ISO/IEC JTC1 SC22/WG20 N1004

March 11, 2003

 

DATE:                    February 11-13, 2003

 

TIME:                     9:00 a.m. to 5:00 p.m

 

LOCATION:         Room 311, Main Building ("Bonbu" in Korean), Pusan National University, BUSAN, Republic of KOREA

                                Preliminary hotel and travel information is attached at the end of the agenda.

 

CONTACT:          Prof. Kyongsok Kim < [email protected] >

 

Meeting web site:  http://asadal.cs.pusan.ac.kr/hangeul/i18n/wg20m23/

 

WG20 Web site:  http://wwwold.dkuug.dk/jtc1/sc22/wg20/

 

 

1.Introduction and announcements by Convenor

 

 

The meeting is opened at 9:35 by Gim who chairs due to the physical absence of Ken. Ken is available on the conference telephone. He will attempt to participate as much as possible on the phone, but that will not be possible throughout the meeting. [In fact, Ken was able to attend the entire meeting by phone.]

 

Ken presents his excuses for not being able to participate in person..

 

Congratulations of those present to Ken for his appointment as Convenor.

 

WG20 notes the ballot comments in N999 and N1007. Gim states that the Korean national standards group voted No on the election of Ken but that this vote was not recorded by the SC22 secretariat. The Korean vote will be registered under WG20/N1015 and has been registered in the internal SC22 registers.

 

Document numbers will be allocated during the meeting by Gim and Keld together for practical convenience.

 

2.Introduction of national delegations, liaisons, and cooperation partners

 

 

Participating in person: 

Norway: Keld Simonsen (RAP)

 

Germany: Marc Kuester (Saphor)

 

Republic of Korea: Professor KIM, Kyongsok/GIM, Gyeongseog  (Busan National University)

 

Republic of Korea: KIM, Dongwon (Korean Agency for Technology and Standards)

 

Participating electronically:

Convenor: Ken Whistler (Sybase)

 

Canada: Alain LaBonté (Secretariat du Conseil du tresor du Québec) (part of the meeting)

 

Sweden: Kent Karlsson (part of the meeting)

 

USA: Mark Davis (part of the meeting)

 

USA: Joan Aliprand (part of the meeting)

 

Keld raises the issue if Ken can act both as a convenor (even though not physically present) and as a US representative.

 

Ken stresses that he is appointed as national representative for this meeting. As to the ballot comments, Ken is the one chosen to defend the US position.

 

Keld is willing to accept this situation for this meeting, though reluctantly. In general, in his views the consequence of being a convenor is not being able to represent a national body.

 

AI: Keld will investigate if that is the correct mechanism in the long term. 

 

AI: Ken will also investigate if that is the correct mechanism in the long term.

 

3.Appointment of chairperson, secretary, and drafting committee

 

Arnold Winkler, while still acting as convenor, arranged with Gim to chair this meeting. That arrangement was confirmed by the SC22 Secretariat. Ken as current convenor reaffirms this choice.

 

Marc Küster was appointed secretary of the meeting.

 

The drafting committee consists of Gim, Keld and Marc.

 

We will sign a list of participants (N1002). The minutes will note also the telephone participants and the duration of their presence.

 

4.Approval of prior meeting's minutes

 

 

No objection was voiced to approval of the Minutes. People can comment until the end of the meeting.

 

[In fact, no such comments were forthcoming, so the minutes are approved]

 

5.Future Meeting Schedule and Plans

 

Planned next meeting is planned for October 2003 in California, co-located with the SC2/WG2 meeting (as per the minutes of the last meeting, N932). Should that date be retained?

 

Keld: This issue was discussed by the last CLAUI meeting with no SC2 participation. The meeting schedule should be coordinated with the requirements of the work. He suggests to take up the issue again towards the end of the meeting.

 

Gim: What about follow-up host?

 

Ken would be willing to host the next meeting in autumn in California.

 

Gim stresses that this is a separate issue from co-location. This finds support.

 

All agree to postpone this discussion towards the end of this meeting. It would be a good idea to speak about this when Alain is

also on the phone.

 

Alain is quite happy to meet next time in California (Mountain View).

