZh1LdyfMkDScfZ9+q6VyQLdumVbNSpeKO/ExTdmn6/DatXivV2Sxp6XKFUacq bPZP3pqWY1GpLDlp6Tm0rZzJuXBM1E2/zV0nzSEt5ABs4nofPGUXst96KLssu0xji7PfiRGN+wBT TKqJiil+2FgN3+Z58em39qv1zJJ5cabYq1lokxZ0mK1UYY4znTGv9FxirUtrn3DVJdXUTLiMNTUe j2HCMMF4jk4k0Qw40/Y4cf+bmi0t6WLQ0ZKEozEibpmYl5nEQNAQmVIyZdx1Sy57sLc+1Dk3TS1R 6lW68mUjLSVLKtNL2nv6e9pLGiO7u9xrli1IkUtZTq5Vq0ua1lS7vC6TZ2lff985JczXNty2scJs z0orddsL09SOfIelcEFuUV2pq2S+L7x87a61bp01M0VncaZl5Kdp0h02U05FhgvLQ+D3XdOfSRZI U0g56UG/P0lK2JuJg6jYm4iJZLGH9xUVmZRx9jmvzktMeSscKoNthcEn9YET6qgH6sCd4A1YqrBI y6hLveqzmYF7coTp5+bmMcbEijSKnjIZGVyrphSZnGHMZskCtb2qYGFNqnxqmyazLC+3PFOrzSzP zSvL1DAXylP4srz8CrsmKXXqDuZyszJPbVTLVNDqhsnb1Aa1TAYfkqfVBpVMpjKoJ19lc7VGlQS0 KmN23pRn8pECG+zDDsbC3cOdT7TERjL2E53cpH4cXu4kxAifVuKBjcd4yo+UwaZLkTkdWbmwp8od ZWbjaTnuHotpUmqymCzsSfiYlebUlUVFleVFRVM1VYWFVeXFcE6wZBchkoNwTiQRO7kg4fVk9nY4 ZtLYbxAlsYoL0Bpn3F6lbrlNWH22OLMSVt/KxOqbcDF4IMDZ9M/WSKxROnijMPpMVloBC9JppGtV cnDtg5/eP/Wso7jYwSx54IMfnDt1wrX+m9uuuGbw5t5S9rbY5F1teUWS/qK85bvf+f6aO8MLP79x zgU/hHW0aWo3my69kThJ1pMkjfkM5mJgPiUywrHhfSa7+gpS52E8ky+BR8GZMghzksWcgodXnpuj G0Q8rFjLynPPWyEzF+en59v0XNWyyjRb1dJKVmMt4LPdVk7a+fSU/7XXp3oPGywGhUSulvf/+uXX Lxh9/eWjA3BacXKdGde17ALw8TzyKnrYq9aWlFg8HpXbak2Ls337s0s1GhUkHiHZVctTNWorPV68 xD19Yr/ByS4pjU+f8PI0ZTHQTy1+WjwlpW6ZPX+53ZeUWN11SZYaF3yBj8vKyuh+mCgzlhvoh7Fm vqe83FgOQXro39vLzHaim8jJ6DjBk4xz1s6iT4RM1sKU060l7C3ZBeqMkpzsknQNO3WNJMlekpVV Yk/ipm5h1Zke0Geoq4rvdy8q4TWMVcJkae0Fc3LGbXmp2mwV7iSVJOPkcdhJnJTur/STf5rRX1pe pXfWFH4+yTGFtdl6HdSivy23TP2WuYdxwP4yjRtInL1+X5Lakk4MR+G5NnG4tCRHeIjh6KuTZx5Q 9yiS0k1Xyo3WrLSMbAMj3W7IqsiBa5A+nr+wtjrjoEqnkEphYzMpd2YVmuVycyHsqz6uluuWboW+ qknWuA16WxtTV+cBHkqXZZnLFpM62MvQ75GXJqFrxjTrsVgtrj75WbRmc4qc/YveXso73JlaucGW fIlcb+GtabyekU6xenuJw+HO0NGCnbTAksYbJJL5t2cV0JEVZOV6a6vSn1TplBIJeG3q/S8pgBm0 wlPwYe5l4iIVjBTX7b7kZEdRnK2PuSokcTboVTm4ouQi1lb0tCQ+fdBr0TLtRGKQsEuWSbol7F2S qISVSNI98Azbp2faKb082HiO57ZaPyE6g441cjqlVcO0K61goPzUm94unA+TLtfR9evWTsAp4YLD fO0F69a6JtatBXeVvUFPd3qwK/93+xZOKzhpU/BJio8HnWz2VcaUVyVceOTcwwXZk3+wzV27cFFf S4leqVFwrEShrV0VXrRl39a5Czbfd/7o7g0lH3Gr15cs9qSyzGfuopq1C7OSLcnyJEeq2W7W66wW 47ztj+3Y8uQVTYsid63jz9+WPb/DA6v44unP2AMQl3IyilEZz9XH2W6vhqTpVXaVR8VpORWdtppp V8WZDq/K62rN1Zv4FtMSunnp84+ewOupPw9N1Ahntuq/tj91pRCuDPIvPjZNwrKVsQdgripFSmpm kqmwuDRdA9s8pyRDo0orsPOFFpVzwZw56dpM3qqWSliuLdudppIr5MbseUWTR1V6FewlvYq7KJEa KVuYq+fkSpXGRPfVxdNvS4h0lBSS+eQWnH9M5ag6wHbDDcHFfs2rJCZVVaVDIi1JXJ1K4kybV5vb amsyLKkRnkU1caYVnkXtiWdRHcQaTjR6WMJJST3y8L/axiwv5Zm+eLHAjS085QQfms3Co45U9Nyw uvicxY3Z6tTCTHtBqirhtKyGhub83rHz8qdOGgvry1NLyqsyK/2VpQ3FKcx7W564stmYW1vgV+tV EolKr5Y6VXjfUE0lw2mqW3rlvkjN+StKdVlV+VOvNCwuW7YBfNgMPpTACkomeaee+SlsBJ6TmfCp IqnirFPjTJpXqW91CjN2xpn02TOe/cz/Z2vMck9i1yScYRJ3lkQyb3v8wi3R8Jz52x+5cGs0NGdq 0lTWUTdnZZXNXLpyQc3KqjTm7eDjV7cuuji+Ofjjq1oXXhy/dNHICnfB0pHFwOKCc0Zgn7RMv81+ CLNsIW+fusW7Z27xDd4sopG4GffxarWKUb1prPbCBjBU89UsJ1y99fOYefSJaBOu38fp1bvVbNBo mSXEzBgk5g9nJgY7Sbx/rxUu4OvXugwTa+HntMu9l/9/3Nu/cOdnP6zpv76jbHVziVkjUWiUapfX V5VVmZeSM799efv8nLJ1V60sXOotSlZIOE6uUShza9pKssp4Q+6CpcuXLshlMpeEz8nTW6ym4qIM p0mempmmS8tPy3Tx6VlF3lV13k1LCjVJJr3eZLfYslLkJqtJl+ZMsRfy6Y4ibxdEaS5zAzuHXUv0 xBgjcvWj8HiWwF2X8Rw5454Ll7E5ZutUd6rZnMrcpTFqpMzfa92emjlulTWfPt2bp9/hHNxLpJLc IZ4L6STvCTZMdMTK2OF6my2u0ew4Y48lt0oeY5pJKTw61BCJ0iJhwRbFmaaYV5l4EBx1TdTBJ4xm 4lCZ8H72P20Jb0qznxoQExmGRGaa9WyBqUjl1trW89wbdw9W12+9uye/vb7SrJRyKQZjbkVzWU9/ Wnl7eUXbnFytUiOXRNOcVr3FkWbw7tgfvvLpnQvgVcustzpTaz3pWenf+kbzcGuOPdeushUKN6Hp 99jrJeOkltyEvnrEaNTOLSDO4ji7PmbRFifOvWKY4T5nc4Y2odDSKVuaS+PM4phXLi5J2NhHhOOg fLLsUJkRH9CPkuJ/pRH0kES8S9Log0PKZx6xJvG9DF4SUswJZ7HXq5Ocnur0tuHmrE3JKfSxcb46 ozQ7Bx4+T9EHSUry0+65KXyqUS5Ty6TbizzJerUsd+nWFczPPdUZ+RbVM3BuSuEeCQlLfka1Z2pt S4tcKZebssFb5qlvcvu5wySb9KO3HmaUSh1Jg1Nk0cPe7DRelWaNsyGv3qtLs7ekqpJbVG2SpaRN vCrDW2mqx4rvpWmGN9IA9Hcp9EjQnNUcXODg8HWkOpm+oeZWVFUJ1+fk8mQ6Y+ESePmgcll7fomV lW/RmqRTR7TWGo+rLF0nf5E7KEsuqnbV2BRTh1LNcoPVyLhkqTquwpljUnCaVMvkHtafZlQozDmp 9B3QBndVi3hXzR4vctO7ajq9IK99OCldnWcpT1xWX5o4bBBuq2e9KEvPquUsckNaSuIuqmckm/WZ pcIddWorvaVeKjNY7JY0B1xfN+vsWCI1nHYb1TCGs19TZxXQVe2H974keO9zkOxHYV6hWJJJegBe //TEzpzcZ0pT4Ysf/Q2W8OZHx0uHm5tbUV2VVFnB5uXimC3mJDYprXJpFae35acXFJtlHeed65Ny qcU59vw0Ndc/yKZd8PrLv+6HFz6JAl79DjG7X3+N2f201qyDlz6F9MWpDvBr1dQx5jFuLzxnPcQx np1ygL2G5MLA4g+p7K4SqZ54juC7x9EJ0amzn4vGs7+NPKZzVBUUVDk0GqROYUxPuQzeThyp6Vnw dhIxOCqcWaW8njMXzsnW67PnFLpqsw2G7NrJLcL7yuOJOz5jmHlfYcjS6XclGqmT1JBrZ05P1xPs M8Lp6QeX5op7NzfOdMeSOyRxZtUjlSVW8XLUA2fduYmzbubQPHpIPDT/pQZOOyuThB1w5gNs5riU aDiZyly3OtJwxUu3LOu8840rqvp8DTaVjJPAdPXulkBT+zZfkee8C9ubNrR4tCqNQnIo1ZmaZMl2 mFd8/6Pv/YAhD65Kysi1JaXnpmcWpmmcLmdd5J7+4L2DlY58XmF1Ebx7cy/C7bOMDCXuFGo4MMsK U+AGvi+zMNWQOOIMcabdq/QWt2Y3pS6RLsG9TS9BSTX0LUa4NsX+OfvT3q6Fy5LsLJdvIx4RJu5F TTqefMnZNbklPZWJm2SCC69qWb2jPSsrccFmJhe2VmY01U/uTWhm3yK9dfP6r+1NzP0AzL3i1G/n SuHWDbco+DQTJ/vwvuJisyrOPkJ/O2fOUkvzW9KbjDOTgSvLab+dOy78du5sZrMm/M/8do47oM4o yy8odyTJp14+c7aMQpHiKM3NKbdr9Pqpk4xbo3ao9EqpRGnQMi9N5X9xzp9/wPRqkgStWp+VPPXK VHFKBsx/G7wPPwdPgTKySXzvUucdYNfBpdnOrvfqk4tb8tTS1JZsq3gzaN/v1bXjrBK/oK0T9jp9 v/Dq/hnz2ZGvFH+5fsoNwvKvqp5RcM+pUgvsjgKrqvXWFWt2tDuERyD4ICkHFoK/Wi24JF0zE/mN y5bN679mAzujmFI0CUuBXZ7QYNxlSrhHLyN7xHk3JcMtYV9mZpmK3haWLcijb19lxDBrKcfaWrPj p25J7bAkFrYuaCqe01K8JHV2rGe9StQcnaD/rIAvpP+jxv6LHfNlW8gkPmXF5SVTatJLcnJLMtRG Z2VO8ZoqWFOCP41ZVdnuNZVnvte23rSsurOxzJjf3taW17W9jZ/xK2ssPmOLfVFz6o0XwmJxzctx LchLnrdxrJ3+/R79O0+WkPJy++fr9fM+JqkK4c/6Drx70XOz/uYwJLsAnn4MUVJbItaT3zn52qy/ BGTO+ENFm0RHCPczUifpJruYFyD9OemQqsku9mKySTYFNBALdyHpk8hJK/sKuVjSBtJOmrlS0sKG yFyuizSzh8HmF8QMdjao52crSZVkHllKbWHl0DrbZJeRi6G/O5lVzKvMq2wKezF7mBvjDkockgnp oEwnu1S+Ur5NoVfcozQrtyufUllVe9QF6vs0OZo3hVHbyHL671qCJwzwRO0mRJPOLITXBVpayV5G 8M9lCekTPjlhtjohxwm1FCQopjm4+W4X0xJiJfvFtBTSz4lpGaSPi2k5+Yx8LqYVpJD5REwrCc/y YlrF3sV6xbSanCs5JqY1pFBaJaa17Lel3WJaRwbl/xDTBqJTJMZpJApFcCZOZYrtYpohcsX9Ypol EsU+Mc2RDMUjYlpCNIqjYloK6WNiWgbp98S0nOxQ/ENMK4hJGRHTSmJQjolpFbNMeZuYVhOX6p2Z vzQ2qU1iGl5N1YViGtaQZrOYNhD5TNpIJJodYjqJaDSXwmgZiZLGSLNLTEOMNB+IaYiRZlpMQ4y0 cjENMdIuFNMQI22XmIYYaS8Q0xAj7TViGmKku1xMQ4x0D4hpiJHuiJiGGBmHxDTEyPi0mIYYGSfE NMQo5QoxDTFKSYwTYpSSGGcSsZrk5D7CwxFYAt+VkGonA6QXVtsICYFsIGHQ1UMqSEaFTz9oBiA1 TNxQspAMwjdPVoBuI+mHspCQCwADYL0ZPvvAsh7qDYJND+gGyCKoPwj6L/ZVK/Q225afsa6dNc4v szlX6DUkjpDau4Uap9sXz9if2c6AMP4AMAyzoW3xYMEDAzA/WhoWtHRWPKSpP/ogNwQMkk2gG5mp c/bSDf8tH9MRDQtt0dHwxAe5AWEMtP8OSPmFXEjocxi0HnEEI7Nm0Au5CJSGhVlSa/d/YwxLoG4v yQdNiBSAVZ8wksVCXdrL2X1IPTAE5X3CGOgcQsIIQ0IqINhSX2wA7RCkB8k2yG0RPU9tItBiGPQB wf84gz7BHxuFVkbEVsOCh095AOdL+8S1QNdpizC/DaCh84oIEQyJcfULK3hA8OWg4JUQKRJaHhI0 g0KLfvAL6hO9DEE7g4KXRsVRDoNmSOgV26TzDM8aAe1xVJgL+jjhYRw77WkEPMDD/HE30VENga0f +g/PWguJvYY+w154YezD4rxwnfUIlqdGPHtG1GtbhXo4602Qd39h9+UJrQ0JLWwT/BARd/VsfydW J+19i+BV/8xOGRCjjT3SWPPQxujMbHCMG0UbunO3i62HYRYYoc0zUfILa4TupqHT5pVYwb0wEr/Q f6/Yv1vwVBh6rIW94YHa9NstrLnT179bWDdDYBOGudIIbRRaGoUWtoGWtrhBiBeN5OmtJvR0NaPn Ns201yWsXfTiNmH2IcFbYSHOIWFdYm1e8BtdIwFhhgNCH7jXe4S6CU83wkmwBE5frBucVYLrq0/Y s6fWzBahr15hTZ2tX8xT217wc0TYtX0zMegTyukqxxkk/D4qzHRY9Dy2FRA+6Uo6c960HFdsPtQq EM7ZIZhXYGYNfXFUw19o+Z/30anWE6cGL+57PAd7T9t/X5z7qZP39HHNneUBOhOcC55CibMzOHOi 