ISO/IEC JTC1/SC22/WG15 RIN: POSIX Rapporteur Group on Internationalization Minutes of Rotterdam meeting Date: Monday 13th May 1991 09:30 - 17:00 Tuesday 14th May 1991 09:00 - 12:30 Location: Digital Rotterdam, Max Euwelaan 60, Brainpark NL-2067 MO Rotterdam, The Netherlands Attendance: Ralph Barker (US) US invited expert; UniForum subcommittee on internationalization; 1003 internationalization coordinator Christian Brack (Germany) Observer Patric Dempster (Canada) rapporteur Dominic Dunlop (UK) rapporteur Akio Kido (Japan) Observer Randall Howard (Canada) Canada invited expert Greger Leijonhufvud (Sweden) UK invited expert; technical reviewer, 1003.2 Yashusi Nakahara (Japan) Observer Nobuo Saito (Japan) Rapporteur Keld Simonsen (Denmark) Rapporteur. Donn Terry (US) rapporteur Erik van der Poel (Japan) Japan invited expert Johan van Wingen (Netherlands) SC2 liaison to SC22 1. Opening of meeting. 1.1. Introductions and roll call of technical experts The meeting convened at 09:30. Those attending introduced themselves. (See list above.) 1.2. Selection of chair, secretary & drafting committee Chair: Keld Simonsen Secretary: Dominic Dunlop Drafting: Patric Dempster, Shigekatsu Nakao, Donn Terry 1.3. Temporary document number assignment Temporary document numbers from the series RRTN (Rotterdam RIN Temporary Number) were assigned to materials provided to those attending the meeting. 1.4. Adoption of agenda and allocation of documents on agenda The agenda was accepted as appearing in attachment 1 to this document. Further levels of headings have been added in these minutes for ease of reference. 1.5. Approval of minutes The minutes of the October 1990 London meeting, as published in WG15 N129, were accepted as a true record. 2. Status, liaison and action item reports 2.1. Action Items Action items were not numbered in the minutes of the London meeting; the following numbers have been derived by numbering the leading set of initials which appears against each action item in 6.3 of those minutes. RIN9010-1 Ralph Barker: Investigate the possibility of links between wg15rin and i18n mail lists and the uniforum-intl mail list. Complete. In the future, the circulation of X/Open/UniForum internationalization material may be limited to matters discussed in joint meetings. RIN9010-2 Ralph Barker: Send Keld Simonsen mail list serializing/archiving software used by UniForum at Sun Microsystems. Complete. RIN9010-3 Dominic Dunlop/Keld Simonsen: Send rapporteur list to Ralph Barker by electronic mail for coordination purposes. Complete. RIN9010-4 Dominic Dunlop: Obtain example costing figures on TCOS institutional representative responsibilities from John Quarterman, and forward to Keld Simonsen.. Obsolete; considered complete. RIN9010-5 Dominic Dunlop: Write up minutes, agenda and attendance list (JTC1/SC22/WG15RIN-N031) and resolutions (JTC1/SC22/WG15RIN-N032, circulate by electronic mail, and forward to convener for entry in document register. Complete. RIN9010-6 Dominic Dunlop: Summarize issues raised in discussion of Japanese locale; circulate to group. Not complete.; overtaken by events. RIN9010-7 Greger Leijonhufvud: Add % descriptor to date format descriptions, allowing for explicit printing of AM or PM (or their equivalents in a particular locale).. Complete. RIN9010-8 Amend 9945-2 draft to replace the LC_DATE concept of abbreviated month name with alternate month name. Complete. RIN9010-9 Greger Leijonhufvud: Investigate the possibility of links between wg15rin mail list and UNIX International mail lists. Not complete; deemed not necessary RIN9010-10 Shigekatsu Nakao: Request that a representative of X/Open Company Ltd. provides a report to RIN either in written form, or in person at the next meeting of RIN. Not complete. It was not clear that Shigekatsu Nakao was informed of this action after the London meeting. The action was transferred to Ralph Barker, who works more closely with X/Open. RIN9010-11 Keld Simonsen: (Ongoing) Continue canvassing European opinions on and solutions to the problem of accommodation of IS646 within the framework of POSIX. Complete. A report appears under 4.5 below. RIN9010-12 Keld Simonsen: capture RIN comments on informally-circulated 1003.2 draft 10, pass them back to Hal Jesperen, and subsequently circulate amended result to selected members of WG15 etc. Complete. RIN9010-13 Keld Simonsen: Send copy of proposal on escape character substitution to Donn Terry by electronic mail. Complete. RIN9010-14 Keld Simonsen: Include issue of collision of Yen-symbol (=Y) and escape character in discussions of seven-bit coded character set issues. Complete. RIN9010-15 Keld Simonsen: Pass additional materials on collating, such as further papers from Alain LaBonte of CSA, to Greger Leijonhufvud for use in production of future drafts of 1003.2. Ongoing RIN9010-16 Keld Simonsen: Promote seeking of TCOS Institutional representative status by EUUG. Complete. RIN9010-17 Keld Simonsen: Arrange to put serial numbers on all postings to mail