UNICODE TECHNICAL COMMITTEE Meeting #65 Friday, June 2. 1995 Novell, Inc., Sunnyvale, CA Attendees: Representatives of Full Members: David Craig, AT&T GIS John McConnell, Apple Computer, Inc. Lee Collins, Apple Computer, Inc. Mike Ksar, Hewlett-Packard Company Don Carroll, Hewlett-Packard Company V.S. Umamaheswaran, IBM Corporation Lisa Moore, IBM Corporation Murray Sargent III, Microsoft Corporation Rick McGowan, NeXT Computer, Inc. Lloyd Honomichl, Novell, Inc. Joan Aliprand, Research Libraries Group, Inc. Ken Whistler, Sybase, Inc. Mark Davis, Taligent, Inc. (afternoon) John Jenkins, Taligent, Inc. (morning) Arnold Winkler, Unisys Corporation Guests: Mr. Tesring Choergual, Director, Tibetan Language Committee Professor Norzhang Wujen, Institute of Social Sciences of Tibet Professor Nyima Trashi, Tibetan University Mr. Mao Yong Gang, Chinese Electronics Standardization Institute Other Attendees: Asmus Freytag, Asmus Inc. Tex Texin, Progress Software Joe Becker, Xerox Corporation Steve Greenfield, Unicode, Inc. Document Registration: UTC65-95-020 Agenda UTC65-95-021A UTC #64 Minutes UTC65-95-021B UTC #64 Closed Caucus UTC65-95-022 Tibetan Proposal (Peter Lofting) UTC65-95-023 Chinese Proposal on Tibetan UTC65-95-024 Peter Lofting's Comments on Tibetan UTC65-95-025 Lloyd Anderson's comments on Tibetan UTC65-95-025A Tibetan Sorting, by Lloyd Anderson UTC65-95-025B Lloyd Anderson's official response to Tibetan Proposal UTC65-95-026 Michael Everson's Tibetan Comments UTC65-95-027 Michael Everson's Cherokee Proposal UTC65-95-028 Michael Everson's Runic Proposal UTC65-95-029 Tibetan Character Set for Information Interchange (code chart) *UTC65-95-030 Tibetan Sort Order Proposal *UTC65-95-031 Tibetan smoothly extended from native to Sanskritic usage UTC65-95-032 Microsoft proposal for 4 Math characters N-1199 Korean Proposal * - Lloyd Anderson E-mail. Chair: Joan Aliprand Vice-Chair: Joe Becker Secretary: John Jenkins Scribe: Steve Greenfield Meeting called to order at 9:30 a.m. Roll Call: Apple, AT&T, Hewlett-Packard, IBM, Microsoft, NeXT, Novell, RLG, Taligent, Unisys At roll-call, McGowan presented a proxy from Lloyd Anderson, authorizing him to vote for Ecological Linguistics. Asmus Freytag had previously presented a proxy from Hangul & Computer, Inc. On behalf of the UTC, the Chair welcomed our guests: three Tibetan experts, and Mr. Mao Yong Gang who had arranged for their attendance on behalf of the Chinese national standards body. Approval of the Minutes ----------------------- The Chair asked for any amendments to the Open Minutes of UTC #64: none was requested. IBM (Moore) moved, Unisys seconded: Motion: That the Open Minutes of UTC #64 be approved. 10 yes, 0 no, 1 abstention, 5 absent The Chair asked for any amendments to the Minutes of the Closed Caucus held at UTC #64. Since only one company was mentioned and only for one vote, Moore asked that "IBM" be stricken. The Chair called for approval of Minutes of Closed Caucus as amended. Microsoft moved, Taligent seconded: Motion: That the Closed Minutes of UTC #64 be approved as amended. Vote: 10 yes, 0 no, 1 abstention, 5 absent Due to the fact that some members of the UTC would not be attending the afternoon session, a request was made to reorganize the Agenda. This was done. Co-Location of December UTC with X3L2 ------------------------------------- No decision could be made because of unknown details re the X3V1 meeting. The Chair polled members on their preference for December 1 or 8 for the UTC meeting. Either date would be acceptable. Action Item (Greenfield): Note co-location of the December UTC meeting with X3L2 as a topic for the agenda of UTC #66. Seventh Unicode International Conference (September 14-15, San Jose, CA) ------------------------------------------------------------------------ The Conference will be held at the Hyatt Hotel in San Jose. Moore (Chair of the Editorial Board for the Conference) reported that there is now a conference agenda, with three tracks (one more than UIW #6). There will be several talks on new script proposals. Freytag said it promises to be a stellar program compared to last year (although UIW #6 was considered to be of good quality). The Conference has broader coverage than the previous Workshops: it deals with Unicode technology and also globalization products in general. For the first time, there will be a formal exhibition, of products using Unicode technology, and (in conjunction with "Multilingual Computing"), by companies that market multilingual products. Freytag continued: The other essential for a successful conference is getting people to come. Member companies need to help Barbara (the conference organizer) with publicizing it. Can your company donate mailing lists, or mail out brochures? Please suggest names of e-mail lists where the conference announcement can be posted. Contact