9Ql7eljY2/4vnSn62X+aTwPiOX7maU69SldeRKjZJ+wPOpvATDvUclDYY/9ZhP5d++LUnvAIo6F7 AE9GtxCrUThnzvr8v48vKymp5NsHeoMjoZENYb5+JDg6EvSHB0aG3fzCwUF+xcDG/nCIXxEIBYKb A33uev/gQE9wYNHIYN9MrVpe1PJUXSu0OVtzbiAYggb5MndJiagvpvqEzUCIDwyE+wNB3s8HAxsH QuFAMNDHh4P+vsCQP7iJH6Els7Ibzj5ifmCYh2Z43/BAGOp3hP3hQIj3D/d5oIERoYPekchwODgQ CLnP2sKSSG++P1TA9wX4xcGRkfCsEfr5oZG+QHCYD/mHQzx4Y2ADv8E/NDC4jd8Cg+dDkZ7wYIAP Qgd9A8MbQzyMByYyJAwA+g0OgxfcfEuY3xDwhyNBGFkw4B/kB8LQR2+oiA8N+cHfvf5RSNMqQ5HB 8MAoNDkcGQoEwTIUCAsNhPjR4AiMmA4YWh8cHNnC90OY+IGhUX9vWPACjRqMDKrwgwPD0Bf4rGdg o9AwdhQObA1D5YFNAXcifHkhfsg/vI3vjUCocdzUncOBLXzQT4MyANOGiv4hPjJKu4EWN4ImNLAd zMMjMKHNdEp+fos/OIR9UQf39vuDMLBA0N0fDo/WejxbtmxxDyX87+4dGfKEt42ObAz6R/u3eXrD G0aGwyHRlKY3+GFwm6hd10gEhriNj4QCMDSICi3m/eCRQHBoIEyj3rNNGHSjb8lCKA0KGfBXXwQ9 s6V/oLd/Vl3gwHDvYKSPLrgRvm8gNDoIHdCxjwYHwKAXrALDYTef6HtkGBybP1DAB4Z6aKVTTQ0n jM86IsGcLg1wUwjWYC/Gb6Z3YfGKbc0VBpA/AL3AEqKrM0gXWt/IluHBEf/sTmHMfhwpBGJmmY9E wqORMCzjzQO9AWrTHxgcPWNCs3by7M361cvIVy8jX72MfPUy8tXLyFcvI1+9jHz1MvK/9zLypU/x md/Nf+nXuJKLs5/GMjPscfYfsUwX4O+xzCLAJ4iPER9h2d8w9yHiA8QJxPuIv6LlBOI9VL6LeAfx NuItxJuIvyD+jDgey1QC/oS5PyL+EMtIAhyLZaQCfh/L8AB+h3gD8VvE62jyGuZeRbyCeBnxEuI3 iKOIXyNeRPwK8UvEC4jncRBHEM8hnkX8Arv9OVr+DPEM4jDip4hDiKcRTyF+gjiIeBLbfALxY1Q+ jjiAeAzxKCKOeATxMOIhxH7EPkQMMR5LLwNEEXtj6eWABxEPIO5H7EH8KJZeCrgP8UOsdy/iHsQP EHcjvo/4Hlb/LuIuxG7EnYg7EN/Bpm9H3IbVb0V8G/EtxC2Ib2K9mxE3Ib6B+DriRsQNiOux6V1Y /TrEtYgxxDWIq7HCVYgrEVcgvoa4HHFZzFYBuBSxE3EJ4mLEDsRFiAsR2xHbEFsRWxCbERFEGBFC BBEXIEYRI7G0SsAwYggxiNiEOB8xgOhHbERsQAQQfYheRA/Cj+hGrEesQ6xFrEGsRqxCdMVSqwGd iPMQ5yJ8iJWIDsQKxHLEMsRSxDmIdsQSRBuiFdGCaEYsRjQhGhENiHrEIsRChBdRh1iAmI+Yh5iL qEXUxKw1gDmIakQVohJRgShHlCFKESUID8KNKEYUIVyIQkQBIh+Rh8hF5MQscwHZCGfMQldyVsxS C3CgkkfYEZmIDEQ6woZIQ6QirAgLwowwYQ8p2EMyKpMQRoQBoUfoEFqEBqFGqBBKbFOBkKNShpAi JAgOwSIYBBHATCOmEJOIzxEnEZ8hPkX8A/F3oVvmE2FGzMeo/AjxN8SHiA8QJxDvI/6KmEC8h3gX 8Q7ibcRbiDexv7/EzE7AnxHHY2ZYWcyfEH+MmecA/oA4FjPXA34fMzcAfod4A/HbmLkR8HrM3AR4 DfEq4hVs+mXES9jYb7Cxo4hfI17Exn6F9X6JeAHxPOII4jnEs1jvF9j0zxE/w8E/gziM/f00Zl4E OIQVnsaOnsJR/wQbO4h4EvEE4seIxxEHEI9h049i03Fs+hFs+mHEQ4j92NE+RAwxjt1GEXsRD2LT DyDuR+xB/AhxX8wEBy7zw5hpIeBexD0xUzvgBzHTOYC7Y6algO/HTCsA34uZvIDvosldaLIbTe5E kzuw7DtoeTvmbkPLWxHfxgrfQtwSMy0DfBOr34y4CfENHNLX0fJGtLwBcX3MtBywCy2vQ1yLGIul dAKuiaV0Aa6OpawBXBVLWQu4MpbSCrgilrIa8DUsuxwtL0OTS717gSf0jfb3dc32Y5pz7E+B/ATk IMiT6nPtMZBxkCjIXpAHQR4AuR9kD8iPQO4D+SHIvSD3gPwA5G6Q74N8D+S7IHeB7Aa5U9Vvvw3k VpBvg3wL5BaQb4LcDHITyDdAvg5yo7LffgPI9SC7QK4DiTOXxJLp7rs4lkRXUhgRihnpSgoiLkCM IkYQw4ghxCBiE+J8xDzE3JiBohZRg5iDqEZUISoRFYhyRFlMT5dlKaIEkYQwIgwIPUKH0MYgBnFG g1AjVAglQoGQx7Q0sjLvauBfQSZA3gN5F+QdkLcher8H+R3IGyC/BXkd5DWQVyEKr4C8DPIEyI9B Hgc5APIYyB3g+e+oqKd3oqe3x4x0hW9D52xFbEFsRkQQ9YhF6IeFCC+iDrEAMR+nbEKkIJIRF2K3 HRjZFdj7csQyxFLEOYh2xBJEG6IV0YJoRixGNCEaEQ2ILIQDB8gj7IhMRAYiHWFDpCFSEVacgwVh 9t4OnAT5HOQkyGcgn0IQ/wHyd5BPQD4G+QjkbxC5D0E+AHkT5C8gfwY5DvInkD+C/AEieATkOZBn QX4B8nOQn4E8A3IY5Kcgh0CeBomDPAJRfRjkIZD9IPtAbhcivAN9fBFiIGZ0A/oRG9EfGxABRB+i F9GD8CO6EesR6xBrEWsQqxGrEF2ITsR5iHMRPsRKhAfhRh8XI4oQLkQhogCRj8hD5CJyMCjZCCdC ipAgOASLYHC7Ee/3gNMgUyBvgUdfAvkNyFGQX4O8CPIrkF+CvADyPHj4UZAruBz71zi3/XLGbb+s eafv0j07fZc07/BdvGeHT71j7o62HZx6hw1w4Y49O17fIbuoebvvwj3bfZLtKdtZ1bbmLb6te7b4 1FsYzebmiG9l5HjkowiXElkZ6YuEIzdHjoJCfndkf+RQhKP/cWdSZM7cpp2RGyNsCpSzJMLoqdoR Ueuaws1BX2hP0CcJVgTZuceDzC+DDMsHGW9wWZAFq33B7Pwmaj0dNKc1kSAfLAlyFzSP+Eb3jPiG m4d8Lwwxm2Aq58OUBtwbff17Nvo2uPt8gT19vl53j8/v7vatd6/1rduz1rfGvcq3es8qX5e703ce 2J/rXunz7Vnp63Av963Ys9y31H2O7xzQt7vbfEv2tPla3c2+lj3NvmXNzGJ3k6+Rq7ITO0My4Wc0 c2fmiUyJujtjNIMdzTiWcSKDG00/kc5eYmP0aZek3ZDG6eGDxY9Ue+oNqbtT96ZK9UKC04wm7Uxi R407jWyJ0Wv8pfGYUUKMdxlZ/Q363fq9em6pfr3+ff20XrJXz+zVPal7Qeft5pbq1utGdJxeRzWc watzlzbptXatR8vN82jrtEu13A1axqt1lzV5tdl5TXWapZr1Gm63hvFqcgua3ldNq1ivCgreV04r 2WklQziGZxjCGACcAry8nzHZm7jHhf8zqZQwzI3jKztcrra4fHpFW1S5bHWUuTqa00E/vctXRWVX R4lv1erOcYa5vov+cfHKaAr9nw8J+St27SIZi9qiGR2dMe6uuzIWdbVFd9K01yukp2magEmXa10o EgqFXSEXfICsC4EmHIEfAQx8AiNhWhIOETBxfclXCCUUWR+BupBZFwrRViMumqNCe/i/+8X8/x7A /9kv6/p1hPwHV+jASgplbmRzdHJlYW0KZW5kb2JqCgozOSAwIG9iagoxMDYyNwplbmRvYmoKCjQw IDAgb2JqCjw8L1R5cGUvRm9udERlc2NyaXB0b3IvRm9udE5hbWUvRkFBQUFBK0NhbGlicmktQm9s ZAovRmxhZ3MgNAovRm9udEJCb3hbLTQ5MiAtMTkzIDEyMzkgOTUzXS9JdGFsaWNBbmdsZSAwCi9B c2NlbnQgOTUyCi9EZXNjZW50IC0yNjgKL0NhcEhlaWdodCA5NTIKL1N0ZW1WIDgwCi9Gb250Rmls ZTIgMzggMCBSPj4KZW5kb2JqCgo0MSAwIG9iago8PC9MZW5ndGggMzQ1L0ZpbHRlci9GbGF0ZURl Y29kZT4+CnN0cmVhbQp4nF2Sz26DMAzG7zxFjt2hIgEKrYSQWhgSh/3Ruj0ADaZDGiEK9MDbL7a7 TdoB9LPz2fkSJyybqjHDEr66SZ9hEf1gOgfzdHMaxAWugwlUJLpBL/eI/npsbRD62vM6LzA2pp/y PAjf/Nq8uFVsjt10gYcgfHEduMFcxeajPPv4fLP2C0Ywi5BBUYgOet/nqbXP7QghVW2bzi8Py7r1 JX+C99WCiChWbEVPHcy21eBac4Ugl7IQeV0XAZju31qUcsml15+t81LlpVLu4sJzRJyckGPiSCIn xHGFvGNNgpyyZo+ccZ40e86XyAfPkVTER8qnO+QTcRYhl5wnrphr5EfWkLeae2J/JVmDecX+M/Sj 2H92QI55X9Kz/4jy7D/GMyr2n1Af9p/iWRT7T0lz4J7oWd39V3Sx9xvEK8Y38DM6oW/O+bHRQ6F5 4aQGA79vyU4Wq+j7BhbGqDQKZW5kc3RyZWFtCmVuZG9iagoKNDIgMCBvYmoKPDwvVHlwZS9Gb250 L1N1YnR5cGUvVHJ1ZVR5cGUvQmFzZUZvbnQvRkFBQUFBK0NhbGlicmktQm9sZAovRmlyc3RDaGFy IDAKL0xhc3RDaGFyIDI2Ci9XaWR0aHNbNTA2IDQ3MiA1NDYgMjI2IDQ5OCA2MzAgMzExIDg3NCAy NTcgNDM1IDUwMyAzNTUgNTM2IDUzNyAzOTggMzA2CjQxOCAzNDYgNDczIDQzNSAzMTEgMjc1IDUy OSA1MzYgNDc5IDUzNiA4MTMgXQovRm9udERlc2NyaXB0b3IgNDAgMCBSCi9Ub1VuaWNvZGUgNDEg MCBSCj4+CmVuZG9iagoKNDMgMCBvYmoKPDwvTGVuZ3RoIDQ0IDAgUi9GaWx0ZXIvRmxhdGVEZWNv ZGUvTGVuZ3RoMSAyNzE3Mj4+CnN0cmVhbQp4nO28eXxURbY4XlV3v713ek9C304nnZAmC0mHEIjk BkIEIhDZTNBIAgkQRBJCwuIoxI3VBfd9iI6ioiNNIhhQxriOOm+eOPocdZ4jb0RRx4x8ZxAZJd3f U9UdCI4zn/d+3/9+n9dJ1TlVdU4tp06dU1X3dnd2dLUgI+pGHNKXXNnUnqPENYTQvyGE7UvWdmro kfFjAD+KkLRwafuyK1NbzQsQUsoREq5YtnLD0i1fPrQCIfMTCBWvXd7S1Pz2lvvNCE2mdYxbDhkb Y9dJkK6DdObyKzvXT/Z+dhjSV0P6hZVtS5rumzI9B6Ep4yHdeGXT+vYc7h4R0vdCWlvVdGVLzoru 1yDdj5BjfXvbmk4EHUWorpCWt3e0tI9/Y9JHkK5FyPY7yMPwRz9GQEWaJhwviJKsqAajyWyx2uwp DqfL7fH6UtPSR/m1QEYwMyuUnTM6NzwmLx/9/+8jHEJpLDyO0vgQSkMofmw4xFrjx2gZheQrEFZ6 IiQ/vehp9HucgzXUh79HbnQae/FYNB3x6DuYhL1oCN2FHGgeuhvbUSZyofloOuaBJoxuwg/E18a/ RBeg29Ej8efwdfE9UH4reh2dhh78kceoFM0C+vmoBX3JfYbq4/cjGW1BBjQRzcEu1ITeh79voQ93 oDvRr/DV8dPQqgNdB/WVo0pUGX8pfgblopv4ncIHyn50G3oei/El8VY0CmWg7SQcfz/+CQqhevQL 9DT0KYwH+GkogK5AN6J7sZd7HbC70KMoho2kgZsivAgtTUcL0Cq0Dm1He9Bb2I5rhQ+EE/GfxY8j EaWgHOhTK/oSl+CZ5DHeGJ8U/whdig6iN2C89G+Av5R/XLg0VhF/KP4ycqLnsIpfwC8JRcItQ9fG H44/AxoZQmNBIrOgncXoevQSehP9H/RXsim+CU1Dc6Hl13A61nAIJP4+8ZKNZCP3LsqH0TZAb7vQ LhSFGTmEnkeHQTZ/QEfRZ9iBU/EMvBjfhv9KjKSZvM09wD3Lvcdj/kmQdxBlgYw60WPoAKzn36K3 sQD1F+JavAK34XvwQ/goiZKvyXe8zF/P/8APCaHY0dgP8Vnxb5EH+dBF6Cq0CWT7C9SHnkX/jv4D /RX9DZ3CVjweL8cP4yg+ir8mCskgs0k7uZs8Rn7JzeJu417iS/jJ/BX8b/mPhM3CDqlJip3ZHbsj 9svYO/Hn4u+A7pih/hCqBoleC1rxGHoRvQu1f4g+Rn+i+gP1T8QL8eXQyhq8Fd+Jf4lfw+/gr2CU iP1lkImkClptIx0gp+vIHeROaP1t+DtCPiIfkz+TbzmBy+DGcau5h7ko188d4T7nrXyIz+fH8rP5 hXwcZqZIuFCYKzwhPCW8LJwQy8VmsV38QrpOukH+t6HcoT/GUGx5LBrrA92VQZOuAkn8HD0Cev8s zMFbINF/hx4fRSdhFnw4gLOh32W4GtfgmfgSfBluwdfhLfh2fC9+AD+Cn4ERwBiIBH0Pk0oylzSR FnID2UJuJs/C3yHyJnmffEAGoeduLsiFubHcdG4hdym3CsbQyW3kbgDJ3sbt4d7m3uWOc19wgzBr