lists. Complete. RIN9010-18 Keld Simonsen: Set up public mail-list, i18n@dkuug.dk,; announce to those who need to know and tell them to tell their friends. Complete. Patric Dempster requested that multiple postings be consolidated into digests (shell archives), but, as there was no interest in this from other RIN members, no action was taken. Keld Simonsen agreed to add Randall Howard and others to the wg15rin and i18n mail lists. RIN9010-19 Nobuo Saito: Produce paper on issues of locale production for group. Ongoing RIN9010-20 Donn Terry et al: (Ongoing) Continue work on questionnaire. Complete. RIN9010-21 Donn Terry Review following topics, and, where necessary, amend draft questionnaire to solicit specific information: -- Need for locale-dependent binary time and date information from system (struct tm) -- Interest in and requirements for LC_MEASUREMENT locale -- Interest in greater flexibility in representation of AM, PM and related concepts, including noon and midnight, in output of date and input to at. Ongoing. -- struct tm: review at this meeting -- LC_MEASUREMENT: see questionnaire -- AM, PM: review at this meeting RIN9010-22 Donn Terry: Advise X/Open that usage of upper-case for LANG codes in XPG is contrary to the recommendations of IS 639:1988, Code for the representation of names of languages; and that @ () should not be used in locale names, as it is an IS 646 national variant symbol. Complete. It appears that X/Open is now observing these guidelines. RIN9010-23 Donn Terry: Investigate the possibility of links between the wg15rin mail list and X/Open mail lists. Complete. RIN9010-24 Donn Terry: Request X/Open to give consideration to the usefulness of an LC_MEASUREMENT locale, and to its potential contents. Complete. There has been no response at yet. RIN9010-25 Donn Terry: Respond in writing to Keld Simonsen' proposal on alternate 7-bit representation. Complete. This issue was discussed under item 4.5. RIN9010-26 Donn Terry: Review following topics, and, where necessary, amend draft survey to solicit specific information. Complete. (Spurious repetition of 21 above.) RIN9010-27 Donn Terry: Request 9945-1 technical editor to report informally on the impact of incorporating the character-related terminology outlined in JTC1/SC22/WG15RIN-N030 into 1003,2. Complete. This issue was discussed under item 4.7. 2.2. Report on internationalization IEEE 1003.x status 2.2.1. 1003.0 Ralph Barker reported on 1003.0, which has recently taken time to consider internationalization. It has been decided that, unlike other topics, internationalization-related material cannot entirely be confined to a single section of the document. The intention is that the guide should address the needs of those not familiar with the concepts behind internationalization, so portability of an application across cultures will be differentiated from the portability of an environment or platform between cultures. Work has also been concentrated on normalizing internationalization-related terminology throughout the guide. The use of language has also been tuned, for example, by replacing must with should, as befits a guide rather than a standard. 1003.0 will go to mock ballot on revision 13 or 13a later this year. TCOS SEC has elected not to form an internationalization working group. Instead, UniForum will manage a coordination function with all TCOS working groups. 2.2.2. 1003.1 Donn Terry reported on 1003.1. The main focus of activity has been on programming language independence; very little has been done on spoken language independence. A topic with an internationalization aspect, the 1003.1a labeled tape transfer format, is now responsibility of 1003.2. (Note that tar and cpio formats stay with 1003.1.) Greger Leijonhufvud noted that TCOS has put the matter of messaging into abeyance until 1992: the conflicting proposals presented during 1990 were judged workable, but inadequate both in terms of facilities and extensibility. (For example, source format; use from arbitrary programming languages; user interaction.) Hopefully, an improved proposal will be seen during the next year. WG15, suggested Greger Leijonhufvud, should follow a similar policy, but keep a watching brief through a standing item (see 6.4). Ralph Barker was concerned that messaging work in the X Consortium , which was unlikely to be compatible with anything in the international arena, might preempt us. It was agreed that we should monitor this issue. Patric Dempster commented that messaging would benefit from an object-oriented approach. In response to a question from Nobuo Saito, Dominic Dunlop and Keld Simonsen commented that, although it was useful for modeling, we should not mandate this approach, as it would preclude many language bindings. (See also 2.5 below.) 2.2.3. 1003.2 Greger Leijonhufvud noted that HLJ plans to speed up the circulation of updates to 1003.2, working to a two-month schedule: 30 days to circulate and comment; 30 days to process comments. It is hoped that draft 11a, due before the end of this year, will be judged suitable for circulation as a CD. (See also 3.2.) 