 

Next meeting of WG20 October, 15th-17th, in Mountain View, California

 

Ken: Meeting modalities should be discussed. We should meet more electronically (NetMeeting and similar) and by telephone. We should have a core group meeting physically at a given place and have other participants dial in from the outside as a matter of course. The telephone costs are very low nowadays from most countries to the outside. There is universal agreement that this is the way forward.

 

Action item on the group: There is a need to meet electronically in the near future to research software options for electronic

meetings.

 

6.Recognition of new documents and assignment to agenda items

 

 

 

 

 

Follow-up documents on N952ff, namely N1012 and N1013 (Hangeul-related portion of CTT in ISO/IEC 14651 (1) and

Hangeul-related issues in the UCA). This will be handled under agenda item 12 (instead of N952, N953, N954).

 

N1014: A few ideas on sorting. Will be handled under agenda item 15 (any other business).

 

N958: Problem statement of expressing DIN 5007 in the framework of ISO/IEC 14651 should be taken up again under agenda item 12.

 

N1010 and N1011: Ballot comments on 15897. They are discussed under agenda item 14.

 

N1016: Input of ITTF regarding the publication of 14652 (under agenda item 13).

 

N960: "Final disposition of comments to 14652" will be added under agenda item 13.

 

7.Approval of Agenda 

 

 

The heading of agenda item 12 is changed to "International string ordering" to cover both Korean and generic string ordering issues.

 

Keld: Under agenda item 15 there will be a discussion on the "TR on i18n functionality" that Pat Hall had in Tromsö committed to work on.

 

Keld: Marketing: How can the marketing of WG20 and its standards be improved? This will be discussed under agenda item 15. Keld will prepare a short document prior to that discussion.

 

Marc: "I18n and cultural diversity": To be discussed. Marc will prepare a short document / text for discussion until February 12th.

 

Keld: Discuss "Browsing and Matching". That will be discussed under 8.9 together with document N994 in which this topic figures prominently.

 

The agenda is approved with the given changes.

 

8.Liaison Reports

 

 

8.1Additions/deletions/changes to liaisons

 

Gim appointed as liaison from WG20 to WG2 (there will be resolution to this effect).

A SQL-liaison would be highly desirable, since there is important i18n work ongoing there right now, but not of the present participants of WG20 attends the meetings of SC32/WG3.

 

TC46/SC2 does not exist any more. It is now TC46/WG3 "Conversion of written languages" (convenor ELOT). SD4 will be updated accordingly. The liaisons will remain unchanged.

 

Action Item on Keld: Will look into the proposed new liaison with Li18nux again.

 

For the liaison to the CDSG, the word "CLAUI" will be cut from the description of the liaison in SD4

 

Keld raises the issue of a liaison with IETF where SC22 has requested a A-liaison. At the SC22 plenary in Hawaii, Mike Ksar has spoken against an A-liaison. SC22 has agreed a C-liaison. Nothing has been followed up since then.

 

Ken asks if that is a WG20 issue or rather a SC22 issue.

 

There is a long discussion on this topic. In the end, the agreement is for WG20 to make a simple statement to SC22 to pursue the liaison (a recommendation) to find out the status of the resolution. This is later revisited and the Ken will contact Matt Deane directly on the status of the liaison even during this meeting (AI).

AI Keld will contact IETF RE: this matter, John Klensin.

 

 

8.2SC22/WG4, COBOL

 

Nothing new (except for a new convenor).

 

8.3SC22/WG5, Fortran

 

No active liaison with Fortran exists any more. This agenda item will be dropped from the agenda in the future.

 

8.4SC22/WG14, C

 

 

This project is very important for practical i18n in C (and C++). Marc stresses that this project is strongly supported by DIN.

 

8.5SC22/WG15, POSIX

 

The Austin group is producing technical corrigenda. There are regular meetings by phone right now on this.

POSIX is the working group mostly related to Linux and Li18nux.

 

8.6SC22/WG21, C++

 

 

N975 points out many differences between the extended characters of TR10176 and C++. The main reason for this is that the C++ standard used the DTR 10176 current at that time.

 

In their revision, the C++ people should do their best to align as much as possible with the current version of the TR 10176. A written contribution of WG20 to WG21 should be undertaken. The same must also handle C because of the interoperability issues between C and C++. There must not be legal C code that does not compile with C++ compilers because of i18n issues.