bn4U38Vfxd/HP84/y78jXCRcCX+PCC8KA8I7whnhjEhEn5gmFogrxCfEP0miNE6qlbZJ70l/k9tx Gs6FnmsjrQXxwhocRfYQB78JD0JGOuaRBUYehnmYC6vib6iCi8G8mGk59M1JvHwK5RR1Pgr8nfh5 VIJfQ5tEwoEl5o+iXvyf5Cj/CrkA/QduxF7+cW6V8BYJoKfAGu0kL5Dn8WT0LCknC8iDHMKf4SfQ Z6Dv69Gd+Aq8Bj2FB/EEfA0uxZvQe8TFzcU3oPL4I4THCp6OTyDoAbqWb0aX/2sriMvQf6IvYz/n TfzVYJ/60d0wo0+jT/CT6HssxL8G68aBNWoCK3MT6PuNiFq9Blhnm2A9esGCrBTfRs9SjyKVipP4 q9AJ9Hf0pXAINGoyWNLjsVb+5/yn8dJ4HqwwWGXoCVh3y9GFsGI+Ay05DGmaugxWugq2pAhWdS1a iJrRNWD1botH4w/Gr49viLeh3wDv93gM/h73wIroB45y9Ab83Yo+xDtgHV74/80LxJrRAPoKe3AW LoL1MCisFXYKe4RnhV8JvxXHgrRvQA+ARv8JtFmFESxB76Cv0HdYhrnxojEoAv0dD32vQytJPXcY TcE+1A5rNgfs+OTkSNZALdeB9B6E9XwY1sYJsBOXoV+hDzDBbhjREmhfhnpqQM6LgHo3zOD1uA9y msFq56I/w7jNeDzphPZ0qOlusFoD0Kf/RJ+DtOOsX2PALlThBVDXd+gS1AwtjEO1eB/MwAFUBpa1 ivs3kHcmtqLJOAM/CnyNsELNKB2VCZ9igsbEZsXHk1buMPiYOOT3gPdKRRfg1dALC4xjCDnxbFQS mwN9eBdzfBT/jvXiPtIS38Kti61Ev0FPwpzo/FqpCiG9cp5eMemC8okTysaXlkSKi8YWFuTnjQnn js7JDmVlBjMCmn9Uelqqz+txu5yOFLvNajGbjAZVkSVR4DmC0ZipwepGLRpqjPKh4LRpeTQdbIKM phEZjVENsqrPp4lqjYxMO59SB8qlP6LUE5T6WUps1cpRed4YbWpQi/62Kqj144UX1wF+c1WwXosO Mnwmw3cy3AR4IAAM2lTP8iotihu1qdHqtcu3T22sgur2GdQpwSktat4YtE81AGoALOoOtu/D7kmY IcQ9dcI+gmQTdCrqC1ZNjXqDVbQHUS5ralNztPbiuqlVqYFAfd6YKJ6yJLg4ioKTo5YwI0FTWDNR cUpUYs1orXQ0aIe2b8zA9pv6rWhxY9jYHGxuuqwuyjXV0zZsYWi3Kuq+6pjnXBIqt0+p2zKyNJXb PtXTqtHk9u1btOjAxXUjSwM0rq+HOoCXZFU3bq+Gpm8CIdbM1aA1cmN9XRTfCE1qdCR0VInxtQSn 0pzGFVpUCU4OLt++ohGmxrc9iuZsCPT6fPrB+FHkm6ptn1cXDEQrUoP1TVVp+xxo+5wNfV5d855f kjdmn9WWEOw+syWJGE0jkZazZQxj5BSrmXNWspj2KDgdFCKqLdGgJ3VBGNN4GrWMR9uXjAcy+NRj 4Io2w4y0RpUpjdutE2g+5Y8KWdagtv1bBBoQHPz6/JymZI6YZf0WUZTqyVlVg/JhPBoOR3NzqYpI U2BOoY+TWLokb8zafjIu2G7VAID4UC3Itql+QgGIPxCgE7yjX0eLIRHtvrgukdbQ4tRepBeE66Ok kZYMDJc459OS7uGSs+yNQdDkZ9lBwRmVQ2f/LVZXytTlE6LY9S+KWxLlNXODNRcvrNOmbm9MyrZm 3nmpRPn4s2VJLJoypY5LJUmMpHKsFJTysrPENFFnjPJZ8C8ypW6OcqCULANr1VFr47REXK8GAv+U p1+SRzD1x09QLgbOsSV7GZ0QPj898bz0eb0zbuegv3yI1MxbuH27el5ZNRig7durg1r19sbtTf3x 7sVBzRrcfpA8Th7f3j61cXhC++OHdqRGq2+qh0EsxxNAWQmavC+It168T8db5y6sO2iF093WeXW9 BJMpjZPr6/Nga0EnRIA/8NQSmvwswTFR6icVegoS+BiHVImPYeSVRSFGuBdwCCmwQfUgT9h6qnyo fJb1ZPnMoXJUAbj1DERjCwO2gC0LIgxO/4zGDZzRBfQD0vgBaIudnF6Es5GEVFx5EEnxD3SltCwi 5kAk9ccHdCWnJCLqEEHqA702kA1lEI1GuXyukKMWGMejUqHCuAKtIC3cUmG5vEz9grPMEDGRFcyp isJLCgZnLzlgTyEqPK8JokMQRFnVfemTVNqEwZceUbMIx4m80o9f0M2iRAQeDk+y0e32oX7SpBv8 mG3puzGH+0mmrvgVXKh0K0Q5RDIRDxSKJmDBa7h8iScMMmiYOeQ91bD6ZMNqz9CsqS1Vn4NAyq3l FeUzB232soLyoXC4fIuQH95yzatb8j0USNby8i2vvrpPJFPm1T2rRBRTBIXrxxbimqhhbk10FKjL QcTFY70yrx6Kx0BSZ/aJ/Hj6qcerG8LsEwhw8IcDKRwnvBj7VffQgQ2x18lEXJb71ut4ZqxPOHRm O9GGjtLz+4z45/xfhXfRGHxEv+CgrT/9QM7rY3gpRXK6U9xOT7hFaMnpFNebOnM+NL4fNNar883z M+qDy41L7csCrTnLxqxL35x+d8BoD/bHj/aN8kco1Fu8vsjFGRcHX8p4KcivzlgdvDbj2uB/ZfxX UAyruabMjMxgmSkSrFFrTFUZU4IrTC3BDaarMraZtmfsVh83PZGRoqiKScwQg17Va3JlSBlB1cRj 9wKP7tUibR7c5tnlIZ5DpAWlwswZfWX+VJya5+DQNEyncrpPixRiHY5ejXgn7gHNHIDN1V943Vdm hdNaXq7i+Sbuxm49xR1x10jZIV++P7vHGrUSaw3+xqahQtBJb97v5iZmsWZu3T6kj6+fOXiyYXCW 9RTAcMcgqhhaHT7ZED6WgB3hY3Z3WUPD6o5wGKzDQZQB8khNnwTyOJKEn/amlGWAeABA6s1eO00d 0S32MpNmL1NZsNC8L3SzEfJMZaqHhpSy8MhPfUI7dOcEdYKpJKME5DjdNCWjOrhbfTJDRQ2gLqhh NW5IyXK5iovGlURC2eyvJDJuXLHGu4VQKJghiU6H28W7YKMk8kENzcCab9eWW2+74KLIwb80btn0 zZNw0HJLsQ9Srrnm2ukFY8bj6NtdN8XRi7GvYu/jj9Nu27rh4sj0VHv+xAUbnml/Zelf3zKtXlKS URbJKlh65eEdG//zCgwWBMcHY1fw18QDYEV8uhFOK8QnIC9/QYRK9pj1c1Qwc3BsIVcScKbwO2NX HDhArcHdYA1eAGvgRAF0Wr+uzDLdcom0wrDCuEd53NwTPGD+QFFFWVTdsksdZ642V1sk2arYHGaH xWEdZx5nudDSZd5gfVc1rFfWe9emb1W2ejeni4rLoRgt5rnmLvMN5jvNvzALZs1kdJhMRovRaXK7 slKsDtzo6HEQhwNpgSywhiaz2YlkMzUH2chkNRHTe6nZPWJUHBCPiLy4pT2ItWBhkAQDTg0n1Kaf 7OjLGJtc/9QIWhtWn2qgmjO8+IfKYe03rAaIwQqUbckPN5ivsb6KbWUI0nTyYCU3FNsCRWx2JJfL nRLg8kkwaLPBfCYmNBi8m7T9+T+6X36p8ZoVfbGfv98x7/Kl5X/4jxXls6dlPntcODT7rese+33a +M1Pxf6EK56qDww9yM3KrJs841KjQNf9XXBaPAEyNqCd+gWywEtylmj3C7hQ2CsQQVA4no5fVbIM CHbPNRyZpiIDNvg0U6FJN3EmXhk5YOPYuecGPMvacAqQk+UnEw6ABhsdMVsWQnygN71MAHfV62Ng H1PveiDiBLB+Ywth5M5AMtzFV5z5khwd0rhi4dDp2PPfxVZ/B23Wx68UR4HVcsLJ0aJfszl9S+B+ dL/jQdeDbnG99Rr3Om2zutm81brVsS1VFtOVLF+qI90R8GZd4b4KyZ0I10vLpfXSBt+GURu07dI2 2zbfZu0+6X7D3bYnpQOu113vu2ylqXW2VqlVvQptkEQOXwQHkZWIz3RlZGdnuiTEiSSUlmfhsvvJ RftDszPyFAKmp89ii5B+PFe3cO8pSijk92aTmr252E7LFEPE3k9u0rXCXD23Mbc9tzu3J1fUck/k klywP0ZsMfqNhUbOCER9oze+khDp6pmDM6kOhVfPPDYEckQVJwfD1qHYYMUg6I+bKQ3VGtAZqjcN WS63RNe7CKvaZkUBiB2SmMX0JpghOq1gFEpD2aUuYeyV3VdO0c3P7dwbeyZ2LXi16bgabyzJiR0q Kzu6f/9//dfTetnChrm3H5qV/44jKP2sAt+Cl+Nl+NbY6th9v9q5Sp/yq5/FfjgzFIvHnRMDTxZR vboidjFZDjNjRdW6OcfyOEcdMFKsyC4fxhlIQRhiRO7UVeVvxgc0vpAnfD+5u8/22BV079AwOHRy 0ApLpcIKawX8XgMOhkiJNWVcaTEhTofd7SItL93Xs2TBDQPbll1QEoxdfBz/9UscwOTo4dg7sUv+ 8mjsiQeWUitCNfw0aLgFzpzr9CxROOg46OEuFPAy4X2B2G1ZsLhRqpWquQXJruy9EpaSsyTRWXL5 0wvTG9Pb07vThXSrZaS6p41U95lWNjd0eSfXNtvvMBMMqqy52SoWJTEY9JJz6/cu/AdsnrNxz+J7 Zq1486VH9q6dcvm0kh7hkCvw8d4t/a0259Dv+ZdjjfmLK2uXm1SQbCXeQlpJD1jSIj1AvRvBpbA7 s3IaV8jxXJVgRbSHHJjXx1Yy89ow0/p5AyoYbBhbmAK7r0qSg7dgb+w4nSd//Atym/AQ8qLf6qM1 pOGgOtoywTzDXG+RvE7k4VxO5LangA+wEwf2cIqkSkZPP8a6Bbl73FE31whgwM25+zHf68SOftLZ h5x0j9gJ3sugFKgFCBXgRdBNoNBzPFzIbZ/vrHDscux1cI2ObsdOxxHHCYeAHFaH5ih08A6vb31P Uuc7aqKlsN+ZyPY7jvjA+PrymXQfCSK3nvQeQ56KQba3BNJjoP+2Ygt8mLo4gzYH9XulbjGYEaJu zxYsKS7JspGrBgzZadkzPIuvvuiqMoNy7bXYx4eOxuZdF05L/Si3+OKpY+/Cbx9999HYNpjm6fEv +Hx+EgqiIrxaXy755DQh3eWbkTotbXrWH6yf2JRx3mrvJaGl3mWhzaHbvXf4dvsOpv7a90aqURRN TpfodWWLo5313nVkM9kt7hdfF40vRj60kvTMorG2MaZMPZwfydQzciDypkfaMs9kkszqdLp7KTRb IhekY5RuTY+m/z2dT08fg4uRDrkW5IeuzQ/oabaKgJ5qhcjjiwRA5Pt5yWhSx1D9hTIGoZhBoBhD J0V3GEaNDcmjlRxTvd+4y0j8Rhw3YqNudkWMvtkRHGkEvbilEGNcPDqwyI0/cePZ7kXuNphjb3Fr 5fDMwO5n9WADNfLhROoY3egPgg2HBQA7WrYnYh4unNio9Bak49X1g4nEQZQZH3guNT0yL7M5kzSE 6+meFWaQM4Ptpwt+dQNdOdmwVaHuj3O43AFmzehs0h1M6bjShCHDdE05HXSqIbMEt8TDv3v7hf4a LjUr9pXBKnHTHm149PCCB25/7aLatpp5+PJxX2WW1lVdNLXYaiB/yr//zvptz8X6b7rxorRSr1xd 3bt14c01aVla2sVTJ8Z+Zy/yZJdPXFAUKs1sAalsgYVWCtpgRU/oOfcIWDHjucJSoUvgCux15uXm djuvKtR6k1uNcSOpMM42ErDi6/TRkoSRyhFRzQEjqBQq7Qqv+DbZd9nJIvsm+177ETtvt6IQ5vrx aN1ASDdsVwn22ioO4jTERL4aLMyglblV2Eh4Z1LlB0tTMQjSLyuiOg8CQzVRN6yXElgv+9Si8fWo IUDdJ5UieAOQlmjDPbHjWJhyRVVj/SUXXjBxTgEfuueKqpJv8yv3xP4PjHEKWG6dWe7puiebZKvL yDL1HjhMPmGWFNmK4N9upTYcwVmP2fBn5b8JDxip9bavmEKt9+DQsfONd8okriRCuGKX3emQCDd1 btWEtKXbXrzn8ck1T8cu7v3V6U+6/oKfxAW/j406/c43sZOxH6jlroS1lw3SdqA0/IuDyBo/rVcb yu5T7jfdbX1CeFx9Xnne1O+TZQeeRi4Uq9XZo54wHRAP+H6tvmF8X/3AeFr6zmRKs6Q5ddAzp262 RSzOF51vOzkn89CjKhg0uwGSm3XYEtprzY1mYvbY6enhgDc1govtiNKkaxEGM0YnYDgvAT1pDOoW WJQ99O7CCt1eZLdTI8gb7B663jINEgrgAmdgthmbfQWjFo1qG7VrFD/KEpB1kyUie9OTayo8c5Bt mWApgewGYZekOzx6jqPCo4+yQAQL2UNXPNv7VwyxXZQdOgEUbFsBRPbkgqewd5gUFis7LzAGBAX2 MtrpXjcF0T5FncSSlYGKMKL0x+g6bGDNm3WQkpk2aqbNm3UQFmKVskMrHHFg71pMZ3g1aghjAVan BpbWioqLEBdgx46UxCnDTb7HnnFf7o39+cZW7Hh3ENvFIZ27rmnywmxu/YLLyssxnlNw/8P7b/sY Tmjh2K9jh6/ZMQ2vvGrTlClrqC5cGlvJLwZdsKF09JBeai8nEVPEUZ42g1SZqhwz0uR2P06Xne5I vVCvXmJakFLvrvctSN+t7k47rZwyfecw2pA5Feajlzc46bSkGCSLVfQEZO8o+2g4o4RsNgtde8qt Vmz1+evK6ZycAn9eDmvOegzmJ+HaB6nHoSMOw5DxlDrd1Cq0qktTWt2t3qXpYkM9tlEblR2yUSm4 