2.3. Liaison reports: SC22 internationalization WG20; SC22 SWG on character sets 2.3.1. WG20 Keld Simonsen reported that WG20 has been approved and formed; its secretariat is ANSI; its convenor is from IBM; and its next meeting is in late August. The brief from SC22 is very broad; there is as yet no programme of work. (All 3 Japanese experts present at this RIN meeting are also on Japanese WG20 working group.) Now that WG20 exists, the group agreed to reaffirm RIN London resolution proposing that WG15 internationalization work is used as a base document for its work. The group also agreed to propose Keld Simonsen as WG15 liaison to WG20. 2.3.2. SWG Keld Simonsen reported that the SC22 special working group on character sets met in October, 1990, but did not reach any conclusion. A group convened to propose NWIs to SC22 has yet to produce any result. At a joint meeting with SC18 and SC2, a request was made for unique character naming. This has been done in DIS 10646; short names, requested at the same time, are judged to be application-specific, and should be handled by WG20 in SC22. Future SWG meetings will occur in conjunction with WG20. 2.4. TSG1 internationalization Activity Nobuo Saito reported that TGS1's main-line conclusion is that, in order to handle internationalization, it is better to define profiles for application portability than to define base standards. This proposal will be discussed at JTC1 plenary in October. Hopefully, some of the work of TSG1 will be passed to WG20 for further development. Erik van der Poel reported that the WG15 ad hoc on coordination had discussed little on internationalization beyond the desirability of international profiles being orthogonal to application portability profiles. See 4.1 below. 2.5. UniForum/X/Open Joint Interest Group Greger Leijonhufvud and Ralph Barker stated that, because of a high level of overlap between the internationalization-related activities of UniForum and X/Open, the two working groups had been merged, so broadening the field of experience available to both organizations and reducing the number of meetings that participants needed to attend. At a meeting in March, the groups reviewed their priorities, deciding that X Consortium-related work was of the highest importance. This led to considerations of RPC and of distributed locales. In April, this decision was further refined: three sub-groups will work on the production of documents by the end of 1991: on program-local locales; on locales in an RPC environment; and on locale registration. These documents will be circulated to RIN for information and comment at the earliest opportunity. Keld Simonsen drew the group's attention to related developments in WG14 and elsewhere. DKUUG has been suggested as a registry for locales, since it has previously expressed an interest. (See 4.2.) Work related to distributed locales has been divided into the following mechanisms: announcement for locales; announcement transport (protocol); data consistency; locale consistency; and possibly query support; negotiation; and management. Data tagging is not being addressed, nor is a general solution to the problem of multi-lingual systems. Currently, X11 is quite close to satisfying these requirements; RPC is not. Distributed locales are required for lightweight processes: the current per-process model is not adequate. The same applies to separate windows (or, indeed, parts of windows) belonging to a single process. An unfortunate consequence of distributed locales is that it appears to be necessary to introduce new interfaces (wisupper(), xisupper() etc.); work to date on per-process locales has managed to avoid this. (Removal of existing functionality is not a goal.) On the positive side, commented Randall Howard, this will allow the introduction of a new, and hopefully more elegant, model. The next meeting of the joint group, hosted by IBM, is in Toronto on 22nd-26th July. RIN members are welcome to attend. 2.6. X/Open internationalization Greger Leijonhufvud reported that work on X/Open is focussing on edition 4 of The X/Open Portability Guide (XPG4), and that this is being delayed by C language- and 1003.2-related issues. The X/Open Internationalization Guide, which references XPG4, is held up as a consequence. It will not be published until XPG4 is frozen, which will happen soon after the technical content of 1003.2 has stabilized. 2.7. PortSoft Erik van der Poel described PortSoft, a pan-Asian grouping formed to promote software portability, and to ensure that Asian needs are addressed by international standards. (See also WG15RIN-N033.) PortSoft is collecting culture-dependent information and national characteristics from all over Asia. It will subsequently make recommendations to standards bodies with the aim of producing an understanding of Asian needs. Later work will produce tutorial materials and promote software portability, both across platforms and across cultural boundaries. Due to the voluntary nature of the group, there is as yet no firm production schedule for documents. 