Barbara at "conference@unicode.org". East Asian Subcommittee ----------------------- Jenkins reported that the Consortium is working with the IRG Editorial Committee on the culling of duplicate ideographs from the current proposal. Mr. Zhang Zhoucai is now Rapporteur for the IRG. The UTC was pleased to hear this. An unsanctioned (and incorrect) CNS mapping table has been posted on Usenet. Jenkins has responded with a Usenet posting identifying the errors and directing Usenet readers to the sanctioned mapping table at the FTP site. He also asked for member companies to be dilegent on this issue. After discussion on how the problem of unsanctioned mapping tables might be mitigated, it was decided to post a Unicode reminder FAQ periodically to suitable newsgroups, e.g., to comp.std.internat. Action item (Freytag and Jenkins): Establish wording for the periodic posting. Discussion on Korean -------------------- Reference: Extract from WG2 document N 1199 (the explanatory text) Because of an error in photocopying, a replacement will be sent by mail to all member representatives and other meeting attendees who are not on the WG2 mailing list. If you require a complete copy of N 1199, please contact Steve Greenfield at unicode-inc@unicode.org. Action Item (Greenfield): Send replacement copies via USPS. The Korean national standard body submitted their Proposal to WG2 at the last meeting. The SC2 Ballot on Korean closes at the end of August. The relevant WG2 documents are N 1198, N 1199, and N 1209. Ksar requested members to encourage those responsible for voting to vote before the deadline. Freytag asked whether any changes had happened to the Korean proposal since the UTC last met. He also asked Ksar what he thought about the likelihood of passage. Ksar replied that the proposal was accepted by WG2 with respect to the location of the new codes (i.e., 11,000 hangul in a block in one location). WG2 proposes that deprecation of the existing codes occur in April 1998. Freytag asked whether there are a significant number of existing implementations that would be affected by the change in hangul coding (i.e., that have implemented the coding in Version 1.1). Umamaheswaran said that there is a standards conformance issue with regard to being in synch with ISO 10646. Ksar said that there seems to be a fear that developers will implement with the old codes (i.e., those to be deprecated in 1998). He pointed out that the deprecation has been announced. Two implementations that would be affected by the change were identified at the WG2 meeting. One would have a problem; the other (MASS, from Singapore) would not. Moore said that IBM would rather like to have the change happen as quickly as possible. McGowan pointed out that Unicode does not have to be on the same timeline as WG2. The UTC may say that the old codes are immediately deprecated as of the WG2 meeting in Geneva. Freytag pointed out that the book will contain code charts only for the new characters, with details on what it means for implementers. In response to a comment from Ksar, regarding the ballot's schedule, it was confirmed that the characters would be annotated "still pending approval by ISO." Plans call for the book to track ISO 10646, corrigenda to it, and PDAMs, and to identify sections suitably with respect to ISO actions (e.g., if approval of a PDAM is still pending). Instructions to Unicode Liaison Representativeto WG2: RLG moved, Microsoft seconded: Motion: The UTC prefers to make a shorter period for the deprecation rather than a longer one. Vote:Unanimous Runic Proposal -------------- McGowan recommended that the UTC support Everson's comments, and request the Scandinavian nations working on runes to take these comments into account. He asked whether it was the sense of the UTC that the Unicode Consortium declare a primary interest re runes. Jenkins said that he agreed that we need to contact the people making the proposal on runes. He would like the Unicode liaison to be assured that the Scandinavians have seen Michael Everson's comments and will take them into account. Freytag pointed out that if we decline to participate, we are morally obligated to accept whatever WG2 decides. NeXT moved, second not noted: Motion: That when the question comes up in WG2, the UTC commit to taking an active part in the development of the runic character repertoire. Umamaheswaran pointed out that there are actually three possible courses of action (not just two): no participation; making contributions via written liaison statements; active party in development. McGowan said that we have incurred something of an obligation to take an active interest in items in which we have already declared an interest (through publication of a proposal in a UTR). Freytag said that the UTC has a role to play as a center of coding expertise. Any BMP