3CUBG05suKiJGsdNL3100f6u7ZgbWPFAOeZiJ25sXrrthqam22MrievCuVt3YStG2L/w0oe+r+ae /cWuR6J7H3iGeupCWA1WkEAueVkfEG1iUM5229zBe+33Ou7JvitXkRzVDmJ/3nTQ/OvAZ8HTplMZ 4mjTfFOL6S7DPfbHMw4apcqgnlkVWpbRHNpi3+LYnHF9plIamipWG2aYZluqA5PhfJ2ZHSo1lgTo abIkUxJVwaYEPKZsY0ZGRlDKzNDHrDGud2xwrh3dlbvVeUPu/c67cp/NeDZo6sa3um/y3Jf7ZG50 jOgOuPRAMOLS0/wRvwt/4sKuYjlQm3VrFsnSPemRLB91SLrbplbUjsGFY3DBGDxmVKAQBF+MA2xh WZQKBoGELTCFXoB4w+v76aycAQVk3oeawo4w23WdpOftQZQ8EpeIGIvYhUMZ4wLVgXm43t2MW92n sIrdhPcFMkhOislIcnyLeMxX5xhqfdhXnSJVDDXAP91LD4eG1an0AP+bvpxccK8JmMEuODJp+mif PzOR9vpYWk8F5AoTHpdRnXGv6c6MVzPeyxADGUYTz/voOPaDTUHF1Lr0ufMqcHL5sXRGVoTdWaT7 wKbgxK0F3wib8ROYQ6AT9A6DZ5QpLqCEDdhMxONF/AnYOMMQXDpU7Sp261CvW4dK3XpJacRNdxdu PWs0RFCvxe1njpx3z/fpGZkRiw/X+uI+khw8u8Zgn2NhmjwZZrcYNJkQRvLeIeGdV8OnoYEZlcz4 m7pisFdYciACOXx9wFRmdBjLKNprpDcZX+0zlDHDgYEfjiiJOwlw1xHw6ZnsToIulpFXEvTZDXX2 hdhnX7XkytIsh3N67OlLN3702Ufv5cS+sy2qayvU0kL4pfq6k998OIQLwnPm56QVaE6HrWbSgvu2 v3DLjrGTJvtdwVHOtKUzajbf/rso2JEdYEfuYR48Dd2v549PmZZC7BGuzFSWEkmt4qabpqdUpf49 VVkgLlDr7QtcCzz1aaekv6fKGIk+aj4EiW5xdZfBYLWY3QHZ1z4Kj7KNNpstIasVM+/djrrpvVF6 RcKuj7Ahq0/9yIgkTchScanaal/qWuppTQMTghpSEj6bHtxABNm2AHYk73DAkuzAYvEzKw5iEjtz sO7W2eDOXbcsXXzd5iXLtvKhB2ubY3+MDcVOxT6snj/0JXew76mH+h5/ZBdYkJvBgszlQ8iFHtTd l9iW2e4WOEX0iuWk3FZDamzHiWShY7PxBhdSnQ6HqogpjpDTieiwzC5dy4zsdeE4rGefh+qqC3Ro p6fHQ9o9JzzkGw/2qIaQIjOPCLQ9Mj4hY9nrrihPCKIjnLwXoKsWQkIww/chFYPszqMhUMI2ebBh DyS28HBUApSbNeFw6xV7LsJe/5yKaR252Ltr/uLL99xNemKeoy0TZ3cdwwM/fEQt5XwYZwXMsRf9 l35xnaXeXu9abmm1t7qu8Wzw3kPuMb5ufd3ze+v7ni/FL+UvU750nhZTxqeMd86wz3BVe+qNrUZp gr3UVerh1gnrLFuEzZZt3ifsj7sO2g+4FDPz+6kRCvfbHRFzsYnmeEdFTMlzv+kQ5pEKcrTbDEgH UqQDHSreCTvqQ+BpeCjS3BKmuWDuCkwUMSW2B6lSAI4/dQnFofsBuh0Iw0EfVQydbDgGe+uhk3Q9 hhMHy9UNmC2ZpHMppQ44AyU8Dz829mfzktmt12y6onYpHMzCJ3/7ZezP2DX48mfk66K5827bc/jB S9sKfvUyDoFhkXDW41R2XbGD+DFM30eo2K/IBlGV+vEoPVV8EI83qGoHDkmZ9PRBT5g88hqXrU1e cB8bAsUGMzzErq/oRQSoMD3uwg6+NHgT9uZ2LSydP41sxd43r7q5XetMWzyftncbQqIXTudG4tEN Bi4khwxwiIVtcLxbV9ImRFRtwsSIQk1sEuqPpuVDLkSiIqufKl+rPK+oagpJ462KXw2SMbwGh85l ZDnfoqxQ15H1/KPKHnW/ckg9pXyvunbxO5Vd6uvKm+rvyQf8+8qH6nHyBf+Z8pVqWqesV68nN/HX KzepO4lUZ2ghK/hlynJ1LdnAS1Wkhq9SatRL5EuUOlXyqAXmCJnAR5SJaoVZ4oiRFxVFdRIf71ak pB/yE55TFcEoSUWi2VjEjulErpVNEQON2CjNBlNE1s3ZEQONIOtB3UoRg8yBrmAiqUimS6SinBrh hAluwAWD1vcGaUZqf3yingetaLysKEUc7+A4nsBcFXEEUALVcEaeEKOqKook+0HL+rGpjz4JP0TG 02u4vksbIgJzyHPnRYQiSZc2wao9vAlm4bBBMxhJPxmv22FzqwMh0oEIFcGZ0UirMY3tgg0/uOJw 2Fr+F2u5z2sdWj20utznscL+EDKsx1ZD561sgUNvz3/KkXyikTIXfIgcP7rPoNHHFw3sk/A8KLya Xp5gTB8b4QC23YafBy8u4Rdig7GPY5/G/igcOuPhvvi+mr/uh400gE7dh5DwCr3TxLfpPlnEdruq ChzheFBHRQWtERRZUeV+/JwelkSHJIkcfRCkOgQBhMRxosopHOi+KAk8TzBBBoMsyXw/ae4VpskA dLvELkCJiRjOv/4ccd/rpZt5T2IznzBwEHnhQJy4rUsIwxPmr7G+yhCZIrK1XH6Vo3F5QoX2K5qB KcpbvXJ2BDHvi8BleENitrKTv1fs4aP8AC/dID7Bf8GfEni2Vkrn0LUyoGcCkiVeoHZym7n7ODg5 qXu4Q9ybnPoSd4Q7o3IXqJM50rEatu4g53rmy8X4F312Q4VInz6kWAwVfKHJBZHRUcFr4N05+hjD 4k1AOA0wCBQMAhGDSbpec0ritMDcPt05sJswOo/wL9nuwzlkAb5l6ANSHbs2duUJ4dBQF9kx9NqZ a0n029hUxLEnAIthJlPA4IxBH+gV63LxcvP63M/5U7DqA05FzBkTyHLZ/c7ZTlLo3OskTqcjmJFl T5E1RxZGJDW7XewWiViTk72XqmziXo9erOqBwnw9vza/Mb89vzt/Z35PvqzlF+aTfEeGhrSUwhSS Qqc07/wrvsT9a/IW/8dX2s54d296GT1U9vooGHGlTW/4E7f6ySd4FnZC16i6U29fNIokrvgTV4QC aDrdHiXvCDlbIJkIBe8mM555asvCtkWbdzY8vHZG7LOYCee8/Mvciy6pmTHmnT3Y3hOePFff8JZw KP2y+xYtezqc/cKm5sOrTTLhX4/9UlAuubBqviIMHYytV4wNsyZflovYPfpx4Tic+ekd6SP6/HuE e+R7jfeaeRlLZtkiebI965V1dmmdbb1zM79N3mbcbL7Rvs2x1bnVvdWz2WeU7LJD8jntPofP4/RJ KXkmxZsnca7svSpGqlXVVE5N3Hin68m71J50UUs/kU7SrdlwcqY+hd5Z0hvvtB/feDewW9VBuj8A 4wDHrpRIKb0AKh6+2MYO+9nNUf2Uol8u29aHq/CNsY2xw7GDsY147Of79n368XPPHSXvHb23vTc8 IbYqdn/soVgbvhUv/3ssHo+fOf0D9UZNIIfLQQ4+9L4+a7OyzbHNtQvdK/5aeY97z/Atp2QpOcYc 02jHaFeX0KVsFmQpRXK7U9zu0SSXyxKkHOE+4R7lTe41g1CBZ2OC51jpS9wnqH2gpxlP4tpAhcNM P16ouz15vAyHanvEXLPIgmdbsEV3eiJwDs3RM+x5Kmf5xrwAfYNYVb7CNJzmzO6RsEXyS4USR2+n +1I3zj0rrBG3B0N0734ssVsApCGxs2qA7QJ7ysfE5k7ut0UbfRQwjq/A/smx334d+8/YVnwVjmDT E81FsT/4Hlv7i9+80bN2D0m99MSXILCFeBW+a9fl0eqOG76KfR/76uu7qQatQIjfLLwJ+uNH1+u5 otWcEhGsVntkgmeCVxfmO5f6npJExZWCdK8/eYG5cNT4NNhw3tFrvd9I950ZKThNTcN+hEMIWZWA 1aW5iMsXsARgZ231ao9NSW42kvsisKWD1LOgChjnMevQ5w3hxCMQOMXgYPKWMKEVUta4zOIinq4x UYTlZCPfxPGcU56s69uu2Jkaixlw6ud/xaNan64PD2EvvqrUeHXPG/6xE+d0XXWN1vf90BMNu2+d sTBmZ9fmGN0Oo60Hu+RCvXrYgv24DBeTYutkPNn2R/x3rEiCS8gkdbblNgFjkuKw2VM4B8Fsd53O gRNSHU7VhZBBDckK21YrOK5g5Z9tq5Ej5HKe3VY78Qkndv7zbXXylYuE1xksK7O52bWqPHzisMHJ AuwNSCHC5l+kKH5q6+GmB2enx45rF19Qvao4dhzM8We7prVvvXXoNjL28YUlVds2D30Ng4bxb4Px L2JPDf9GX0P4uM9koxZ/QL/GmxeRYJOTAt5pqbhXfVF9Q/mN+pGqzuUaOWKSPEq1eIm8VhQOKJ/w g/wZ/ltRmCXNkpeK1/A38Q/wDwr3i/dL98uqn7eLYT4s5Iq5Uq5cYKrhawT1rP8GLy3yBoEX6esp 1DuDfVEN4KGv1H1CgVzml7DUAs45hLsR1SfYqVb8LLlOmG+2nlrtgYWS0J7k8YPtT6gXHva9dGhv 9CqB5KsXdFOCOkC3RvqwbaAt0/HC2F1gbt6JfXs9bEdO4bWxq4cuxx9viz0NTW+EHe69cBrJxhMP otEgowabWgEbdaNTdBkjXESOeCLBKjJVnuqpCho1rmD0XKVxdPfoXaMfFR+Xdhv3i/uN0dFHRh8d bUajC0bXQsGLoz8ZLY6mJ/oKSHezQkEK8JIv3UXPqKpEnwXoo3jJarNlp6alhbLBDosWa8hu0xeW NNpwmw3b+km1bvGlhtLTIK8tDTeCfYG8Z7NCoWx6kO1FKDt5F0OhPg76nQ2k2XolhHIImdmRbH3C BZGC7LezP8nmLNn+7O5sDmVr2YXZ8Ww+25vzafnw44LkSwvJS2yw6WFwi6dg61F+TlnZNjHxPHPY W3aE6YMAHKYHYnpDwJ4HuF1MebPPKu85Pd6IuR0DS+8urH7ksq5HckCb07Mvnrg8P3Z8VMW4yuV5 seN86LYn582fP2/RZVX3DtWTRT/PL5+24+4YIdUPLBxTfcN9Q2dgzsaAfh9k7yF16MECpZAvFGqV dqVb2alIIhZIFs8RCcmK2+3jNwlY6Md5uipKGi5Em6ith6SNM9eSdtJNdhKeeOWhp5PvkFxct4/o 9DEWbCnYc6ypLVXHks8M2YOPBtjqltCn3viT2Ez+5tgs/uXTp3+YBL1aDN7pMHgnDXZC1eNH1Yxa IK2V1xpvlG8w3ui+IVUR3WKq3W1PzbHleHJ8OaPkaYZL+XnKQsMK/mf8VZ5O3wHzAeuvTa9bf289 bjVzaaLG7tb8vjK6RLIIxq60PFGxU4dkr5mdglOoN0qh3ijXlWfhEMaadxFkZ9sXEL+mccSnZRRm kAxvdo+KLapfLUz4+b7Axl3neSX2/sPJwdXsbaeEdwLnNGijz0hBEYZfdwD/VBKwgYfKyCQlEXvp uddURKeVuvjSEq6CbGyI7dr/eWzP0wMHb/4dtuHiMbGP/E91v/zZFy80PD+FpH431L9w20t42buf 4eZF0z97q3TlNaf+Gvsh9sP0yCFqu2NLyA1sbmv1IrFQ0AUilCKugp/NE74UW5GdIyKReMEKRuSk 8ABC9FWvI5jDjfKyq9nD6XLr4GADVdak84EMCOzpqjNoK3bejv96/HhsiXTJx98v+ZjayxC06WFt Vum5FcJs1mQhr7MW7cgKTon44bx1Egv8A+gTjHGj8ou1ibaGWOWJlqBV+hQXg4YU24IlxBMzHwc7 ffrJj4Vdf0icpvl6sDUutEv3SCnulIXycji48BisjLVKrrJ8aRVE5oRsktkkGg0G2J4RHHIh5oQQ jkMl/+Jux2imdsFkMp71RUZ8AvbV5/uis4+pznNHI+94VtPHUiOcD1vc4JL4+tjxzIvLpneG6TOq He823D/bT0Y93TK+9obemJ8PPfjslOU3/IxK9A6IngaJcmgue5dEHw3WSRdqBdItRIUB4YjwjSD4 hUZhk9ADGQI4Cg6GyoVw4j45kBlBXj7Z63Mek11tdjALfwf2Coe+rx5uC3uhgkzdCWdmlYRG3IHw Z+9AhugNCHtvhTIPP2E3wJws5ENwGk3VnUKOryAi0UikkUwjcJwf9AHkacc034TI/TwW4fgvq0YD +Hli53yKT81AeYZfG4zQ+xO6K12LqEgwOJDXkIVyDRE0wbAFKcm3BVVsMrK6DIo7wmOkYBGpoDx0 mYXZq0aput2AVN6gKgohWARcKaO3V7onLSdiMPnZuz28CaybVa1QZ7NFXagbeFJm4Ok64fhDpBCU pVu3GEsQ1rAOq8NrfLXHE/ZSSYTh2NsAStDgZW87sjQTL10z9jIMXWCXyOEG+tQo8a4iDqS46UPd FHCsz8Xm4ew3JrhFs/UtHIiB9Ib+tH+qKy+PjKI3fGTE7Eto3UGkQNcr6PwrtQrpVqLKgHJE+UYR /EqjsknpgQyBEyUk8JwFYR0dQUeBswFUQhREiVeJFML8sFbwXvkftAJsM+wb2GtJ1vJhHaGvUuDE VIOyevkDmI+d+WEGH4IuYvQQzPpjwjNIQBfovloJ9u+EhxMCknnBJxFu5K2BOPbgyJemYrTNmUMJ ZWTn5YDzITgpHxWe+WH6d3T0W+Jf8HeyN1rS0EP0GeBpfayhrDT1wlRip/fWiVvr7ySxhJ9omphS kjqVrzHVpExNvVO6T1GNZugMGnmXnWIwWJA6fJdtHQ2rhD0KM+If32SXzxwcKv/8H+6x0Wqw3VPq dAO9x07cYgv0FputI1jZ9uQ1tjPF4T53jb0Fe6/rfTkWGzp46T7dHpm+oeH6G5a1bIZt54k7Y8dj