2.8. Coordination letter Donn Terry has built up a list of organizations interested in internationalization, some of which may be doing work which overlaps That of RIN. He suggested that the group should circulate its minutes to those groups so that they know what RIN is doing, and learn the names of groups that RIN believes is doing related work; Ralph Barker suggested that RIN should encourage other groups to send us their minutes or corresponding documents so that RIN can know what they're doing. (See action items, 6.3.) 2.9. Other internationalization activity There were no reports. 2.9.1. INSTA Nordic culture This formal group under the Nordic Ministry Commission is collecting Nordic locale information. As yet, there is no planned deliverable (document etc.). The group intends using the RIN questionnaire as a basis for its work, and to coordinate with RIN. 2.9.2. WG14 C language At last year's London meeting, WG14 resolved to produce an addendum to IS 9899. The three topics to be covered are 1) (UK) More precise definitions of undefined behaviour; 2) (Japan) multibyte support; and 3) (Denmark) alternatives to trigraphs. The UK is handling technical editing. A PDAM may be available after this week's meeting in Tokyo. 2.9.3. UNIX International Yashusi Nakahara walked the group through the UNIX International internationalization/localization guidelines, (WG15RIN-N038) which are targeted at the designers of internationalization interfaces. The document is closely related to the X/Open Internationalization Guide, but, rather than being XPG4-specific, addresses the internationalization of all existing software. In particular, it addresses GUIs: while concentrating on Open Look, its findings are likely also to be relevant to other environments, such as Open Look. Erik van der Poel drew the group's attention to the document's requirement for ISO 2022 support in electronic mail, and hoped that UI's work is being coordinated with the Internet Engineering Task Force working group. (This is the group which coordinates the production of requests for Comment.) 2.9.4. JUS Userreq is JUS' working group on application portability, particularly of Japanized applications across programming environments. Taking a practical approach, the group is building a set of internationalized tools based around gcc. 2.9.5. C++ (WG21) Since internationalization is an important part of the C++ standardization activity, it was proposed that Keld Simonsen, who attends WG21 meetings, should informally pass relevant information back to this group. A formal liaison is not considered necessary now, but may become so in the future. 3. Balloting activities 3.1. 9945-1 ballot internationalization issues There was little discussion on this topic, as there is currently no balloting activity. Patric Dempster reminded the group that, as Canada had voted in favour of the acceptance of 9945-1:1990 on the understanding that file name interchange issues in archives would be addressed in the future, this matter should be covered by 1003.2, since responsibility for the pax format has been transferred to that group. CSA may elect to do further work on this project itself. 1003.1a will introduce a second alternate TZ format, answering concerns raised by Canada in London. 3.2. 9945-2 ballot internationalization issues 3.2.1. POSIX.2 draft 11 3.2.1.1. Definition of byte Greger Leijonhufvud stated that, for technical reasons, current draft (10) of 1003.2 includes all definitions from 1003.1. This includes the definition of a byte, which 1003.1 took, in turn, from ANSI C. This definition implies that the basic character set should be a single byte wide. This is too restrictive for POSIX. The current proposal is that a byte is the smallest addressable object of an architecture, and that it should be at least an octet. Support of 8-, 9- and 10-bit bytes was judged to be important. (Donn Terry mentioned that there had been some discussion of 7-bit bytes, but that architectures using them had been judged to be of historical interest only.) Bytes wider than 8 bits present problems for octal, decimal and hexadecimal escapes (\177, \d127, \x7f). The proposal is that such escapes should refer to octets -- the low 8 bits of each byte -- rather than to bytes. Donn Terry queried the usefulness of standardizing the backslash notation at all, since most uses of it would not be portable. GS contended that we should do the best we can for the most common case: 8 bit bytes, and that the notation was needed in localedef in any case. The group agreed. 