 

Action item on Keld as liaison officer: Provide a paper to WG21 that they should align with TR 10176 with specific proposal to make specific changes during the revision of C++.

 

Send this document also for SC22 consumption.

 

We should discuss the issue of inter-language interoperability (C,  C#, Java) for a future revision of TR 10176.

 

8.7SC2/WG2

 

 

The two parts of 10646 will be consolidated into one document (no more two parts, there will be one ISO/IEC 10646:2003). The next meeting will be in California.

 

8.8SC35

 

 

WG20 notes these documents.

 

8.9CEN, CDSG

 

 

 

CEN/TC304: the status is unchanged.

 

Keld:  SC22 has more or less canceled the cooperation agreement.  However, a few projects have not been completed.

 

Marc: No need to hurry a decision from SC22 either way. Next time we are likely to know more.

 

Marc is the chairman of the CDSG. The membership is European and  international. WG20 is invited to send representatives 

to the CDSG. It is unclear if WG20 can send one or more representatives. Keld and Marc are suggested for membership. This suggestion is later extended to John Clews and Keld (in that order) as members since John has long been active both in the CDSG and WG20..

 

There will be a link from the WG20 site to the CDSG.

 

Action item: Marc will send link to Keld.

Comments on N994 are invited from all parties / Ken will bring it also to the attention of the Unicode Consortium. The work

Programme in N994 is a living document that will continuously change in the view of new development.

 

Browsing and Matching has been brought up by Keld. In N994, §6.3.6, representative. Should we handle this in the WG20?

 

Marc: Will there be the necessary industry involvement in WG20?

 

Ken: The CDSG primarily suggest scoping actions and that are necessary preconditions for future work.

 

(6.3.9 From Ken's perspective scope too broad. The scope should be  restricted to a subset of Indo-European languages,

e.g. Romance and / or Germanic and / or  Slavic languages. Indo-European in general is too generic. That would

give a proof-of-concept that could be extended).

 

Agreement: Let things be done in CEN and follow the developments with interest. No specific action needs to be taken at this

point in time.

 

8.10TC37

 

Some work has been ongoing for approving new language identifiers. Some general discussion on the structure of the standards.

 

8.11Unicode

 

 

N1005: Unicode 4.0 is reaching the final stages of publication. It encompasses all the changes since Unicode 3.0. The datatables will be released in April. The repertoire of Unicode 4.0 is exactly synchronized with 10646:2003, the unified version.

 

The book will be published in late August or early September.

 

The Unicode Consortium has started using a public issues review process. There is now a link from the homepage "Issues for

Review." Anybody who is interested is invited to send in input to the document.

 

8.12W3C

 

 

Report from Martin Dürst in N996.

 

Pointer from Marc to the mailing list of the W3C i18n list to which he subscribes. The W3C is pursuing a very open process.

 

Action item: Keld as the liaison to W3C will try once more to get on the W3C mailing list.

 

 

9.Review of prior meetings action items  [SD-5]

 

 

All action items done or closed except A0022-02 Simonsen "Extract e-mail thread that shows how WG20 could support the C++ effort for new library functions on character sets and locales." This one is now closed too.

 

A0022-06 is done.

 

10.CLAUI meeting in Antipolis, France, October 2002

 

 

Discussion postponed until Alain is online.

 

988 JTC1 resolution: resolution #21: dissolution of all tech. directions.

989 JTC1 resolution #22: cancel current form of cooperation and find an alternative way.

 

Keld: CLAUI asked wg20 for free availability on web of 3 TR's. We should ask SC22.

We need to decide that now.

 

928: Report from the CLAUI meeting in Paris Dec. 2002,  where a generic agenda was agreed

959-CLAUI

929, #4

 

AI: Ken: require of SC22 secretariat to make our technical reports 10176, 14652, and 11017 publicly available on web.

 

11.ISO/IEC TR 10176 - DTR voting results

 

 

No comments were received during the ballot.

 

How about those 1482 characters listed in Annex A of the Annex as listed as possible identifiers by the Unicode Consortium,

esp. combining diacritics?

 

Ken: We should maximize the interoperability issues between languages. We should examine the situations that presently exist.

 

Ken: We should make recommendations for specific issues, e.g. problems with normalization.

 

Gim: 11XX characters (240 conjoining jamos) are needed for individual letters and / or historical characters. They are rarely

needed, but they are needed in general.