f4+diH10af2DJPex2e27njrw8EN0NWUgxL0Pq8mKZz5rf5PHuD8e18dabREVQyRhWSXf4dMqKTVc qF5orMN1pBW3kk12+RP+iPEb/qiRVwv4h6XnSSeSkYrn6YqiyCIuMD7MNucWqxWpt/K7wDBroXxY kjj8rKIWWy3JjRCFehZ9KmWxWjRLoUW3bLKIFp9uxwP2I3Zil+Ri1G3cSR/7x4/qCtQhGd00hcO9 GP+TZ1tGsEW29T8bfrZFLWhDuMN6ErRhdcepBppzsnww3EFv0oa+PQbxIIV49fA+FceP6GbFE4Fz tFoIUJVkun+iD3ITW1dEH97gxDUFAQuiGsoMViMEE3ssU4+KS/C4UlESYOeBpXHFAWcGvmVxwdja 2DZuVWzFrV1puO8P+M32Ag6TL38dG/OAxDTxToSEJtBEK5jDTXpxjpCjXuhu4VuMQq67zD3NVe9a 7hLK3ONSt6TeJ9xtEPw2eimTYs+yWGXvP7xpldIdwFqgMEACNruGNGuhlVjpAtF+8g7m7AXM8Guu gSK3i71gINK/YOIKZRKhtyawU7yTpD/XeG1/Y17p0pnXL3506F2c8/HVpdMWlZevnDtpv3AoLfRy 7Pi/77++Z0lNrp9/+UyJ2b7gtT17Diy1m5m/gb1XJozUi7br4yVZUiSrW3YpF8oXKtIlygLr3dZ7 bPc6H3A9bn3O9XvnZ+Ip0WAyGmGwUlaKYjRoprfpxSvZoWfoqbWpjalce2p3KtFSC1N7UgdS+VRM YMTeQu+Al/PSMfvoxerIMXecakjYo0HmStkz95SADRbU2SckVjNJvup0B84xpNx69cZuH84pvPaD Z3734UZHOhxNPj88fuGVy+5+hguficVOf3R3fdMD8zeeouO7F2YyCONT8L+DJnGi7OXcMm+XCQfu CdEbQXq+o1fF7JyXO3dehCuSZIckyZxMiMQpPCEKJHh6S8zTW2K+SHyb7o1h0F7dUGtoNHDthm4D 6TEMGIhmKDQQg6wkK2W3lea5cyNKEbOMA+wuaEefOkIMbGE0DEsCUsw2M4eGEu/PwmfLNa+eO7zB AjRnR2Qtca8+8JxCb9rZq/SJx5hTGFX3AUOJ3G0oYQO7wJcfkedCJHAurojTOb6au1HeKffIvfIx TnyVe1v+SObglCZHuInybPl2bpfcw+2Vo9yLsiHxFYXikgjRi9lXFI7qpoKiCNFoJDlKIOcecDL5 ETIPIkZdPUqDFEQykSQP4dzSGJItTSTF0iyiS5eRBZLiIKnSTDJVul96SvoN+ZB8QY5LfyeGbJIj zZDWS1ulp4kIytAxfMQKh9Hw9TnYYHp7Tn0Itt2LNVKHU2K/H9oHqpDHvft9NffCmSo6+y7wVofZ 7C+l37o4sd9SJhFsTz7FMBKJPk+HuTWo7Am0lk2fVR/VU0GWmJNEHTJEDRIiB06OB52njywGK4ao /WFGKIwTj9Tp3sNJRAmOSsSBMRFzBQJukSe4n4R0C5KlLAXJojBNblcwU4ossI6KL6k/Br4QpqSW a4R25uDELZJXPfP0sI5ASLyQRg1m4lsnn7N4MPngITnpmL4moyUeuAfpl0i6bSU6nibW4mbcJrbj TWI3lokWyIlIOtCR/vjuXluJgZKXWL2RWVKDtIS0Sh3kGulmaT95XlLSiVXKI5pUQQqTM1crbSfd 6q2G0+SEZEl+J4PeTNFbcHCMmB7ycMDp4sjQfH7smY+5ydjAPXf68zPB7+hXZ+lXYcitL73ZXbDI Uv6t7JXZN0of+bT87C8FxAdjF4v3wnkQ09dph79yi5A0KTYLTTn3JdwffSm1QIQs4ddQ/QI0g18T HwT8bsDvEstQPdmDrqA4wDUQ/PwaNJ1LR1sAnwJ4Jf8puhRgIcAdpAy05lN0M6TnA94l7kG3QV33 0fpoXQCboHwFhNuBZpuUjjYCPgbKFkN9t0MIQfo24L2DBqAxMPgpegjCFh6hDKjjTgh3QLgX8kBP 0Tj0NxIkF5BvuFl8Kt/Iv8F/I2wSN4unpQvkKfKr6jL1M/UzQ73hO2OBqd48yfxXy1prinW7TbX1 2K+yv5ayPOWRlPcdNY5rHT3OGc4HnG+6LnSb3de7/+b5q/d5Jq0CVMx+7oJ6OCuk5iMkfk9ehr0c /YbSXO4rhJLlK1jMMSmrLMUxLhl1JnEO1aGrkzj9LYX3kriARqEvkriIPOxXMyguoUN4VBKXUSH+ IIkraDuRk7iJ7CG3n53XEqE8icO0CquTOEG8sDaJc2iMcFUS55Eq7E7iArII0SQuIqPwYhKXUIPw myQuI4+Yk8QVNFXUk7gJzxdvhJoxz0FbZumjJM4jn/QFw6m0VJkkcR65ZCPDRcgX5YwkziO7nMNw icpNnpDEQVbyZIbLkG+UFyRxHnnkxQxXqPzlq5I4yF8+lsRpPX9J4iB/+VQShzqVWUkc5K8M1wPy V7qTOMhfuTeJg/zVOUkc5K9el8RB/obEWFQ6dtOBJA5jN73KcAPk201/TOI8GmU6wXAj7ZvZnMSh P2YXw81U08y5SZxHaeaEHOjrhHbzgiQO9ZiXMzyFytC8OYmDDM03MdxB+2P+eRKH/pifZrgT8h3m 15M4jzRzYr5cjP77JA70FpnhXkpvyUjiQG8Zx/BUOqeWBUkc5tTSyPB02h/L2iQO/bFczXA/o78l iVP6hGwz6ZxankniMKeWQwzPpfKx/C6Jg3wsiX7msXr+ksRpPacpLjP5W81JHPpvTWM4G5d1XBKn +VUUNyboL0/iNH8lw9m8WDcncTA81lvQk3DAL4KtSCFYHA3NQ8tRC8CZqA2tgtCJNqB2ljMFUh2A 07gJ8lsZRT6UVKKV8KehOZC3DPg70RqWagHYAtRrIW5mvzXSARRNQDsZeFdC3o9bmTCCRjtLNQEt YPWsSbapoRKorRCNBSwH6mhFS6C0Dcrb0FKoa/RP1vLP6siH0a8fwZE3onfzzquplY2rCUInk0Ez 1HglwA50BeTRtv/n8qO1rmI1JvjmQ6oVUlRiGpoLWBNLJVpeBbkFrAaN1b2cjUmD8behLijtZL2l 1Pn/4578I928s1gVo1zH+roM0rNhrEuZ3GlpHpvtNrQ4OZZZrGQ5k2IT9GUM5NWyljpYSSuT4VyI u9iIEvOiwVyUofEwG/VsNBqT7QaAXUyPEjJKzMFS1tdOltcGcTPLb2ftbTgrKQ1yOlifOpMyWsVk mUg3sZraWetXMpkPS30xq2N4RlYmx7nqbC8SHMP96BhB2870sBl6vIS1kZDHOtZvKpGfHkMiTWmX QGtdTCLNbGX9WBKUYyXDcoB+NECqgYuT/f7pulf9P4z9XO3NZ+e+g+nX8FwOa89PjWCkbp/fr4kj 5oiOJDGWTtbesF7S+hNjbYacdWzkbWzV/StNaDpv1luSK+XH64VKtRPouhgn7e3as9qcqIdSrgSK f6VD+U9qRYWF47R5y1u0mW2r2jo3tLdoU9o62ts6mjpb21bla5UrV2pzWpct71yjzWlZ09KxtqU5 v7KjtWnl5LaVzcMsE1iORrMmLGjpWAOcWkl+4VgtZ2brko62NW1LO0efIxlJkV+0nhXksermJYha 12hNWmdHU3PLlU0dV2htS/9p/7TWVVonlM1f1drZ0qzN7WzqbAHmVc0FbR1aG5R0aEvaulZ1drS2 rMn/Z5WczZtHo6qOpnWtq5Zps5cubV3SouVpc9oWQyuzWpcsb1vZtGaMVtsE1S1pbdLmNnWtaoax aGPLxhfVt3VpVzZt0LrWtECPYARL21Z1ap1tWnPrmvaVUACd0to7WiFzCZS0AGxao7W3dFzZ2km7 vngDG8hKaHMVrQIKaB0dLLe9o625a0knHe265dCRES0AbF21ZGVXM8yLNtyJtlUrN2g5raO1lisX Q90jqFf9y9YZeTMdfUfLGjpKKp5zDSSknaxrIhtRTiu00tlyJZVlRyu02ty2btXKtqbm84XQlBg6 TMfZeWnr6mzv6tSaW9ZSMQPN8paV7edLKB/scAusWLpeqRcZ6cnOL+lEXdgEGv3leTTncpey1Tyy LJFTzfg7zytJ5nFbucPcq9yLEO8bWX5e/v/uPP535/G/O4//3Xn8787jf7zzOGvBW/+pbU+UXARw OcC1wE9zus6j/cfSC5kE1pxHNZxXDb5gJVicU0D/JeSdb/fPLxvmWZP0EW0/WeO50gUMG0mTyJnG UmuZxzm//PySWqiDjrqL2QIqqw3nUf9U+UhJtf1TGbbxfn4SP5Gfwo/jx/M6fwFfw5eNpP7J8nk/ 6VPP5Vb/w3gSOTU0hccCzciyc7k1bBW0g6TbfkRxNh/b0J+4IOjziPKzeRcxK9H6I505l/vf1av/ puz+2/UNawSH/uVn37zuShP3NNoLgSArxBqEHo5e/+jc032SqUjvB2h3MNjrChcdjA8AMqGY5efd WdT9AvcUWoSKIfup3vk0+6k+vaqIweKJCVgwlsFeOVEsOYr8lT5gK4BAkCWJzYZwK4RdEF6EIEKH nkKfQIhD4LgnuEd6q/1Qw2NQkaXSwT2GMPTyMfQ2hDgEDnr/GIzlMfRNMoeHXv2iTzHS5n/BuFK5 XwCXBWIrhG4IeyG8DUFAbRDvghCHwAH2CJQ9ggj3CPdwr9VvrVS5n6NNEAh3P7Jg+iMIA9y9fVYm m/v6LClFeqWVuwvVQiAoys1EAxAIVHsbsN2GCJDX9OaNZSKs6VPNRVag3wGd3gEd2QFN9kCMWVqH QOl39KW4aPXX91psjO9nvYWRBNJn9RTVghTWI8y1cKtQEPm5jQBHAVwCMB3gYq4ZmVg/9T6Ltagb 2qsA8grOCXbaz1VyLvDSfq6K89E3cIGsq9ecaKerNye3CEY8hfMwEgtnQhGAMif1Fvm15zmdCX9r n2Kg/dvaa3UWHeZu5CTkAKpuoHL7LYc5FWZWZSOZ16eYinZWGrl5MMx5IBY/9BGDlFexilb1QkWV Nm4ql4ZcUHYFl46cAKu5UQw+zj0MS9rPPdQXSvMPPM/dwbhup5VC85MSqjWpz2QuGqhUOPraW5S7 BSbgFtb4zr7Q+CJUGeJyUCEEAjLeBNgmpvTbAdsOs7YdZmo7zNR26NR20D7EbYMS+msOBdxVqJ1b h3ZC2AU4VStnLwj0IEMyc4oOcl7OA4KxPg+ixJDr61PMtGeeXnsKI/P0Gc1FFYe5NaDna6BOnevs c3uK2p7nctlQxvR5UilDey+o62HOnZgaYHTRKTnMpYEgqGDSuVG9Tn+00g9pqsh+hMlb5AgVEnmX /AedbvrLugz+Jgl/m4T/noDxAXIksSjI7yg8WplGPoPKFpGP0S7ACHmevAJbaj/5iPTTXpAPyUFU AfADSDcDPAiwGOCh3sAb/n7S3wcA+v5Ar8lFB0te6Q0XJBF/VhJxpyYRu6uoMou8TF5CaVDF7wFm AnyJDKAMgC8C9AAcIJ3oDYD7SQnsMvzk2SR8lbxAVZw8Rw7AHtNP+nrNtAvRXomCvb0iBc/0okSq tsD/AnmGPIV8QPrL3pAPcp/oC2X6Lc9DfZg8Rjp70/32SpU8jOvwSSDqQR9QiOzkkd5SWsnO3hc0 /0Gyk+zUPaV6lp6n7+YKswrzCndzWhYcyUu13VqlldwCBmQXgfVLdkBcijQC2gNBh7CTbOvlS6OV QzAmOi6CuiHuYVgjxO0MQxBbz5aeYFgFuRHNhkCgjo0QNkHohnAt4iG+CsLPIFwN4RqW0wmhC8I6 sCbtwNEOHO3A0c442oGjHTjagaOdcbSz1rsgUI5G4GgEjkbgaGQcjcDRCByNwNHIOGh/G4GjkXHU AkctcNQCRy3jqAWOWuCoBY5axlELHLXAUcs4dODQgUMHDp1x6MChA4cOHDrj0IFDBw6dcRQCRyFw FAJHIeMoBI5C4CgEjkLGUQgchcBRyDg04NCAQwMOjXFowKEBhwYcGuPQgEMDDo1xWIHDChxW4LAy DitwWIHDChxWxmFl89MFgXIcBY6jwHEUOI4yjqPAcRQ4jgLHUcZxFDiOAsdRsm4fd6TyNWA5AixH gOUIYzkCLEeA5QiwHGEsR4DlCLAcSQ69kwmDgNpshLAJQjcEyjsAvAPAOwC8A4x3gKlXFwTKGwWO KHBEgSPKOKLAEQWOKHBEGUcUOKLAEWUcPcDRAxw9wNHDOHqAowc4eoCjh3H0MMXtgkA5/udK+T+e GnItrpPB15JuPJrBTehrBjeiDxi8Bu1j8Gq0m8GfoesYvAqVMrgOhRiE+hjsRH4Z9/pLLZUuMAGz ISyC0AZhF4S9EF6EIDHsbQifQIiTEj2Dt0izpV3SXulFSdgrHZWIRZwt7hL3ii+Kwl7xqEi0ylRi YnYUTAu6lcWbIP4GAjgRiCsYVkEi0G4E7GwJ/EVIRLcNat/k4rdz8Yu5eG8uvjUXVyrkQswzS6eh UgIdx3W6MTTJ/wGE0lD2JLBMtxz42u3vDY3z9+MXEmC0Hgb4NYR9EHZDuA5CKYQiCHkQsiD4WV4u 0NfpGckqX4CQDSEAQaNNIBd9am63yfpBYsK7+14zIfq7lr3ZOcD3fG92IYD+3uzZAJ7rzV7sr1Tw AZRNd0V4P8zcUwD39vqPQfEvE+DpXv/zAJ7o9UcANPRm5wO4tDf7t/5KE56P/DxlnZeEc2HcFM7p 9S8Asot7/aP99F247BClzoWGsqB0NK5DxwBmJbkyEy0Fe/0TAWT0+ssotYyy6cRjEeWx7gkQKOT6 oEPfHMR1PNYN/kH/Hf6vgf3PIFhQjw+1fh7A21n9eIGu+l/I+zkQV/p7K1VKD/5hXxJGKdzv3521 