3.2.1.2. Charmap Implementors have been worried by the possibility that users might change their charmap. Consequently, the current draft requires only that the conformance document specifies the extent to which users may change charmaps (if at all). This leads to the concept of private and public charmaps: private are user supplied, and may not be supported fully or, indeed, at all; public are implementor-supplied, and shall be fully functional. The limiting case is an implementation in which just one public charmap is be supplied and that private charmaps are not supported. All ballots received on the possibility of allowing changes to the comment character were negative. However, this functionality is likely to stay, as it is considered that to satisfy these objections would raise a greater number of others. 3.2.1.3. Regular expressions There was a serious error in the definition of longest leftmost match for regular expressions in the last draft of 1003.2. This will be fixed. The issue of when '$' and '^' are special in regular expressions is contentious. Some want 'ab$cd' to be allowed ('$' not special); others want it to be illegal (as it is in extended regular expressions). Traditionalists counter by saying that this would break too many existing scripts, and will probably win the day. RIN is happy with this situation. The result of the application of a regexp to a sequence of characters containing an embedded null is currently permitted; there has been an objection to this, as current practice in the C language and utilities written therein is that null is special. This suggests that the issue is language-dependent: RIN is in favour of putting language in the LIS which does not require that null is special, but allowing bindings to make it (or perhaps some other character) special if they wish. tr no longer knows about multi-character collating sequences, or, indeed, anything much relating to regular expressions. 3.2.1.4. Contents of locale In LC_MESSAGES, there has been a proposal that, to help shell scripts, yes_expr and no_expr should be shell expressions, not regular expressions. This is judged to be elegant, but of little use to the other languages which access locale information. In LC_MONETARY, there have been several proposals that the current keywords be replaced with a COBOL-like picture. However, this would erroneously imply that a real editing mask is being specified. Greger Leijonhufvud has canvassed a half-way house (an ordered list of keywords), but it appears that the consensus is to leave things as they are. 3.2.1.5. Other issues There has been some question over whether the day of the month in the default output of date has a leading zero, a leading space, or is left-justified. (Implementations vary.) The resolution is that it has a leading space. [Secretary's note: KS thinks a leading zero was agreed. Other recollections are sought.] pax -e, which would encode the names of characters outside the portable filename character set in the names of files written to an archive, is likely to be dropped on the grounds that it is too exotic to invent at this stage. Keld Simonsen and Patric Dempster objected, saying that there appears to be a requirement, at least from Canada and Denmark. Dominic Dunlop suggested that there is a real need for at least a prototype implementation before further standardization work can proceed. Greger Leijonhufvud and Ralph Barker added that pax is a minor aspect of the far larger problem of data announcement; solving the problem for pax is insufficient. Donn Terry commented that the coming tape interchange format may solve the problem by including a mechanism for character encoding announcement. Discussion of this issue will continue off-line between interested parties. 3.2.2. localedef Despite changes in the current draft, some balloters persist in saying that localedef should be jettisoned in its entirety. (Arguments are that it's out of POSIX' scope; it's administrative and hence the responsibility of 1003.7; it's not prior art.) A particular problem is the substitute command, and its use of regular expressions. It has been suggested that string-for-string substitution would be adequate; however, the CSA -- and, by implication, most western -- collation standards cannot be met without regular expressions. Given rationale that regexps are not necessary for practical national collation sequences, Greger Leijonhufvud would be happy to drop them. (See actions.) There will be a BNF (yacc) grammar for localedef. 3.2.3. POSIX.2A draft 6 There was no discussion on this topic. 3.3. Other POSIX documents 3.3.1. 1003.8 Erik van der Poel raised the issue of character set translation in transparent file access, a topic which 1003.8 has chosen not to address so as to allow it quickly to produce a standard. While SC21 addresses this topic through facilities in FTAM, and the more restrictive flavour of TFA presumes underlying facilities similar to FTAM, it is not clear that all involved in the various aspects of the work are aware of the effect of their decisions on others. After contributions from a number of experts, it was decided that data interchange was a suitable topic for the standing issues list, and that Keld Simonsen should search for information on internationalization-related topics in SC18 and SC21. This may lead to requests for the establishment of formal liaison in the future. 