L2/02-251 Title: WG20 Meeting Summary Source: Ken Whistler Date: June 19, 2002 Unicadetti, Here is a short -- well, kinda long -- summary of what took place at the meeting last week of WG20 in Tromsø, Norway. The attendees of the meeting (June 10 - 12) were: Arnold Winkler -- Convenor John Hill -- SC22 chair (first day only) U.S. -- myself Canada -- Alain La Bonté Ireland -- Michael Everson, Christian Cooke Germany -- Marc Küster Norway -- Keld Simonsen S. Korea -- Kim Kyongsok U.K. -- Pat Hall This was actually the best-attended WG20 meeting in quite awhile, despite its location well north of the Arctic Circle! The attendees enjoyed absolutely beautiful weather, along with the midnight sun. The local host was Trond Trosterud, who did a fine job setting things up and shepherding us around. Following is a summary of the decisions made regarding the various technical work at the meeting, --Ken ************************************************************ 1. Amd 1 to IS 14651 -- International String Ordering The disposition of comments on this document went fairly well, but with some interesting twists. The comments from Norway were all turned down. These were requests to: a. align with IS 12199 and CEN/ENV 13710 EOR for some unspecified Latin letters (which turned out to be a-ring -- big surprise -- and Polish l). This was turned down in the interest of stability of deltas -- including the delta for EOR (European Ordering Rules) -- and because it was pointed out that there was agreement that those standards align with 14651, rather than the other way around. b. several comments having to do with the tailorable template needing to "specify a fully deterministic ordering". This was turned down as the result of a misunderstanding of what the goal of the table is. Some bugs reported by Sweden were accepted to be fixed. The more complicated suggestion to do syllabic sorting for *all* Brahmi-based scripts was turned down as underspecified and probably unnecessary as part of the basic table. The first U.S. comment (secondary weight fix for script numerics bug) was accepted. Surprisingly, however, the fix for Hangul syllable ordering was rejected. This because Kim, there representing ROK, was adamantly opposed to it. He admitted that it would fix a problem in the current table, but he was concerned about possible side effects. Even after extensively chalktalking this, and explaining that the "side effects" that he was most immediately worried about are already present with the current treatment, he was still uneasy about the idea of splitting the jamo weighting into two groups -- low for the initials and high for the medials and finals. We will need to review this in the UTC and figure out what to do. Presumably some offline but detailed discussion with Kim, giving him the option to discuss it over with the Korean NB and present objections, might work best to move this forward -- but it definitely won't be in the current table for the Amendment. The disposition of comments document itself is non-problematical, and should not need further review. No significant changes will be made to the text ballotted for FPDAM. I will deliver the updated table to the editor by July 15, and the Amendment will go out for its DAM vote. (No further technical changes -- just an up-or-down yes/no vote.) ************************************************************ 2. Future amendments to IS 14651 There was consensus that it will soon be time to push ahead with the next amendment. I set the expectation that the next amendment should not settle on the Unicode 3.2 repertoire, but instead aim up for the complete Unicode 4.0 repertoire. That might make it possible for us to finally synch 14651 and 10646 at a significant synch point which also matches our publication plans for Unicode. We should come in with the draft table for the Unicode 4.0 repertoire at the next WG20 meeting. That will significantly grease the skids on this. The next meeting could then issue a new project subdivision and send out the PDAM for ballot at the same time. ************************************************************ 3. DTR 14652 -- Cultural Convention Specification As predicted, this was enormously contentious again, with the committee split right down the middle along foreseeable lines. The U.S. provided the bulk of the comments on the ballot, and again, many of the comments were rejected. The strategy I took in the face of this was to push adamantly for the cosmetic fixes that will further indicate the controversial status of the TR, but not to press for technical fixes. This approach allowed us to not cannibalize the committee time needed for the rest of the projects, and left the DTR text little changed technically. The *big* fix was in response to Germany's main concern. We *finally* got the entire Conformance clause removed from the text -- not just marked as controversial. This will render the final document much less likely to mean anything for procurement purposes, since there is no longer anything present that one could claim conformance to, even if the front page was saying it is a TR rather than a standard. With that out of the way, Germany agreed with the publication of the existing text -- although they still indicated their strong displeasure with much of it. Essentially, the committee decided that the thing is now sufficiently branded as problematical that it can simply be published, and we won't have to go the painful process of yet another DTR ballot that wouldn't actually fix the technical content we've been objecting to. The disposition of comments document on this one *is* a problem, however, and the committee was very explicit about what the editor had to do here. The draft disposition of comments is being reviewed for a week by the people who attended the meeting. This is to doublecheck that the editor correctly recorded what he was supposed to during the meeting. Once we have agreement that he got the DOC correct, the editor will then roll the changes into the actual DTR text. Unlike earlier versions, *this* time the DTR text and the DOC will both be posted up for a one month period when anyone on the WG20 NB list can examine them to ensure that the editorial changes in the DTR text actually do correspond to the