zf8A1IWzDvjv8+f7b8nrlyH7Zuj3NtZEr/86rZ88paf4u/2F/s68Y/41/hn+Jv8cf0MW5Pf6L/O/ QLuJ6nEdeeqAvxYqnA6jyOr1X5jVz7pY7d/g1/3Z/jLtBSpfND5Rb2neC1QCqCjR+hiQb25WP9Xx +aX92KbnSiekndKl0mRpohSUMqRRUrrkkO2yVTbLRlmVZVmUeZnISHbQt3PC9I0Ah2ilQORpzDPc SmhMEq+GECwTNANFU7gaUjN3Mq6JDixBNYu16Km5wX6sXrwwKgQn46i9BtXMmxwdH67pl+JzoqXh mqhUe2ndPoxvqYfcKNnaj9G8un4cp1k3ptLfe96H0Y03px5EGHtvvLm+Hnlcays8FfZJtrLqqp+I GpPxiF+49IxE06N318yti+5Jr48WUSSeXl8TvZb+GvRBYiGmqVUHiZmC+rqDfDuxTJ1D8/n2qnog O8bIQJvNQIayKQAyeTLSKBnYk8mUDOYoQRcCdqALUAB0qgmFGF1INTE6HlO6fR9oU6v2aRqjyULo A0bzQRYaQQMaA7xV+0IhRhXUcB2lwnVBjXVsNKvI7weSPD8jgXOwn1Xkx6yxaME5kqwkSclZkhLW FofP0fgTNI6cYRpHDtCE/x8/LZPDuG9s18ZX6A9sNwantkBojO5Yu9wT7V6safs2diV/eTvUuHjJ cgqbWqJdwZaq6MZglbZv7Cs/UfwKLR4brNqHXpk6r27fK3pLVe9YfezUYFNVfV9FeV3leW1tO9tW XflPVFZOK6ujbVVU/kRxJS2uoG1V0rYqaVsVegVra2or1fvaun0ymky/A81gHzGooMONqYH6yS5r +ySq0AcnBjwbUw/xCD+BDOH6qDE4OWqCQIvyKvMqaRGsM1pkpr+inizybJwYSD2En0gWWSHbFpyM zr6sR4noT3bVRANzF9ZRVYnqTT89Z2vohxV70NTWKviHdCcL8DeSEq35yU/nT326urrW0KgrvAah mmju3JroOPr1ZEmCphqr6iEvfziP41jePkWZ2h8fgMIwdAJ30uYoFsb0Z6Z0FU5dEukReyRCjwqd fb70orbD4ME3QYBzHFnXW8COz2RdX0YWPb909hWUJCAcVyns9QWK6LuGpcBKYVYC6rY8QHZm7czb WdqT1ZPXUyrS3+raDZn+3dSV9hbs5lBneM2wIADtrEeJX7+C9h7uTUtnDfdQJByuD69hX0BAPxb1 8HfWQehnBbsmWesaVn3n8IQk8tckK4GZSLTeNczWlWRihV2MKVFJInU2OveBFEL/F/IQwxIKZW5k c3RyZWFtCmVuZG9iagoKNDQgMCBvYmoKMTc0MTIKZW5kb2JqCgo0NSAwIG9iago8PC9UeXBlL0Zv bnREZXNjcmlwdG9yL0ZvbnROYW1lL0NBQUFBQStBcmlhbC1Cb2xkTVQKL0ZsYWdzIDQKL0ZvbnRC Qm94Wy02MjcgLTM3NiAyMDAwIDEwMTFdL0l0YWxpY0FuZ2xlIDAKL0FzY2VudCA5MDUKL0Rlc2Nl bnQgLTIxMQovQ2FwSGVpZ2h0IDEwMTAKL1N0ZW1WIDgwCi9Gb250RmlsZTIgNDMgMCBSPj4KZW5k b2JqCgo0NiAwIG9iago8PC9MZW5ndGggNDI3L0ZpbHRlci9GbGF0ZURlY29kZT4+CnN0cmVhbQp4 nF2TwW7iMBRF9/kKLzuLKrFxnCKhSBSKxKIz1dD5gJAYJtLgRCYs+Pv6vpu20ixAx859L8fWS77Z b/ehn/K3OLQHP6lTH7ror8Mttl4d/bkPmTaq69tpXsl/e2nGLE+1h/t18pd9OA2rVZb/Ts+uU7yr h3U3HP2PLP8VOx/7cFYPfzaHtD7cxvGfv/gwqSKra9X5U+rz2ow/m4vPpepx36XH/XR/TCXfgff7 6JWRtaZKO3T+Ojatj004+2xVFLVa7XZ15kP33zPrWHI8tX+bmKI6RYvC6jqxES4X4AV5B7ZkAy6Z L8GO+1JbCZsn8BP3C/CS+1vwWriy4GdhJ7Ub7kv/LfNL8Asz8q4dM8jrggxPTf8K79L0N8L0r1Cr 6e8kT3+zAdPfwk3T3z6D6W/ho+lvcQ+a/lb60H8h+/R3OLumv5M+s7/0p3+Jsxv62wpM/xJ5M98/ 8ob+7gU8+8PZzP6yT//KgenvcG9m9pc8/R3OYub7X4Ppb8WH/iXuytDfSn/6l2sZpHliMFKY+c9R Ve0txjSm8mHIfGIy++C/vp1xGFElvw+hFtc5CmVuZHN0cmVhbQplbmRvYmoKCjQ3IDAgb2JqCjw8 L1R5cGUvRm9udC9TdWJ0eXBlL1RydWVUeXBlL0Jhc2VGb250L0NBQUFBQStBcmlhbC1Cb2xkTVQK L0ZpcnN0Q2hhciAwCi9MYXN0Q2hhciA0NgovV2lkdGhzWzc1MCA3MjIgNjY2IDU1NiA3MjIgNjY2 IDc3NyAzMzMgNjY2IDMzMyAzMzMgNTU2IDM4OSAzMzMgNTU2IDYxMAo1NTYgNjEwIDI3NyA2MTAg NTU2IDI3NyA4MzMgNzIyIDcyMiA3NzcgNzIyIDYxMCA2MTAgNTU2IDg4OSA2MTAKNzc3IDMzMyAz MzMgNjEwIDI3NyAyNzcgNTU2IDI3NyA2MTAgNjEwIDUwMCA3MjIgNzIyIDcyMiA2MTAgXQovRm9u dERlc2NyaXB0b3IgNDUgMCBSCi9Ub1VuaWNvZGUgNDYgMCBSCj4+CmVuZG9iagoKNDggMCBvYmoK PDwvRjEgMzIgMCBSL0YyIDQ3IDAgUi9GMyAyNyAwIFIvRjQgMzcgMCBSL0Y1IDQyIDAgUgo+Pgpl bmRvYmoKCjQ5IDAgb2JqCjw8L0ZvbnQgNDggMCBSCi9Qcm9jU2V0Wy9QREYvVGV4dF0KPj4KZW5k b2JqCgoxIDAgb2JqCjw8L1R5cGUvUGFnZS9QYXJlbnQgMjIgMCBSL1Jlc291cmNlcyA0OSAwIFIv TWVkaWFCb3hbMCAwIDcyMCA1NDBdL0dyb3VwPDwvUy9UcmFuc3BhcmVuY3kvQ1MvRGV2aWNlUkdC L0kgdHJ1ZT4+L0NvbnRlbnRzIDIgMCBSPj4KZW5kb2JqCgo0IDAgb2JqCjw8L1R5cGUvUGFnZS9Q YXJlbnQgMjIgMCBSL1Jlc291cmNlcyA0OSAwIFIvTWVkaWFCb3hbMCAwIDcyMCA1NDBdL0dyb3Vw PDwvUy9UcmFuc3BhcmVuY3kvQ1MvRGV2aWNlUkdCL0kgdHJ1ZT4+L0NvbnRlbnRzIDUgMCBSPj4K ZW5kb2JqCgo3IDAgb2JqCjw8L1R5cGUvUGFnZS9QYXJlbnQgMjIgMCBSL1Jlc291cmNlcyA0OSAw IFIvTWVkaWFCb3hbMCAwIDcyMCA1NDBdL0dyb3VwPDwvUy9UcmFuc3BhcmVuY3kvQ1MvRGV2aWNl UkdCL0kgdHJ1ZT4+L0NvbnRlbnRzIDggMCBSPj4KZW5kb2JqCgoxMCAwIG9iago8PC9UeXBlL1Bh Z2UvUGFyZW50IDIyIDAgUi9SZXNvdXJjZXMgNDkgMCBSL01lZGlhQm94WzAgMCA3MjAgNTQwXS9H cm91cDw8L1MvVHJhbnNwYXJlbmN5L0NTL0RldmljZVJHQi9JIHRydWU+Pi9Db250ZW50cyAxMSAw IFI+PgplbmRvYmoKCjEzIDAgb2JqCjw8L1R5cGUvUGFnZS9QYXJlbnQgMjIgMCBSL1Jlc291cmNl cyA0OSAwIFIvTWVkaWFCb3hbMCAwIDcyMCA1NDBdL0dyb3VwPDwvUy9UcmFuc3BhcmVuY3kvQ1Mv RGV2aWNlUkdCL0kgdHJ1ZT4+L0NvbnRlbnRzIDE0IDAgUj4+CmVuZG9iagoKMTYgMCBvYmoKPDwv VHlwZS9QYWdlL1BhcmVudCAyMiAwIFIvUmVzb3VyY2VzIDQ5IDAgUi9NZWRpYUJveFswIDAgNzIw IDU0MF0vR3JvdXA8PC9TL1RyYW5zcGFyZW5jeS9DUy9EZXZpY2VSR0IvSSB0cnVlPj4vQ29udGVu dHMgMTcgMCBSPj4KZW5kb2JqCgoxOSAwIG9iago8PC9UeXBlL1BhZ2UvUGFyZW50IDIyIDAgUi9S ZXNvdXJjZXMgNDkgMCBSL01lZGlhQm94WzAgMCA3MjAgNTQwXS9Hcm91cDw8L1MvVHJhbnNwYXJl bmN5L0NTL0RldmljZVJHQi9JIHRydWU+Pi9Db250ZW50cyAyMCAwIFI+PgplbmRvYmoKCjUwIDAg b2JqCjw8L0NvdW50IDcvRmlyc3QgNTEgMCBSL0xhc3QgNTcgMCBSCj4+CmVuZG9iagoKNTEgMCBv YmoKPDwvQ291bnQgMC9UaXRsZTxGRUZGMDA1MDAwRTEwMDY3MDA2OTAwNkUwMDYxMDAyMDAwMzE+ Ci9EZXN0WzEgMCBSL1hZWiAwIDU0MCAwXS9QYXJlbnQgNTAgMCBSL05leHQgNTIgMCBSPj4KZW5k b2JqCgo1MiAwIG9iago8PC9Db3VudCAwL1RpdGxlPEZFRkYwMDUwMDBFMTAwNjcwMDY5MDA2RTAw NjEwMDIwMDAzMj4KL0Rlc3RbNCAwIFIvWFlaIDAgNTQwIDBdL1BhcmVudCA1MCAwIFIvUHJldiA1 MSAwIFIvTmV4dCA1MyAwIFI+PgplbmRvYmoKCjUzIDAgb2JqCjw8L0NvdW50IDAvVGl0bGU8RkVG RjAwNTAwMEUxMDA2NzAwNjkwMDZFMDA2MTAwMjAwMDMzPgovRGVzdFs3IDAgUi9YWVogMCA1NDAg MF0vUGFyZW50IDUwIDAgUi9QcmV2IDUyIDAgUi9OZXh0IDU0IDAgUj4+CmVuZG9iagoKNTQgMCBv YmoKPDwvQ291bnQgMC9UaXRsZTxGRUZGMDA0QjAwNjUwMDc5MDAyMDAwNDQwMDY1MDA3MjAwNjkw MDc2MDA2MTAwNzQwMDY5MDA2RjAwNkU+Ci9EZXN0WzEwIDAgUi9YWVogMCA1NDAgMF0vUGFyZW50 IDUwIDAgUi9QcmV2IDUzIDAgUi9OZXh0IDU1IDAgUj4+CmVuZG9iagoKNTUgMCBvYmoKPDwvQ291 bnQgMC9UaXRsZTxGRUZGMDA1MDAwRTEwMDY3MDA2OTAwNkUwMDYxMDAyMDAwMzU+Ci9EZXN0WzEz IDAgUi9YWVogMCA1NDAgMF0vUGFyZW50IDUwIDAgUi9QcmV2IDU0IDAgUi9OZXh0IDU2IDAgUj4+ CmVuZG9iagoKNTYgMCBvYmoKPDwvQ291bnQgMC9UaXRsZTxGRUZGMDA1MDAwRTEwMDY3MDA2OTAw NkUwMDYxMDAyMDAwMzY+Ci9EZXN0WzE2IDAgUi9YWVogMCA1NDAgMF0vUGFyZW50IDUwIDAgUi9Q cmV2IDU1IDAgUi9OZXh0IDU3IDAgUj4+CmVuZG9iagoKNTcgMCBvYmoKPDwvQ291bnQgMC9UaXRs ZTxGRUZGMDA0QjAwNjUwMDc5MDAyMDAwNDQwMDY1MDA3MjAwNjkwMDc2MDA2MTAwNzQwMDY5MDA2 RjAwNkU+Ci9EZXN0WzE5IDAgUi9YWVogMCA1NDAgMF0vUGFyZW50IDUwIDAgUi9QcmV2IDU2IDAg Uj4+CmVuZG9iagoKMjIgMCBvYmoKPDwvVHlwZS9QYWdlcwovUmVzb3VyY2VzIDQ5IDAgUgovTWVk aWFCb3hbIDAgMCA3MjAgNTQwIF0KL0tpZHNbIDEgMCBSIDQgMCBSIDcgMCBSIDEwIDAgUiAxMyAw IFIgMTYgMCBSIDE5IDAgUiBdCi9Db3VudCA3Pj4KZW5kb2JqCgo1OCAwIG9iago8PC9UeXBlL0Nh dGFsb2cvUGFnZXMgMjIgMCBSCi9PcGVuQWN0aW9uWzEgMCBSIC9YWVogbnVsbCBudWxsIDBdCi9P dXRsaW5lcyA1MCAwIFIKPj4KZW5kb2JqCgo1OSAwIG9iago8PC9DcmVhdG9yPEZFRkYwMDQ5MDA2 RDAwNzAwMDcyMDA2NTAwNzMwMDczPgovUHJvZHVjZXI8RkVGRjAwNEYwMDcwMDA2NTAwNkUwMDRG MDA2NjAwNjYwMDY5MDA2MzAwNjUwMDJFMDA2RjAwNzIwMDY3MDAyMDAwMzMwMDJFMDAzMj4KL0Ny ZWF0aW9uRGF0ZShEOjIwMTAwNDE5MjEyNzMxKzAyJzAwJyk+PgplbmRvYmoKCnhyZWYKMCA2MAow MDAwMDAwMDAwIDY1NTM1IGYgCjAwMDAxMDQ2MDAgMDAwMDAgbiAKMDAwMDAwMDAxOSAwMDAwMCBu IAowMDAwMDAwNTI3IDAwMDAwIG4gCjAwMDAxMDQ3NDQgMDAwMDAgbiAKMDAwMDAwMDU0NyAwMDAw MCBuIAowMDAwMDAwNzYxIDAwMDAwIG4gCjAwMDAxMDQ4ODggMDAwMDAgbiAKMDAwMDAwMDc4MSAw MDAwMCBuIAowMDAwMDE4NTE1IDAwMDAwIG4gCjAwMDAxMDUwMzIgMDAwMDAgbiAKMDAwMDAxODUz NyAwMDAwMCBuIAowMDAwMDE5NjI0IDAwMDAwIG4gCjAwMDAxMDUxNzggMDAwMDAgbiAKMDAwMDAx OTY0NiAwMDAwMCBuIAowMDAwMDE5ODYxIDAwMDAwIG4gCjAwMDAxMDUzMjQgMDAwMDAgbiAKMDAw MDAxOTg4MiAwMDAwMCBuIAowMDAwMDM2MjQ1IDAwMDAwIG4gCjAwMDAxMDU0NzAgMDAwMDAgbiAK