4. New business 4.1. Example national profiles and locales 4.1.1. Japanese national profile Yashusi Nakahara introduced a Japanese National Profile draft, a mature document, and a draft guideline for national profile, which is preliminary. (At the request of the Japanese Rapporteur, permanent document numbers were not assigned to either document.) A national profile has been difficult to define, as it must cover a larger problem area than simply POSIX. General issues presenting difficulties in connection with the Japanese profile are discussed in 1.5 of the Japanese National Profile. 2 raises issues specific to 9945-1. None of these difficulties is insuperable, however. Once work on the national profile is complete, it will be used as a basis for a JIS POSIX standard. Erik van der Poel pointed out that, while discussion on a number of locale-related issues had started on the mailing list, closure had not been reached. For example, it seems unnecessary to require, or, in some cases, even to allow, constructions such as , where both characters are part of the invariant part of IS 646. It seems desirable that the guidelines should contain more examples. An opinion was needed from Greger Leijonhufvud, but, by the time of RIN's discussion, he had left the meeting. There was a discussion of short character names, particularly in the context of Japanese glyphs. The draft guidelines distributed at the meeting not currently discuss this issue; suitable material should be added. (No permanent document number was assigned to the draft guidelines, as thieir authors wish to develop them further before submitting them as an official document.) The issue of short character names is within the scope of work of WG20, and will be handled in that forum by Keld Simonsen. (See also 2.3 above.) Erik van der Poel also queried the allowed content of comment fields in charmap files. RIN believes that comments are merely a sequence of bytes -- and so can shift to any character set -- but cannot continue onto subsequent lines. This should be checked. Erik van der Poel was strongly against the inclusion of LC_CTYPE in the locale. Surely, he contended, the issues of character set and locale should be orthogonal. Keld Simonsen agreed that locales should be written so as to be independent of particular character sets. Summarizing discussion, their inclusion is historically due to X/Open and judged to be useful, but might not have been done in the same way, were we starting design today -- but we would probably still want to duck the general issue of data announcement in each data set. Johan van Wingen considered it essential that repertoires were identified in the locale, and pointed out that the issue was not POSIX-specific. SC2, in its current revision of IS 2022, is addressing it, but in a rather narrow, data-stream oriented manner. 4.1.2. Danish national profile (Keld Simonsen) Keld Simonsen briefly described current status, stating that updates had been incorporated into IEEE P1003.2, draft 11. 4.1.3. Guideline for national profiles (This had also been discussed under 4.1.1 above.) Yashusi Nakahara introduced a draft guideline for national profile, pointing out its similarity to OSI profiling work. Questions discussed during its formulation had included the issue of whether subsetting or supersetting should be allowed (not desirable); how orthogonal should be national profiles and application environment profiles (as much as possible); the geographic or ethnic scope of profiles (not solely national); the portability of national profiles (as great as possible); and the code set independence of national profiles (highly desirable). Ralph Barker commented that it would be desirable to synchronize the national profile-related aspects of 1003.0 with the content of the draft guideline. In particular, the definitions of 1003.0 may be of use in the guidelines. Patric Dempster queried whether code-set independent profiles were practical from the point of view of portability. Donn Terry and others argued that they were, and that the invariant subset of IS 646 was the least bad answer to the problem, and one which worked (except for a slight wrinkle involving ! in EBCDIC). The group concluded that the issue should be documented in the guide. Randall Howard queried allowing charmap comments using some richer codeset rather than simply invariant 646 characters, but was rebuffed by others who contended that such asceticism would seriously limit acceptance of the locale concept by making the maintenance of locale information more difficult for many developers. Erik van der Poel and Patric Dempster suggested that the contents of locale description files should be non-normative annexes to national profile standards, so as to avoid mandating links to any particular code set. Dominic Dunlop countered that, while apparently increasing the portability of the specification, in practice this would render it