Email Address Internationalization (EAI) WG minutes Meeting: IETF72, Monday, July 28, 1740-1950 Place: Newcastle room, Citywest hotel, Dublin Chair: Harald Alvestrand , Xiaodong Lee Minutes: Andrew Sullivan Version: 1.0 ======================================================================== 1. Administrivia Scribes appointed, agenda modified. 2. Status of core documents Documents were approved by the IESG and are in the RFC Editor's queue. 3. Review of Last Call comments and the last version of -downgrade There were no comments received. There was one issue from Philadelphia about -list-* headers. There's been no discussion on the list so far. Barry Lieba: for most list headers, the answer would be to allow multiple URIs. This could be dealt with by putting both in. The List-ID header is supposed to be a machine readable thing for machine consumptions, so we can say it doesn't meed to be internationalised and skirt the problem that way. Randall Gellens: propose to revise list-* documents to say, "You can do UTF8 in any context where you otherwise normally could." The Chairs asked whether this creates a normative dependency. Randall Gellens: no, informative. Pete Resnick: List-IDs look mostly like domain names. Does that mean this is Punycode? John Klensin: The text in the IDNA documents says, "If it's not expected to be a protocol element, then it's an A-label." That means you'll see Punycode there. Dave Crocker: I looked at List-ID recently and they're not required to be domain names. They're just required to be unique. Pete Resnick: The uniqueness is often generated partly by domain names. A-labels are fine. Cyrus Daboo: I like to use List-ID in Sieve, which means that Punycode will have to be typed in. The Chairs expressed a worry that this will introduce a new downgrade mechanism. There was no proposal yet, so this issue was referred to the list. An alternative was suggested: if 8 but characters appear in the list- headers, just downgrade. Barry Lieba opposed to that. People use list- headers in filtering, so getting rid of them would be bad. The Chairs told the room that they can't send -downgrade- to the IESG until the issue is resolved. 4. eai-downgraded-display The Chairs noted that this document was to be added under the new Charter, but the new Charter isn't ready yet. They asked the Area Director whether it could be added without the Charter changes. The AD gave a free hand to the Chairs. 5. pop and imap pop and its relationship to downgrade were discussed in Philadelphia, and there were no subsequent remarks on list. Randall Gellens discussed proposed new text. There is a new issue after the most recent revision: should the language listing have the human-readable portion of the language name in the language thereby described, or in some other language? Barry Lieba: I don't care, but why say anything on this topic? Cyrus Daboo: Do we want test in the -pop- document about how the client is storing the data? Randall Gellens: No, just one additional justification. Cyrus Daboo: The -imap- document should probably say something similar about cached, downloaded documents too. Randall Gellens asked Pete Resnick about how UTF-8 messages are downgraded in -imap-? The answer: it's handled as encoding, because IMAP requires all ASCII. *** The Chairs asked for a hum on whether the language name should be expressed in the language so named. The hum indicated agreement. 6. mailinglist The issue of the relationship to UTF-8 mailto: URIs was raised. Barry Lieba: The definition ot UTF-8 mailto: belongs in its own document. He expressed ambivalence about the ability to move mailinglist ahead without the problem being addressed. Randall Gellens: the mailinglist document will not define the UTF-8 mailto: He was leery of suggesting that the WG take on such a significant new document. On the other hand, work will not get done without the UTF-8 mailto: Chris Newman put on his tech hat, and said that he thought it was ok to work on this because without it the specification is not complete. He then put his AD hat on, and noted that while he would not have trouble with the WG attacking the problem, there are several thorny issues in the IRI space, and so such an effort might be troublesome.m He left the decision to the discretion of the WG. Ted Hardie suggested there are two approaches: - offer a UTF-8 scheme for an IRI scheme for mailto: - take on the IRI form Since there is open IRI work, the latter at least won't happen until the IRI work is done. He noted that would be a blocker, so the WG should not waste time. The Area Director noted that the latter is good advice. The Chairs asked Randall whether mailinglist is blocked Randall Gellens: there's currently a lot of hand-waving in mailinglist about this topic. Perhaps the answer is to wave more? Barry Lieba: Just hand-waving is fine as long as this work is experimental. 7. Status of implementations and implementations advice. See meeting materials and info at http://cslab.kr/eai-test/index.php/Main_Page. The Chairs noted that draft-dainow-eai-email-clients was to have been adopted by the working group, according to the Philadelphia meeting. Since the AD earlier allowed adding a document without the recharter being complete, this draft is also added. If you have implementation experience, it would be helpful if you sent text about it. Claudia Galvan: is there any advice to give for interactions with IDNA, particularly with respect to bidi in the local-part of the email? Derek Williams said he wasn't sure, but he'd send a note to the list. Claudia Galvan: in the documents, there are some parts that include both local-part and global-part. Is there a dependency? [scribe's note: Noise seemed to be "no", but no clear conclusion.] 8. Future new work, next steps, new charter The Chairs reported that they need to draft a new Charter. Given the current state of work, finishing by the end of 2008 seems too optimistic. Pete Resnick (pop) had received only one comment, all editorial. He would like comments from implementors. Derek Williams asked whether anyone else thinks an LMTP document is needed. Alexey Melnikov said yes. When SMTP was published, he thought that LMTP could have been included. Perhaps the best approach is to catch this when moving to standards track. Also, on another note, the sieve working group is rechartering, and eai-related things are planned for the new charter; therefore, people were encouraged to go to the sieve meeting. 9. Summary and action items - downgrade: one item to go to list - imap needs more comments - pop needs the language issue addressed - mailinglist: additional handwaving needed [scribe's note: this may be the first time that is an explicit request from a WG to an editor!] about future UTF-8 mailto: - Charter: Chairs to write and propose to working group.