MDAwMDAzNjI2OCAwMDAwMCBuIAowMDAwMDM2OTUzIDAwMDAwIG4gCjAwMDAxMDY2NTkgMDAwMDAg biAKMDAwMDAzNjk3NCAwMDAwMCBuIAowMDAwMDQyOTUxIDAwMDAwIG4gCjAwMDAwNDI5NzMgMDAw MDAgbiAKMDAwMDA0MzE2MyAwMDAwMCBuIAowMDAwMDQzNDYxIDAwMDAwIG4gCjAwMDAwNDM2MjMg MDAwMDAgbiAKMDAwMDA2MzcwNCAwMDAwMCBuIAowMDAwMDYzNzI3IDAwMDAwIG4gCjAwMDAwNjM5 MTUgMDAwMDAgbiAKMDAwMDA2NDUxNCAwMDAwMCBuIAowMDAwMDY0OTUzIDAwMDAwIG4gCjAwMDAw NzM1MTcgMDAwMDAgbiAKMDAwMDA3MzUzOSAwMDAwMCBuIAowMDAwMDczNzM3IDAwMDAwIG4gCjAw MDAwNzQwOTEgMDAwMDAgbiAKMDAwMDA3NDMwNSAwMDAwMCBuIAowMDAwMDg1MDE5IDAwMDAwIG4g CjAwMDAwODUwNDIgMDAwMDAgbiAKMDAwMDA4NTIzNSAwMDAwMCBuIAowMDAwMDg1NjUwIDAwMDAw IG4gCjAwMDAwODU5MTQgMDAwMDAgbiAKMDAwMDEwMzQxMyAwMDAwMCBuIAowMDAwMTAzNDM2IDAw MDAwIG4gCjAwMDAxMDM2MzEgMDAwMDAgbiAKMDAwMDEwNDEyOCAwMDAwMCBuIAowMDAwMTA0NDcy IDAwMDAwIG4gCjAwMDAxMDQ1NDUgMDAwMDAgbiAKMDAwMDEwNTYxNiAwMDAwMCBuIAowMDAwMTA1 NjcyIDAwMDAwIG4gCjAwMDAxMDU3OTcgMDAwMDAgbiAKMDAwMDEwNTkzNCAwMDAwMCBuIAowMDAw MTA2MDcxIDAwMDAwIG4gCjAwMDAxMDYyMzMgMDAwMDAgbiAKMDAwMDEwNjM3MSAwMDAwMCBuIAow MDAwMTA2NTA5IDAwMDAwIG4gCjAwMDAxMDY3OTkgMDAwMDAgbiAKMDAwMDEwNjkwMSAwMDAwMCBu IAp0cmFpbGVyCjw8L1NpemUgNjAvUm9vdCA1OCAwIFIKL0luZm8gNTkgMCBSCi9JRCBbIDxBMjI2 MTFFN0ZBMTZGOTNEODIxMjVBRTE2RDE3Mjc0NT4KPEEyMjYxMUU3RkExNkY5M0Q4MjEyNUFFMTZE MTcyNzQ1PiBdCi9Eb2NDaGVja3N1bSAvRERDN0FCREQwMkZGMUVFQTFGN0RGMjEwNDlERkUwRjIK Pj4Kc3RhcnR4cmVmCjEwNzA5MgolJUVPRgo= --Apple-Mail-4--807036317 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii ------------------------------------------------------- Rafael Marin Lopez, PhD Dept. Information and Communications Engineering (DIIC) Faculty of Computer Science-University of Murcia 30100 Murcia - Spain Telf: +34868888501 Fax: +34868884151 e-mail: rafa@um.es ------------------------------------------------------- --Apple-Mail-4--807036317-- Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Tue, 6 Apr 2010 17:19:45 +0200 From: Alan DeKok Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Message-ID: <4BBB5111.4000904@deployingradius.com> Sam Hartman wrote: > There is an additional concern. At this point in the process, we have > not learned the identity of the peer. I don't know the full details of the IEEE EAPoL state machine, but IIRC, there are provisions for an Access-Point to send an EAP identity request. The peer should then respond with an identity that contains a realm. > So, we don't even know what EAP > server to contact. So, even ignoring the credentials matching to method > issue you describe and the problem that RADIUS does not provide the > information we need described in the paper, we're absolutely stopped by > the fact that we don't know what party to contact in order to find out > what EAP methods might be available. If WiFi roaming groups can use EAP, I don't see why there are any technical problems here. Alan DeKok. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Tue, 6 Apr 2010 11:49:36 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "Alan" == Alan DeKok writes: Alan> Sam Hartman wrote: >> There is an additional concern. At this point in the process, we >> have not learned the identity of the peer. Alan> I don't know the full details of the IEEE EAPoL state Alan> machine, but IIRC, there are provisions for an Access-Point to Alan> send an EAP identity request. The peer should then respond Alan> with an identity that contains a realm. At this point in the GSS exchange, we've not committed to using EAP. At the GSS layer, you can't commit to a particular way of describing the initiator (peer)'s identity until you've committed to a particular mechanism. You're right though that there are no technical problems here. The analysis document proposes simply that we need to rethink how GSS deals with negotiation. The draft GSS mechanism solves this issue by requiring the client to commit to using EAP before deciding what EAP method to use. At that point, all these issues go away, and the typical EAP solutions apply. As I've discussed this issue with members of the GSS community, they tend to agree that while GSS normally discourages multi-layer negotiation, that performing a multi-layer negotiation is exactly the right approach here. One reason advanced for this is that the negotiations are with different parties. The initiator and acceptor (NAS) decide to use EAP; the peer (initiator) and EAP server decide what EAP method to use. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Tue, 6 Apr 2010 09:27:18 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "David" == David Mitton writes: David> Sam, Let me start by saying I do believe you have an issue David> here. David> My original objection was to how the problem was described David> "the NAS needs to know what EAP method to use" Even given David> some situational license, this is not a NAS problem in David> general, but with your NAS application. (and my comment David> about layering wasn't meant to be a suggestion, but a warning David> why that would be a bad idea) Here we're agreed. All that paragraph is saying is that to implement GSS-API style negotiation, where before anything else happens, the acceptor provides a set of mechanisms, the acceptor needs to know what mechanisms to provide. Since that part of the acceptor inherently has to be the NAS, then the NAS would need to know the EAP methods to implement that strategy. For a variety of reasons, we both agree that will not work. Some of the reasons are stated in the rest of the paragraph in question. Your concern about making sure the credentials are appropriate is also a reason that things might fail. There is an additional concern. At this point in the process, we have not learned the identity of the peer. So, we don't even know what EAP server to contact. So, even ignoring the credentials matching to method issue you describe and the problem that RADIUS does not provide the information we need described in the paper, we're absolutely stopped by the fact that we don't know what party to contact in order to find out what EAP methods might be available. I think you and I are agreed that providing a list of EAP methods to the peer from the NAS before identity information is exchanged is not a valid technical approach. --Sam Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Mon, 5 Apr 2010 22:44:31 -0400 From: David Mitton Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method Content-Type: text/plain; charset="us-ascii"; format=flowed Message-ID: <20100406024453.8D0031617623@smtprelay03.hostedemail.com> Sam, Let me start by saying I do believe you have an issue here. My original objection was to how the problem was described "the NAS needs to know what EAP method to use" Even given some situational license, this is not a NAS problem in general, but with your NAS application. (and my comment about layering wasn't meant to be a suggestion, but a warning why that would be a bad idea) But if I back up a second, I don't believe that