 

Ken: They are handled in 10646, Annex B.2, as combining diacritics. That could be the case for column names in database that

identify old Hangeul concepts. Otherwise there will be unnecessary constraints.

 

Gim: There is a general need for this.

 

Keld: This would endanger portability because it implies normalization. There is no normalization specification.

 

Ken: This is a problem of what is "portability." For Java, it is the Java VM that guarantees portability.

 

Keld: Source code portability must be guaranteed.

 

Ken: It is the language standards and conforming applications that must guarantee source code portability. The interoperability issue is of central importance.

 

Keld: The portability issue is an important issue for each language standard for source code portability.

 

Marc: The main issue is whether normalization would add significantly to the implementation of compilers.

 

Ken: There are free libraries available that do normalization.

 

Gim: The definition of normalization is complicated. There are many options available.

 

Keld: Normalization has always been controversial.

 

Marc: Normalization is used in Java and Java is used a lot in embedded devices.

 

Ken: There is an entire database that does normalization and that fits in embedded systems.

 

Marc: As editor I'll do what the group wishes. However, I think the division of Annex A is unnatural.

 

Keld: Normalization issue should be avoided. We can consolidate the state of the art information from the programming languages. Many programming languages are happy that they can avoid normalization.

 

Marc: Is the gcc actually used as a compiler on an embedded system?

 

Ken: We should take concrete actions to come forward with concrete changes. This should be done on the text of the standard itself.

 

Action item on Marc as editor and Keld to come forward with a paper that will show how the table can be changed.

 

Keld: We must consider the issue of casing for languages such as COBOL and Fortran. This should be specified in 10176 rather than 14652.

 

Ken: Write up a document.

 

Keld: We should also address the issues of i18n. APIs.

 

12.International string ordering - discussion of Korean ordering rules

 

 

Issues of Korean ordering:

 

N1012 and N1013: The articles by Gim.

 

N1018: Mark Davis' "response" to N953, the precursor to N1012.

 

During this discussion, Ken, Kent Karlsson and Mark Davis are online. The complete discussion takes about one hour.

 

Gim gives a brief presentation of his two contributions, in particular of N1012. The document shows how the present algorithm

can result in unacceptable comparisons.

 

The real discussion starts with 5.1 (p. 9): The UCS has no state except for very rare exceptions (e.g. Thai, Lao). CTT and UCA should produce exactly the same results with regards to Canonical Decomposition. Mark Davis will send written answers to 5.1 of N1012.

 

The final outcome of the discussion is as follows:

 

- The first step will be to prepare a requirements paper on Hangeul ordering

 

- Technical discussion on the implementation will ensue

 

In practical terms, Mark Davis will prepare a draft requirements paper (action item). Gim will collect questions on N1018.

 

All email discussion will take place on the normal WG20 list.

 

German concerns: N958: Germany has been invited to send a contribution on this issue to Alain (action item).

 

General issues: N1006 (ballot results for FPDAM): The vote was accepted without problems.

 

Schedule for 14651: The repertoire of the next amendment will be aligned with 10646:2003. The schedule will be adapted to that.

 

Keld has an action item to contribute a paper on a Norwegian alignment of the DTT in the direction of EOR.

 

13.Resolving publication issues for TR 14652

 

 

ITTF does not like it to mention the objecting member bodies in Annex D. They would prefer to remove the attribution of specific issues to given countries.

 

Gim supports the view of ITTF.

 

Ken points out that the removal of country attributions to the issues would suggest that these issues were universally agreed to. Ken suggests instead to follow the edits suggested by ITTF but to insert a warning note that these issues were raised by one or more country during the ballot period. Gim seconds this view. Keld will prepare such a text by email. There is universal agreement that this is the way forward and that it both fulfills the requirements of ITTF as expressed in N1016 and stresses that not all issues are shared by all countries alike..

 

Relating to the revision of Keld said that we should begin revising 14652 in the light of what we said we would do in N960, Ken said he was not very interested, and Marc said that he thought it should wait for the publication of TR 14652 and give users sufficient time for feedback. All agreed, however, that the editor always could come forward to the WG with a new draft. The WG could then decide how to handle this draft.