candidate is our primary mission. For anything outside the BMP, we can use the written comment method. NeXT moved, Taligent seconded: Motion: That the Unicode Consortium's default position should be to declare an interest. Vote: 11 yes, 0 no, 1 abstention (Ecological Linguistics by proxy), 4 absent Cherokee Proposal ----------------- Moore reported on Michael Everson's proposal on Cherokee. The proposal is similar to the original Unicode proposal. There has been no official response from the Cherokee Nation as yet. Another Cherokee reviewer suggested some modifications. Michael Everson told Aliprand that Gloria Sly was working to coordinate the opinions of three bands of Cherokees; Moore has also been in touch with Gloria Sly. Aliprand said that the UTC should take no action on Cherokee until there was a response from the Cherokee Nation. Action Item (Aliprand): Write to the Cherokee Nation to encourage them to reach closure. We should stress that the code chart is correct, rather than the text itself. Whistler suggested that the Society for the Study of Indigenous Languages of the Americas also be contacted. Action Item (Aliprand): Get information about the Society from Whistler. Tibetan Proposal ---------------- Collins demonstrated software from Otani Daigaku (Otani University, a Buddhist institute in Japan) to the Tibetan experts. The software is based on an encoding similar to that published in Version 1.0. (Tibetan was withdrawn from Version 1.1 to allow further study). The Chair deprecated the fact that proposals and comments on them had been received so close to the UTC date. This was exacerbated by unanticipated delivery problems: Steve Greenfield sent Peter Lofting's May 22 proposal by first-class mail on May 23, but not all member representatives had received it by the end of May (including at least one in the Bay Area). The Chair announced that, for future script proposals, firm deadlines will be established for submissions well in advance of the UTC. Submissions received after the deadline would be disregarded. The Chair identified the UTC documents relevant to the discussion on Tibetan, namely: 022: "Notes on Changes to Tibetan Script Encoding" Dated 22 May 1995, from Peter Lofting Sent by first class mail to member representatives on May 23 023: "Explanation of Mr. Lofting's Comments on Lhasa Tibet Proposal" Dated June 1, 1995, from Chinese delegation 024: "Tibetan (U+xx00-->U+xxFF)" Dated 31 May 1995, from Peter Lofting 024A: "Questions on Lhasa Tibetan Encoding" Dated 31 May 1995, from Peter Lofting Contains Peter Lofting's questions about Document 029 024B: "Updated Tibetan Code Chart" Dated 31 May 1995, from Peter Lofting 025: "Comments on Tibetan Proposal" Undated (June 1?) from Lloyd Anderson, Ecological Linguistics 025A: Memo beginning "Peter Lofting was very kind ..." Dated June 1, 1995,from Lloyd Anderson, Ecological Linguistics 025B: Memo beginning "The following paper will be distributed ..." Dated June 1, 1995,from Lloyd Anderson, Ecological Linguistics Apparently superseded by 030 and 031 026: "Toward Consensus on Tibetan Encoding" Dated 1995-05-30, from Michael Everson 029: "Tibetan Character Set for Information Interchange" Coding chart, undated, from Chinese delegation Also called "Lhasa character set", "Lhasa Tibet proposal", or "Lhasa Tibetan encoding" 030: "Tibetan Sort Order Proposal to Follow Chinese Great Dictionary" Dated June 2, 1995,from Lloyd Anderson, Ecological Linguistics 031: "Tibetan Smoothly Extended from Native to Sanskritic Usage" Dated June 2, 1995,from Lloyd Anderson, Ecological Linguistics [The prefix of these all documents is UTC65-95-] The key documents are 024B (Peter Lofting's latest code chart) and 029 (the Lhasa code chart). The Chair called on Rick McGowan to summarize the issues regarding Tibetan. McGowan said that he saw a small number of basic questions: * What order should we put the consonants in? Anderson favors Chinese order with provision for f after ph and v after b to get a normal ordering that is reasonable for Sanskrit. That is, the order in Document 029 with provision for f and v. * Encoding of independent vowels. Lofting encoding has pre-encoded initial vowels; Lhasa Tibetan encoding does not. * Half numbers. Not in Lofting encoding; are in Lhasa Tibet encoding. Mao pointed out a defect in Document 024B. Although it says it is "a combined Unicode & Lhasa Tibetan code chart," it is missing characters zz01, zz03 and zz04 (using Lofting's notation as given in Document 024B). Jenkins asked MGowan to give a short history of the work on Tibetan. The original Unicode encoding for Tibetan, published in Version 1.0, was maximally decomposed. There were certain questions about it from British experts, and it was not included in Version 1.1, but was withdrawn for further study. The revision was published in Unicode Technical