is true either. If your NAS did the appropriate thing by the EAP/AAA architecture, it would simply pass through the EAP method negotiation to the client. And it would be a negotiation between the EAP supplicant/client(peer) and EAP server(authenticator) where the method selection would get resolved. The NAS is a transparent man-in-the-middle. I have no experience with GSS-API, so I cannot directly evaluate how this differs from "the traditional GSS-API solution", but I can guess you have an issue, where you want to present a dynamic list of known EAP methods without prior knowledge. Even if you could get the active EAP methods, you have another problem, in that just because an EAP method can be serviced on a given AAA authenticator, it doesn't mean it will support your credentials for the given user. For example, my typical RADIUS server setup will have configured EAP methods: PEAP w/MS-CHAP and EAP-POTP (RSA tokens), but if use POTP and your account is not in the RSA authentication server database, you will fail. Typically, for remote access, it's a matter of knowing what auth protocol you have to use to succeed. The server itself supports a number of user groups, methods, and credentials for various situations. Dave. On 4/1/2010 04:47 PM, Sam Hartman wrote: > >>>>> "David" == David Mitton writes: > > David> Since there isn't an author attribution on this document, I'm > David> not sure whom to direct this too.... > > David> In section 6.1, Negotiating the EAP Method, the statement is > David> made that "the NAS needs to know what EAP method to use." > David> This is incorrect. The NAS does not, in fact it really > David> should not, it would break layering. > > >That statement is made in a paragraph explaining why negotiating the EAP >method to use at the GSS-API layer would be problematic in an EAP/RADIUS >environment. In context, in order to negotiate which eap method to use >at the GSS-API layer, the NAS needs to know which EAP method to use so >that it can advertize it in the GSS-API OID. The paragraph then goes on >to explain some of the reasons why this is impossible to implement with >the current RADIUS protocol. > >The analysis was never contemplating having the NAS inspect EAP >messages. The text suggests extending RADIUS to advertize a set of EAP >type code before EAP starts. It points out that doesn't actually work, >although it misses some of the critical reasons it doesn't work. (It >does include enough to demonstrate that approach is problematic though). > > You're absolutely right that having a NAS >inspect the type codes of EAP messages, or restricting the set of EAP >methods the NAS can work with would defeat the entire point of making >the NAS EAP-method agnostic. > >The conclusion that section then draws is that the traditional GSS-API >solution to the multi-layer negotiation problem is problematic in an >EAP/GSS environment. > >The specification (draft-howlett-eap-gss) discusses this issue in more >detail and concludes that multi-layer negotiation is required. Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Thu, 1 Apr 2010 20:16:43 +0000 From: David Mitton Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method Content-Transfer-Encoding: 7bit Message-ID: <797367747.69273.1270153003260.JavaMail.mail@webmail09>
Since there isn't an author attribution on this document, I'm not sure whom to direct this too....
 
In section 6.1, Negotiating the EAP Method,  the statement is made that "the NAS needs to know what EAP method to use."  
This is incorrect.  The NAS does not, in fact it really should not, it would break layering.
 
Think about current RADIUS/EAP authenticators, e.g.  Wireless APs, Dial-in servers, or VPN Servers, 
they do not know or care whether PEAP/MS-CHAPV2 or EAP-LAST, or EAP-POTP is going to be used for a session authentication.
They must know how to encapsulate an EAP message in RADIUS and forward it.
The RADIUS server decapsulates the frame and directs it to the appropriate EAP authenticator. 
If said EAP mechanism is not installed, or the response is an EAP NAK,  the NAS will just receive such an EAP message in
an Access-Challenge and pass it back to the client.  If the client knows another method, it's free to try again, within the current session.
 
The NAS depends on the RADIUS Server /EAP Authenticator interface to deliver it what it needs to know via RADIUS attributes.
It should not be inspecting the contents of the EAP messages for type related information.
Not only does that break layering, but it imposes yet another dependacy on the EAP methods that can be deployed.
 
If this didn't work, then no new EAP methods would work until all NASes are upgraded.
Trust me there are enough problems in the deployed implementations already.
 
Dave.
Return-Path: Received: (from root@localhost) by example.com (8.11.6/8.11.6/SuSE Linux 0.5) id h4CMt1U02501 for root; Mon, 12 May 2003 17:55:01 -0500 Date: Thu, 1 Apr 2010 16:47:19 -0400 From: Sam Hartman Message-Id: <200305122255.h4CMt1U02501@example.com> To: RANDOM_USER@example.com Subject: Re: Feasibility Analysis, Negotiating an EAP method MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: >>>>> "David" == David Mitton writes: David> Since there isn't an author attribution on this document, I'm David> not sure whom to direct this too.... David> In section 6.1, Negotiating the EAP Method, the statement is David> made that "the NAS needs to know what EAP method to use." David> This is incorrect. The NAS does not, in fact it really David> should not, it would break layering. That statement is made in a paragraph explaining why negotiating the EAP method to use at the GSS-API layer would be problematic in an EAP/RADIUS environment. In context, in order to negotiate which eap method to use at the GSS-API layer, the NAS needs to know which EAP method to use so that it can advertize it in the GSS-API OID. The paragraph then goes on to explain some of the reasons why this is impossible to implement with the current RADIUS protocol. The analysis was never contemplating having the NAS inspect EAP messages. The text suggests extending RADIUS to advertize a set of EAP type code before EAP starts. It points out that doesn't actually work, although it misses some of the critical reasons it doesn't work. (It does include enough to demonstrate that approach is problematic though). You're absolutely right that having a NAS inspect the type codes of EAP messages, or restricting the set of EAP methods the NAS can work with would defeat the entire point of making the NAS EAP-method agnostic. The conclusion that section then draws is that the traditional GSS-API solution to the multi-layer negotiation problem is problematic in an EAP/GSS environment. The specification (draft-howlett-eap-gss) discusses this issue in more detail and concludes that multi-layer negotiation is required.