[lemonade] concerns with draft-ietf-lemonade-reconnect-client-04.txt

Chris Newman <Chris.Newman@Sun.COM> Wed, 16 May 2007 07:28 UTC

Return-path: <lemonade-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com) by megatron.ietf.org with esmtp (Exim 4.43) id 1HoDvh-0001qn-Uc; Wed, 16 May 2007 03:28:13 -0400
Received: from lemonade by megatron.ietf.org with local (Exim 4.43) id 1HoDvh-0001qh-AH for lemonade-confirm+ok@megatron.ietf.org; Wed, 16 May 2007 03:28:13 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org) by megatron.ietf.org with esmtp (Exim 4.43) id 1HoDvd-0001q8-Rt for lemonade@ietf.org; Wed, 16 May 2007 03:28:09 -0400
Received: from brmea-mail-4.sun.com ([192.18.98.36]) by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HoDvd-0005JA-Gn for lemonade@ietf.org; Wed, 16 May 2007 03:28:09 -0400
Received: from fe-amer-05.sun.com ([192.18.108.179]) by brmea-mail-4.sun.com (8.13.6+Sun/8.12.9) with ESMTP id l4G7S8Gi021654 for <lemonade@ietf.org>; Wed, 16 May 2007 07:28:09 GMT
Received: from conversion-daemon.mail-amer.sun.com by mail-amer.sun.com (Sun Java System Messaging Server 6.2-6.01 (built Apr 3 2006)) id <0JI400B01HWHAA00@mail-amer.sun.com> (original mail from Chris.Newman@Sun.COM) for lemonade@ietf.org; Wed, 16 May 2007 01:28:08 -0600 (MDT)
Received: from [10.1.110.5] (216-165-236-126.championbroadband.com [216.165.236.126]) by mail-amer.sun.com (Sun Java System Messaging Server 6.2-6.01 (built Apr 3 2006)) with ESMTPSA id <0JI400GNNI2TLM00@mail-amer.sun.com> for lemonade@ietf.org; Wed, 16 May 2007 01:28:08 -0600 (MDT)
Date: Wed, 16 May 2007 00:28:06 -0700
From: Chris Newman <Chris.Newman@Sun.COM>
To: lemonade@ietf.org
Message-id: <7CB41CC1AD859AC968313FAB@[10.0.1.21]>
MIME-version: 1.0
X-Mailer: Mulberry/3.1.6 (Mac OS X)
Content-type: text/plain; format="flowed"; charset="us-ascii"
Content-transfer-encoding: 7bit
Content-disposition: inline
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 82c9bddb247d9ba4471160a9a865a5f3
Subject: [lemonade] concerns with draft-ietf-lemonade-reconnect-client-04.txt
X-BeenThere: lemonade@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Enhancements to Internet email to support diverse service enivronments <lemonade.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/lemonade>, <mailto:lemonade-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:lemonade@ietf.org>
List-Help: <mailto:lemonade-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/lemonade>, <mailto:lemonade-request@ietf.org?subject=subscribe>
Errors-To: lemonade-bounces@ietf.org

I'm concerned there are architectural issues with 
draft-ietf-lemonade-reconnect-client-04.txt that may create problems with IETF 
consensus on the document.  I wanted to verify that the WG has consensus on 
these specific architectural issues in addition to the document as a whole 
before I start IETF last call, as it may be more difficult to resolve these 
issues in an IETF-wide or IESG scope than in a more narrow WG scope.  My 
intention is not to force any specific changes, but rather to encourage making 
changes prior to IETF last call that might be helpful to speed IETF consensus.

1. Implicit server behavior changes

This extension has several cases where client commands trigger a server 
behavior change as a side-effect of a client command.  There has been 
significant discussion in other IETF contexts (EAI design team, IMAPEXT, ENABLE 
extension) that such behavior is undesirable and that if a server behavior 
change is to occur it should be explicit and early in the protocol exchange.  I 
would like to confirm the WG has consensus on this specific issue as I suspect 
it is likely be problematic during IETF last call.

2. Change in IMAP untagged response model

For better or worse, IMAP is a cache update protocol model rather than a 
request/response protocol model.  That means an untagged response is not bound 
to a particular command, but rather the portion of the client's cache state 
that is updated can be determined solely from the untagged response itself (the 
SEARCH response being a notable exception).  This extension changes that model 
by binding the VANISHED response to a specific command with the TAG correlator. 
Are there servers which execute client commands out-of-order so that such 
binding would be necessary?  Does the WG believe changing the fundamental IMAP 
model in this case is the best way to accomplish this specific task?

3.  VANISHED response sometimes but not always implying EXISTS change

The VANISHED response is used in two different ways: one to replace the EXPUNGE 
response to indicate a change in mailbox state at that moment (including an 
implicit EXISTS update), and one to indicate that messages were removed before 
the mailbox was selected in response to a QRESYNC select modifier or VANISHED 
UID FETCH modifier.  I suspect the logic the client must use to distinguish 
these cases is a bit subtle.  Perhaps a simplification to VANISHED to address 
issue 2 & 3 should be considered?

4. Complex interaction with CONDSTORE and UIDPLUS

This extension requires CONDSTORE and requires the server to advertise 
CONDSTORE.  That means there's a silly state where the server advertises 
QRESYNC but not CONDSTORE.  Perhaps there's a simple way to eliminate the silly 
state?  Furthermore, there is a great deal of variation in server behaviors and 
client codepaths between un-extended IMAP, IMAP + CONDSTORE, IMAP + CONDSTORE + 
QRESYNC and IMAP + CONDSTORE + QRESYNC + UIDPLUS.  Is the WG really comfortable 
with this complexity?  Perhaps QRESYNC should require UIDPLUS in addition to 
CONDSTORE?  Perhaps QRESYNC is really a refinement of CONDSTORE rather than an 
extension to it, so perhaps QRESYNC should subsume CONDSTORE and obsolete it?

                - Chris Newman
                Applications Area Director



_______________________________________________
lemonade mailing list
lemonade@ietf.org
https://www1.ietf.org/mailman/listinfo/lemonade
Supplemental Web Site:
http://www.standardstrack.com/ietf/lemonade