far less useful and usable. Dominic Dunlop queried the ultimate disposition of the guidelines, which are currently solely for WG15 internal use. Yashusi Nakahara and others see the document as eventually becoming a TR. However, there is no need to start this process just yet. The registration of locales is also an issue. Keld Simonsen believes that this is within the scope of work of WG20. X.500 may also be relevant, both for the provision of a locale enquiry service, and for naming the domains to which locales belong. Keld Simonsen, with support from Erik van der Poel, resuscitated the proposal for a replace statement in locale definitions, facilitating the harmonization of locales. WG15RIN-N035 also addresses this issue. Those close to the IEEE balloting process warned that the introduction of such a new feature -- or, indeed, any new feature -- at this stage would be counterproductive. Randall Howard suggested that the best way to gain acceptance of additional functionality would be to implement, then place the result in the public domain. 4.1.4. Coordination ad hoc Yashusi Nakahara considered it appropriate to discuss national profiles at profile coordination ad hoc meetings, even though the primary function of these meetings is to coordinate AEPs. It is not clear that those concerned with AEPs are cognizant of the function and implication of national profiles, and of their (orthogonal) relationship to AEPs. 4.2. Collection of national locales and profiles Erik van der Poel stated that PortSoft will attempt to collect locales from its members for contribution to the informal archive maintained by Keld Simonsen at DKUUG. The matter of what to do with locales once they have been collected is to be discussed in WG15 plenary, having first been reviewed by a break-out group. 4.3. Revised questionnaire Rather than discuss the document in detail at this meeting, Donn Terry requested that all present critique the current content, responding by electronic mail in good time for detailed discussion at the next meeting. Those responding should bear in mind that the function of the document is to discover areas that RIN is not addressing, rather than to determine the level of usage of those features that have already been proposed. Patric Dempster queried the continued omission of questions on units of measure. Donn Terry suggested that he should propose a suitable addition which he could incorporate in the questionnaire. 4.4. ISO 10646 Johan van Wingen described the status of SC2 activities. Unicode will never be an ISO standard, as insuperable differences related to the concept of characters exist between Unicode and ISO coded character set standards in general, making it impossible to fit Unicode into the ISO framework. Voting will be closed on DIS 10646 on 6th June, meaning that SC2/WG2 cannot currently work on the document. It is expected that there will be several negative votes requiring changes for their resolution. Objections include: the details of the encoding, and of the unification or otherwise, of south-east Asian characters; the retention or deletion of the full set of C0 and C1 control characters; and the concept, inclusion and handling of floating diacritical marks and the resulting implication of open repertoires. On the last point, Johan van Wingen's opinion is that requirements for floating diacritics emanating from the US Research Libraries Group are spurious when considered in the light of that organization's current practice, and argues that national member bodies should oppose their inclusion. Patric Dempster commented that Canadian sentiment in favour of Unicode is influenced by considerations of storage requirements, and that he would welcome counter-arguments. Keld Simonsen considers that, for POSIX internationalization work, there is a pressing need for an international standard in the area of multi-octet coded character sets. Johan van Wingen contends that it is more important to do it once and to do it right than to to do it now. No consensus was reached on which of these views RIN supported. 4.5. Invariant ISO 646 support Keld Simonsen informed the group that TR10176:1991 recommends that source code for all SC22 languages should be writable using only the invariant portion of 646, if only as an alternative to a more agreeable format using a richer character set. After a rehearsal of the issues, particularly as they pertain to 1003.2/9945-2 commands, Randall Howard expressed the view that this should be handled by a filter in the input stream, not by the little languages themselves. Dominic Dunlop queried whether RIN, given longstanding existing practice to the contrary, agreed with the recommendation of the TR. To retrofit a solution would be of no use to the existing body of scripts (programs) which do not confine themselves to the invariant set; almost all new programs would be created by developers using a richer character set, particularly since the population of 7-bit terminals is diminishing. Donn Terry added that the imposition on implementors of 9945-2 conforming environments would be considerable, leading to objections. These and other reasons for the rejection of DS' initial proposal for a solution, submitted to 1003.2 during the mock ballot three years ago, are not documented. RIN should agree on a position at or prior to its November meeting, and, if appropriate, submit it to 1003.2 for formal comment. 