14652 will thus be published with the following changes:

–country names will be deleted;

–a phrase will be added "... these issues were raised by one or more countries"

 

14.Registration of cultural elements, ISO/IEC 15897 – ballot resolution

 

 

 

The comments are run through for Germany, Netherlands, and the UK. For details, cf. the disposition of comments.

 

For the US comments Joan Aliprand joins the teleconference.

 

Joan highlights several key points of the US comments:

 

·         Should be aligned more with ISO/IEC 2375

·         Should be brought up-to-date with the new POSIX standard currently released

·         Should be better usable.

 

The results of the discussion are captured in the disposition of comments.

 

There is a lot of discussion on the question if 14652 commands "consensus" in the JTC1 sense.

 

A major point of discussion are US comments TECHNICAL #50 and TECHNICAL #51. The final result is acceptance in principle of those comments with clarification on what is meant by "technical grounds" -- this applies only to non-conformance to the standard as such. The editor will clarify the text suggested by #50 accordingly.

 

After this discussion, Joan signs off.

 

The discussion is captured in the DoC.

 

15.Other business

 

 

Present by phone: Ken, Alain

 

15.1A few ideas on sorting

1014 is briefly discussed. Marc is invited to provide a fuller working document for the next meeting that will elaborate on details of the scope (action item).

 

15.2Browsing and Matching

Alain: We should look at the issue of searching in WG20.

 

Marc: This is being handled on the European level (cf. N994, the Programming for Cultural Diversity). However, the mechanism is open for all.

 

15.3i18n and Cultural Diversity

 

Resolution "WG20 notes that internationalization depends on mutual respect between and peaceful coexistence of the different cultures of this world." This resolution is briefly discussed and no opposition is voiced at present.

 

15.4 Functionality of i18n

 

First start by accumulating information on existing i18n functionality in programming languages inside ISO and beyond. Such

work would be placed on the website. Building on that consider new i18n functionality that should be required.

 

Popular libraries such as ICU are also within scope.

 

AI on Keld: Will raise this issue at the next C and C++ WG meeting and will collect such information and place it on the web.

 

15.415.5 Marketing

 

Suggestion of Keld to open up the mailing list for WG20 to the general public. Ken points out that there might be a wish to create a dedicated mailing list rather than opening up the WG20 list in general. There will be more discussion on this on the WG20 list.

 

Alain points out that during the last JTC1 plenary, Frank Farrance  presented an overview of popular standards and their references on the web. 14651 was amongst the most often referenced standards.

 

Marc stresses that one of the best methods to attract users is to provide interesting contents. The comparative i18n capabilities of programming languages can be part of such content.

 

The Unicode site in 2002 had an average of 11.000 hits per hour.  

 

16.Review of Priorities and Target Dates

 

There are a number of action items from this meeting. However, the only topic that needs concrete target dates is 15897. Ken suggests that Keld writes up the disposition of comments and send it to the editing group for review. After that, the DoC will be declared final. The editor will then have the opportunity to implement the DoC and send it to the editing group for review.

 

Ken: We should go afterwards for a CD-3 because of the large number of fundamental opposing comments primarily for the US. Especially the versioning mechanism and the role of the JAC is at present undetermined. Keld: We should opt for an FCD. Marc seconds this provided that the DoC will be satisfactory to the editing group.

 

Marc: The editing group should be empowered to decide if the DoC is satisfactory and that the DoC is satisfactorily implemented. Then we can see if the document might command more consensus. Then we can decide whether to go for a CD-3 or an FCD.

 

Editing committee for 15897: Keld, Alain, Kim, Ken, Marc, Joan

 

Dates: 2003-02-21 to finish the DoC, 2003-02-28 for the review of the DoC, the editor will deliver the final DoC by 2003-03- 03. The resulting draft by 2003-03-17. Comments on that draft by 2003-03-31. Then the editing group will come to the decision whether to go for CD or FCD. Editor to deliver final text for ballot to convenor by 2003-04-10.

 

Allocated document numbers: N1019: preliminary DoC, N1020: final DoC, N1021: final draft for ballot.

 

The next SC22 plenary is on: 2003-09-15 in Oslo. Keld will represent WG20 at that meeting.

 

 

17.Review of Actions Items from this meeting

 

 

18.Approval of Resolutions

 

All resolutions were approved unanimously unless otherwise noted

 

19.Adjournment