Report #2. (UTRs 1-3, which contained script proposals, were distributed world-wide, with particular efforts made to send them to standards organizations, leading universities, and national libraries in countries which used the scripts.) The revised proposal was countered by a different proposal from the British. China's initial proposal for the encoding of Tibetan resembled a font encoding, because was a large set or precomposed entities. Discussion with the British followed. China's current proposal (Document 029) and the Unicode proposal are converging. It should be possible to agree on a workable model. The character repertoire is relatively stable, and is well-defined. There are a few fringe question, and internal question of alphabetical order. Umamaheswaran suggested that a small group consisting of the Chinese delegation and UTC attendees most interested in Tibetan work on a consolidated proposal for Tibetan. Ksar supported this suggestion. Umamaheswaran added that the focus document should be the Lhasa character set, Document 029; McGowan agreed. Whistler suggested that the intended deliverable be clarified. McGowan said conceptually it should be a repertoire and ordering to carry jointly to WG2. Practically it would be the code chart (029) with a list of the changes to be applied to it. The members of the delegation from China, McGowan, and Collins went to another room to work on the Tibetan encoding. Meeting Resumed at 3:40 Status of Version 2.0 Book -------------------------- Davis reported that a review group consisting of most of the original authors met on Thursday to go over the text which incorporated the changes from the editorial review meeting in March. While the changes had been made to Chapter 1, Chapter 2 had been substantially rewritten, and much of the original text had been dropped. Since work was not proceeding sufficiently well, members of the group decided to do something about it. The goal of the authors is to have a first working draft by June 17. This first draft will merely be an amalgamation of the source works and all the material from UTC decisions. Once this has been done, the authors will prepare a coherent second version, which will be the version sent to members for review (and to Addison-Wesley for copy-editing). Aliprand had sent call for volunteers to corporate member representatives, and a couple of people expressed interest. Sargent asked about the size of Version 2.0. Davis replied that it would be one volume, somewhat thinner than the combined 2 volumes, with an accompanying CD-ROM. There was discussion on whether the CD-ROM should include character images. Davis replied that ideas could not be put into effect without volunteers to work on them. Umamaheswaran asked whether the Consortium had approached AFII. Davis replied that the Consortium has talked with AFII many times, but without results so far. Ksar said there is an AFII CD-ROM which includes ideographs; Freytag pointed out that they are bit-maps. [TrueType fonts are needed for publication of Version 2.0.] Ksar said that AFII fonts are available to companies that are members of AFII; a restriction is that they cannot be put into products. Action Item (Ksar): Talk to AFII re acquisition and use of their font. Action Item (Freytag & Sargent): See what Microsoft can do re providing a TrueType font for hangul. Aliprand passed around the current NISO Press catalog, which includes a listing for the new version (listed as "Version 1.1, Second edition"). NISO is the National Information Standards Organization, which develops standards for libraries and the publishing industry. [Version 2.0 already has an International Standard Book Number: 0-201-48345-9. The ISBN is an exact identifier for ordering books.] In response to a question from Ksar, Davis replied that once we have a first draft (target date is June 17), the authors will be better able to estimate a likely publication date. Aliprand pointed out that several authors will be away for the last half of June and early July, so that if the June 17 date is not met, all dates will slip out considerably. WG2 and Helsinki Meeting ------------------------ The Chair had sent out a request for member representatives on the WG2 mailing list to review WG2 documents, and propose discussion by the UTC of any they considered important. None was proposed (other than N 1199 on Korean). Davis mentioned the proposal to WG2 for the addition of Byzantine musical symbols to ISO 10646. Aliprand asked whether they fell within the scope of ISO 10646. Umamaheswaran replied that there were rules for this. Davis reiterated the Unicode position on music, which was explicitly excluded because it was considered a two-dimensional representation with a complex syntax and rendering requirements. Umamaheswaran asked about basic publishing needs. Davis replied that