[EAI] Minutes from Stockholm

[Harald Alvestrand <harald at alvestrand.no>]

In a combination of distraction and confusion, the fact that the EAI 
minutes were embedded in the Jabber log escaped the chairs' attention 
for a while, and the minutes were thus severely delayed.

The attached is the first draft of a minutes-format minutes.

Please comment

             Harald

Email Address Internationalization (EAI) WG minutes

Meeting: IETF75, Monday, July 27, 1740-1940
Place: Room 300, City Conference Center, Stockholm, Sweden
Chair: Harald Alvestrand <harald at alvestrand.no>, Xiaodong Lee <lee at cnnic.cn>
Minutes: Chris Newman <Chris.Newman at sun.com>
Version: 1.0
=====================================================
1740 - Scribe, blue sheet, agenda bashing

 Chris Newman volunteered as scribe.

1750 - Documents in Last Call
  draft-ietf-eai-imap-utf8-07
  draft-ietf-eai-pop-06
  draft-ietf-eai-downgraded-display-01


IMAP document:
 Alexey: reviewed with 1 issue: lack of SASLprep
 What to do with invalid UTF-8 and strings that fail SASLprep
 This is section talking about usernames
 Pete: fine. will just change this.
 John Klensin: concerned about reliance on stringprep as we're moving
   away from that in other way.
 Chris: suggest indirect reference through SASL PLAIN.
 Pete: has anyone implemented this IMAP spec?
 room: no hands
 Pete: this is a concern, but OK for experimental status.

POP document:
 Alexey issues: language range reference needs to be clarified to be
    "basic language range" from the lang-range RFC.
   Issue 2: comments about SASLprep for APOP and USER/PASS.  Same as IMAP.
   Issue 3: inexact message sizes issue.
 Randy: message size will be inexact with UTF-8 mailstore when it has to
   downgrade.  Clients that require reliable sizes may be broken by this change.
 Joseph Yee (in chatroom): rfc1939 (pop3) is concern on list, not on stat
   (if I recall correctly) regards to size
 Conclusion: need more text describing the issue of message sizes.
 Barry:  may want to support multiple usernames but best to describe that
   in framework or usability document

Downgraded display:
 No issues raised on downgraded-display.

1810 - Deployment experience
  What implementations are done, what tests have been done,
  what has been deployed in (test) production?
  What areas of trouble did arise?
  How should we document this experience?

  draft-yao-eai-deployment-03
  draft-yee-eai-pop3-scenarios-00
  draft-ietf-eai-email-clients-00

 Discussion of deployment experience.
 Problems talking to non-EAI mail servers, downgrading raises same phishing
 problem with non-downgrading.

 Joseph Yee confirmed that 30-40 participants had registered in the testing
   platform

 John: Concerned whether downgrade is sufficiently interoperable and proven
   to overcome significant disadvantages of new email syntax.
   Long discussion of interoperability concerns with downgrading.
 Ted: what use case is a concern?
 John: concerned with environments where anti-spam games based on behavior.
   If we're introducing distortions in such tests we should at least
   understand them.
 Ted: concerned about feature-creep for how downgrade interacts with
   non-standard systems delaying work.
 joseph.yee (in chat): On the 'business card' example, users type only one
   address, not the compound, downgrade's big main purpose is for the sender
   address
 John: concerned that downgrade is the high risk area.  Particularly the
   <u <a>> syntax. Draft standard test will involve testing this against
   broken mail related stuff in the wild.
 Dave Crocker: we really believe we do not understand all of the problems
   this may cause and would like to understand them better before promoting
   this.
 Ted: charter says we can go forward with downgrade as informational
   Charter does require mailing list problems be considered before advancing
 Barry: thought double-address would be transitional initially.
 Harald: may not be very transitional. We need to have confidence in
   downgrade or willingness to set aside downgrade before moving the set
   to standards track.

Moving on to mailto URL issue (19:03)
  
 Discussion of schemes to do mailto iri and uri.
 Is mailto making progress?
 Alexey has encouraged document author
 Ted: if we do a new IRI scheme, this could be non-blocking.  But if we do IRI
   form of mailto then mailto-bis is blocking. Suggest a conf call gets set up
   with dedicated time of experts such as Leslie and Pete.
 John: also interested in this call.
 Alexey will collect together assorted experts to talk with Martin about
   mailto-bis.

Mailing list draft:
 Draft has expired.
 Randy: No, nothing has been worked on. Major issue in mailing list is the
   mailto: issue. The rest of the mailto issue is handwaving about ideal
   solutions.
 Chris: I volunteer to take EAI issue discussion to my engineering team to
   provide some feedback on the problems.

Conclusions on drafts:
   documents in last call: imap & pop have small details.
   downgrade-display has no small details.
   need more input on list that documents have been reviewed.
   Will pass these 3 to Alexey hopefully relatively soon.
   Need shepherd write-ups for these from WG chairs or their delegates.
 Deployment experience: No concrete actions.

Progressing on standards track:
  We do want downgrade experience and experience with what happens in the
  wild with downgrade.
  Who's doing testing, what are results?
  People who volunteer are extremely welcome.

  Mailto: AD action to figure out where it is.
  Mailing lists: Randy action to refresh document.
    Chris action to discuss with team to determine problems.

The WG concludes at 19:28.