disposition of comments in detail -- not just the people who were in the room will be doing the checking. I'll notify people when it is time to do the checking. This won't be open season to create new comments -- just time to verify that the disposition of comments was correctly applied before publication. After that, DTR 14652 finally goes out for publication as a TR. Hopefully, it will just get drop-kicked into a black hole after that. ;-) ************************************************************ 4. CD 15897 -- Registration of cultural elements The disposition of comments for this standard was as controversial as for DTR 14652. Once again, the U.S. provided the bulk of the comments. The other comments were limited and fairly easily handled. The editor rejected many of the detailed U.S. ballot comments rather offhandendly, and once again, rather than fight doggedly on each one, my strategy was to simply force the editor to document each decision clearly, so that they can be checked against the revised text. In this case, the editor also had to go on record as taking into account the *extensive* suggested restructuring of 15897 to align it with 2375 -- the work done by Ed Hart and Joan Aliprand, with Michael Everson's assistance. (Michael is the editor of 2375.) WG20 is on record as wanting the registration process to follow the 2375 model. So we should watch the draft to see that it does in fact take into account the detailed restructuring Ed and Joan had drafted. The draft disposition of comments, as for 14652, will be reviewed by the attendees of the meeting for a week. After that, the editor has to produce a revised draft. And once again, there will be a one-month period where anyone in WG20 can check whether the Disposition of Comments was correctly applied to the document itself. ************************************************************ 5. Registrations under the current IS 15897 More controversy -- this over the batch of European locale registrations from TC304. WG20 went explicitly on record as not simply accepting the disposition of comments by TC304, but instead requesting that TC304 more thoroughly address the procedural concerns of the Netherlands and other NB's who objected to the entire process. Also, there were technical errors noted by Ireland that we also wrote into a WG20 resolution. This was an effort, in the absence of any current independent review body for these registrations, to put in place a precedent that national bodies, through WG20, could be expected to have strong interests in seeing comments actually resolved and problems actually fixed, rather than docilely accepting a blanket disposition of nothing. ************************************************************ 6. NP for a TR on I18N functionality This is the "son of 15435, I18N API's". We got a very clear, unambiguous ruling from John Hill, chair of SC22, that 15435 was dead, despite the editor's attempts to breathe new life into it. Later in the meeting, the issue of responding to SC22's invitation to instead start a new project for a TR on current practice in I18N functionality came up. I resisted this one by insisting that any such new project must have a decent founding document -- that it would be ridiculous to simply hand a blank check to start up some new project without any working document delineating the work to be done in some detail, and also that simply putting forward the last draft for 15435 made no sense, since that project was dead, and we weren't going to simply turn a bad, failed project into a TR and keep on going. So instead of starting ballotting for an NP for a new TR, the national bodies are now invited to submit anything they want addressed in such a new TR. And the person who will coordinate those submissions will be Pat Hall of the U.K. He agreed to pull together any input, and then draft a white paper based on that input and anything else he wants to bring to it. My guess is that Pat will, in fact, get rather little systematic input on the topic. I would suggest that the UTC not contribute -- we aren't under any obligation to do so. In any case, the issue is postponed to the next WG20 meeting, when we will see if enough content is assembled to warrant consideration of an NP vote. If the material is really weak, we might be able to just stop it right there, before the NP ballot even gets issued. ************************************************************ 7. New editor for TR 10176 -- Guide for Programming Languages Marc Küster of the German NB was nominated to be the editor of TR 10176, now that Arnold Winkler is stepping down. I expect with Marc as editor we may still get a sympathetic hearing for the notion of separating off the annex for identifiers (the only part that has needed updating), or even making it just refer to the Unicode recommendations. ************************************************************ 8. The future of WG20 and CLAUI Another hopeless deadlock. This time, rather than just issue a vanilla, "We have no consensus to do anything" resolution, WG20 drafted a rather detailed document as input to CLAUI, including 4 different options that could be taken, organizationally, to the problem of working on 14651 and other WG20 projects, with or without the cooperation of SC2/WG2. One proposal which received consensus, in outline at least, was to try holding WG20 meetings in conjunction with SC2/WG2, so as to minimize the travel burden of holding separate meetings. WG20 also went on record in that document as strongly concerned that CLAUI, as currently constituted, was not functioning well. Even Keld and Alain acceded to that. The notion of holding any CLAUI meetings colocated with JTC1 plenaries was also supported. ************************************************************ 9. Future meetings of WG20 WG20 decided to continue to honor the earlier decision to hold the next meeting in South Korea. Kim had already asked the Korean NB for their agreement on hosting. The next meeting is scheduled for February, 2003. WG20 also decided, as part of the discussion about the future WG20, to hold the following meeting colocated with WG2. The next available candidate would be the October, 2003 WG2 meeting now scheduled for the Bay Area -- probably in Mountain View.