there would probably be rendering issues. Davis said that there was a need for a Unicode liaison statement to WG2 on Byzantine musical notation. It was the sense of the meeting that the UTC would have no objection to putting Byzantine musical notation outside the BMP. Ksar raised the issue of appointment of a Unicode liaison representative on a longer basis than meeting-to-meeting. Davis replied that he would encourage as many people as possible to have a long-term relationship with WG2, but there is the problem of corporate support. Freytag pointed out the desirability of institutional memory at WG2 meetings. Umamaheswaran said that the liaison should have an additional function: to review WG2 mailings and bring documents of interest to the attention of the UTC. Aliprand, Davis, and Freytag were of the opinion that UTC member representatives who receive WG2 mailings are free to propose that the UTC discuss WG2 documents that they consider important. Umamaheswaran said that WG2's concern is that the reverse is not working. Freytag replied that members must make a judgement of priority level on what they consider important. Davis said that the "high ticket" items for the UTC are Korean and Tibetan. Freytag pointed out that countries are obliged to respond only to ballots; they are encouraged to respond to papers. No other organization or constituency responds to every single paper. Standing Action Item (Aliprand and Greenfield): Issue call for WG2 documents to be discussed before each UTC. Include discussion of WG2 documents as a recurring item on the agenda for UTC meetings. It was agreed that "unicore@unicode.org" should be added to the WG2 mailing list, so that all regular UTC attendees would receive the WG2 document registry as well as other WG2 documents. This is on atrial basis, and will be reviewed at UTC #66. Action Item (Ksar): Arrange for "unicore@unicode.org" to be added to the WG2 mailing list. The responsibility to share review of WG2 documents is shared by *all* UTC member representatives. It is not the ex officio responsibility of office holders (such as the Convenor of WG2, the UTC Chair, or the Unicode Liaison to WG2). Tibetan: Report of Ad Hoc Committee ----------------------------------- Collins and Mao reported on progress. * The participants are very close to agreement. * They will prepare a modified code chart. * Collins and McGowan recommend a letter ballot. Collins summarized decisions that had been made. The Lhasa encoding was used as the basis. Sanskrit precomposed vowels have been removed. Subjoined letters will be used. Ambiguities re characters already in the Unicode Standard have been resolved, e.g., swastikas will be encoded in dingbats, not in Tibetan. Freytag, who will be the Unicode liaison at the WG2 meeting in Helsinki, requested a document defending the changes that have been made. The UTC needs to say to WG2 that it is confident that this is the best solution. Freytag added that China may also wish to make such a statement. Ksar said that he wants the concerns of the British experts, Mr. Ross and Mr. Lofting, to be addressed in the document. He added that, at Helsinki, he does not want to see a proposal from China that is different from that submitted by the Unicode Consortium (the encoding being worked on today). Mao replied that from now on, we are working on a joint proposal. Freytag asked whether the formatting codes would be put in the Tibetan block or the General Punctuation block. Collins replied that it is likely that there will not be any formatting codes at all. Freytag remarked that if we can do Tibetan without formatting codes, we would prefer it. Collins pointed out that Unicode encodes Indic scripts without the use of formatting codes. It was agreed that the code chart would be faxed as a letter ballot as soon as possible, with the explanation to follow by e-mail. At the same time, a copy of the code chart will be faxed to Mao Yong Gang in Beijing, and to Lhasa as well if possible. Freytag described the procedure to be followed if the Tibetan proposal is approved by the UTC in the letter ballot. There will be one WG2 document which is the proposal itself, with a cover letter from the Chinese national body, and the Unicode Consortium. [Separate cover letters had initially been proposed.] The procedure was approved by consensus of the UTC. Mao pointed out that the Tibetan proposal will have to be approved by various departments in China. Freytag said that the Unicode Consortium does not want to submit the Tibetan proposal without support from China. It must be a joint proposal. If either the UTC or China fails to approve the Tibetan proposal, Freytag will not submit the proposal to WG2. The only thing needed for the WG2 meeting in Helsinki is the final code chart with relative assignments and the final names list. Freytag will provide the