Minutes from the EAI working group Chicago, IL, USA 2007-07-26 1. Agenda bashing, etc: Nothing to report. 2. Status report: a. RFC 4952 published 2007-07-26. (Jubilation) b. SMTP believed to be ready except for MIME type choice. c. UTF8HDR has open issues, but believed to be ready except for MIME type choice. d. DSN is a dependency with issues. e. DOWNGRADE is in a reasonable state, but there are issues to discuss (in particular, "unknown header" handling, multiple Downgrade: headers, and downgraded no-Alt cases) f. Other drafts are dependent on completion of core. Remark (John Klensin): There were late comments on the framework document that suggested it could be more clear. The design team is aware of these, and will be incorporated in later revisions; but if you had comments or think the document as published is not good, send comments _now_. 3. SMTP a. List of issues closed. A question was put to the room: Does anyone object to these issues being regarded as resolved? There was no response. b. Unresolved issues requested. There was no response from the room. A question was put to the room: How many people have read the document? About one half of the room responded that they had read it. A question was put to the room: Who believes this document is ready for Last Call? About one half the room indicated agreement. A question was put to the room: Who believes this document is not ready for Last Call? Nobody responded. 4. UTF8HDR a. The Chair listed the issues closed. A question was put to the room: Does anyone object to considering these issues closed? There was no disagreement expressed. b. The remaining issue is issue 1485: body part for UTF8SMTP. A question was put to the room: Are there any other issues for this document? There were no additional issues raised. A question was put to the room: How many people have read the document? There were approximately 10 people responding. A question was put to the room: Once the MIME type is selected, is this document ready for Last Call? Nobody objected. John Klensin made a remark that it would be worrisome if this document were Last Called prior to the DSN document; the Chair replied that they would be Last Called together, because there is a normative dependency between them. c. MIME type selection [minute note: out of order from slides; the MIME type selection discussion happened before DSN discussion]. The suggestion from the Chair is that there are some bad options that have been eliminated, but that there is nevertheless not a remaining clear winner. The issue is that the WG needs to pick -- not design -- a method. Several methods suggested. John Klensin remarked that it is important to find "least liked" options, but that nobody has ever died because of a bad decision on similar topics. The Chair asked John Klensin to send an outline to the list of what "very bad" options were, and Klensin agreed. A coin toss established the Condorcet method as the way by which the choice would be made. 5. DSN open issues: report by Alexey Melnikov a. Issue 1 is just MIME type issue. See above. b. If DSN cannot be delivered on the next hop, then what? there are three options: 7 bit, downgrade, or discard. Eric Allman observed that it is important not to make downgrade an implicit requirement, because it's possible not all implementations will do downgrade. Some remarks to the effect that detailed design in the meeting is not a good idea. After considerable discussion, the Chair asked for a hum on the suggestion that downgrading was off the table, but that some versions of encoding are possible. There was a noticeable response in favour, and silence opposed. 6. Downgrade unresolved issues a. Chair asked how many had read the latest version? There were not many indicating having read. b. There remains an issue with UTF8 addresses with no alternate. There were previous suggestions merely to bounce the mail, but that seemed unsatisfactory, which is why there has been a suggestion for group syntax. The editor asked for feedback. General sense seemed to be that MUAs would not mishandle empty groups, and that the important thing was not to require too complete behaviour lest the feature be completely unimplemented. c. Unknown header fields discussion. There seem to be two options: either a field is processed as Downgrade: [old field], or it gets a special Downgraded-* prefix for every field. The room seemed to converge on documenting the problem and moving on, but no clear statement seemed to garner clear support. the Area Director pointed out that, to the extent this issue impinged upon DKIM, it would need to be addressed, though he did not have a recommendation on how to proceed. The Chair observed that Downgrade: fieldname: value received no remarks in favour, but Downgrade-fieldname: value was preferred by 7 participants. 7. POP: Randall Gellens a. Some meeting participants indicated they'd read the document recently. b. It appears that a full MIME parser is required in POP servers depending on which commands the client has issued (and whether the message is 7 bit or has been downgraded). A large number of slimy, worm-like issues are under this rock. Discussion needs to go to the list. 8. Mailinglist a. Nobody replies to question to the room: Who has read the latest version? b. A number of issues outlined, with not much discussion. Must go to list. 9. Working Group timeline a. The Chair plans to issue a 2-week Working Group Last Call on SMTP, UTF8, and DSN in August. If that goes well, it should be possible to go to the IESG in September. WGLC for Downgrade in September, going to the IESG in October. Then remaining additional drafts can be discussed in the December meeting in Vancouver. b. Given that the entire work is to be Experimental, how can the results of the experiment be collected, and how can the results be documented? 10. Closing a. Actions to be taken: MIME decision ASAP, fix drafts after that in August, and WGLC after that. b. Other business? None. Respectfully submitted, Andrew Sullivan