cover letter if both China and the UTC have approved the proposal. Mao was asked whether he thought China would be able to approve the Tibetan proposal by before June 26 (date of the WG2 meeting). He replied that China hopes that the joint proposal will be accepted on June 26. Winkler asked whether Peter Lofting had been invited to attend the UTC. Freytag replied that he had been invited, but the Consortium was not willing to pay his way. Mao reported that China had written to interested parties (as instructed by WG2), but had received no replies. Freytag said that China will be able to say at the WG2 meeting that its National Standards body practised due diligence. Action Item (Aliprand): Inform Peter Lofting and Michael Everson of our action. The justification document is to be sent in reply to any comments from them. Since Aliprand will be at a conference for part of the letter ballot period, Mark Davis volunteered to be responsible for the letter ballot. Action Item (Davis): Arrange for the letter ballot. It was noted that the WG2 proposal must include the proposal summary form (N 1116). Since one of the questions on this form relates to the availability of fonts, it was agreed that China would be responsible for this form. Action Item (Mao Yong Gang): Supply completed WG2 form N 1116 for Tibetan proposal. Mao said that China was encouraged by the comments from Bhutan which had been reported at the WG2 meeting in Geneva. Ksar replied that (with respect to the instruction to cooperate with other interested parties) WG2 is satisfied that China had tried to comply. China is not responsible if other parties fail to respond to China. Proposal from TCVN/TCI on Vietnamese Currency Symbol "Dong" ------------------------------------------------------- Submitted by James Do, on behalf of TCVN/TCI, at UTC #64. Freytag said that this symbol was approved by WG2 in Geneva. In addition, it appears in some Vietnamese 8-bit character sets. Vietnam is also of interest from a world marketing perspective. Taligent moved, Unisys seconded: Motion: That the UTC accept inclusion of the dong currency symbol as a precomposed character for The Unicode Standard, Version 2.0, with the goal to be in synch with ISO 10646. Vote: Unanimous (10 members present) In response to a question from Umamaheswaran, Davis clarified that there are two categories of "proposed". Both will be marked in the code tables the same way, but their different status will be explained in the text. Microsoft Proposal on Mathematical Characters --------------------------------------------- Reference: Document 032 Sargent presented a proposal to add four more mathematical operators. He said that Unicode has made mathematical formulas easier to do. Freytag said that there is a certain logic to the proposal. They deserve to be processed as part of the examination of whether our symbol set is sufficient for mathematics. He likes the superscript and subscript operators; can be used for decompositions. Davis asked whether the superscript operator (for example) needs to have a visual form or not. Sargent replied that it is nice to have a plain text symbol that can be used without building it [a formula] up. It is an issue of syntax. Further discussion clarified that it is the superscript or subscript operator that produces an effect, but it has a fallback visual appearance if the effect cannot be produced (e.g., the character affected by the operator cannot be displayed superscripted). Freytag suggested that we might also want to use this technique for the current formatting characters, providing a visual fallback realization for them. Texin said that it is an issue of the font and the operator. We don't have a definition for the interaction between font and character in Unicode. Needs the mathematical sense for what the operator means. Whistler said that the proposal is suggesting a new class of characters "Operators" with a new semantic. Based on a comment by Freytag, RLG moved, Novell seconded: Motion: That Murray Sargent develop a proposal (with full algorithm) for mathematical operators, and that this is to include rules for the effect on all Unicode characters. Amended upon suggestion of Freytag: This proposal to be submitted for UTC #66 (scheduled for Sept. 29). Freytag said that the proposal should include both description and rationale, and that these should be separate sections. All 65,000 characters need to be classified as to their function under the algorithm, similar to what has been done for the BiDi algorithm. Vote: Approved by consensus. Action item (Sargent): Prepare proposal on mathematical operators for UTC #66. The Chair thanked Novell for its hospitality, and Lloyd Honomichl and his Administrative Assistant for making the arrangements. The next UTC meeting will be in Boston, Hosted by Lotus. Meeting adjourned at 5:57 p.m.