4.6. Internationalization background paper Due to time constraints, there was no discussion of this issue. 4.7. Byte terminology See 3.2.1 above. 4.8. Removal of pax -e translation See 3.2.1.5 above. 4.9. Multibyte issues Yashusi Nakahara suggested that future revisions of IS 9945-1 should be harmonized with Japan's proposed addendum to IS 9899. Dominic Dunlop pointed out that future revisions of IS 9945-1 would probably be language-independent, and so effort expended now on harmonization with an amended C standard might be wasted. Clearly, a future C binding to IS 9945-1 services should be harmonized. It is unfortunate that WG15 is currently without a timetable for the bringing forward of language-independent material. 5. Review/approval of resolutions and action items 5.1. Recommendations for WG15 The resolutions appearing in JTC1/SC22/WG15RIN-N037 were approved. 6. Closing procedures 6.1. Future meeting considerations -- request for invitations Dominic Dunlop and Patric Dempster commented that meetings needed two or even three days. Johan van Wingen suggested that a meeting could be held in conjunction with SC22 plenary in August. Ralph Barker proposed a meeting adjacent to the Toronto X/Open/UniForum meeting in July. Yashusi Nakahara argued in favour of synchronizing with WG15 meetings so as to minimize travel costs. It was agreed that the next meeting should be from 3rd-5th October, 1991 in Stockholm, prior to the WG15 meeting. (Note that the 3rd is a Sunday.) No further meetings were scheduled. 6.2. Permanent Document number assignment The following document numbers were assigned: N033: Introduction to PortSoft N034: Japan UNIX Society: Introduction for the WG15 Profile Coordination ad hoc N035: Proposal for building another locale N036: Minutes from Rotterdam N037: Resolutions from Rotterdam N038: UNIX International internationalization/localization guidelines [Secretary's note: My notes suggest that N038 was assigned a duplicate number N034 at the meeting. ] 6.3. Review of Action items Readers should note that items have been renumbered since the circulation of a preliminary version of this list. RIN9105-1 All: critique new draft of questionnaire; respond by 1st August. RIN9105-2 All: Participate in UniForum/X/Open joint internationalization activity if sufficiently interested. RIN9105-3 Donn Terry, Erik van der Poel, Yashusi Nakahara: Monitor progress of coordination ad hoc, ensuring that issues pertaining to national profiles are brought to the group's attention where appropriate. RIN9105-4 Donn Terry: Circulate revised questionnaire document to group as soon as it becomes available. RIN9105-5 Donn Terry: Circulate revised tape archive format document to group as soon as it becomes available. RIN9105-6 Donn Terry: Circulate revision (action RIN9105-4 above) of questionnaire document to WG15 members for completion, and pass to Dick Weaver as an expert paper for consideration by WG20. RIN9105-7 Erik van der Poel: Circulate materials from PortSoft as they become available. RIN9105-8 Erik van der Poel: Determine whether substitute is necessary to implement Japanese collation. (See also RIN9105-20 below.) RIN9105-9 Erik van der Poel: Encourage production and contribution of national profiles by PortSoft members. RIN9105-10 Greger Leijonhufvud, Yashusi Nakahara: Do not require spelling out of ISO 646 invariant character names in locale guidelines; do not suggest in examples that this is necessary. RIN9105-11 Greger Leijonhufvud: Pass IEEE ballot objections to pax -e to Keld Simonsen for review. RIN9105-12 Keld Simonsen: Circulate organizations on Donn Terry's address list with RIN minutes; require organizations to let RIN know if they want minutes of future meetings; request them to send RIN their minutes. RIN9105-13 Keld Simonsen: Continue collection of national profiles, locales and charmaps in electronic form. RIN9105-14 Keld Simonsen: Formally submit for discussion at November meeting the proposal for 646 invariant character set support in 9945-2. (Already circulated by electronic mail.) RIN9105-15 Keld Simonsen: Informally provide feedback on internationalization-related issues from SC22/WG21 (C++). RIN9105-16 Keld Simonsen: Introduce to WG20 the work done in RIN on short character names. RIN9105-17 Keld Simonsen: Pass INSTA Swedish contact information to Greger Leijonhufvud. RIN9105-18 Keld Simonsen: Identify and report on activity on internationalization-related topics in SC18 and SC21. RIN9105-19 Keld Simonsen: Update wg15rin and i18n mail lists: add Randall Howard (and others). RIN9105-20 Patric Dempster: Clarify, through discussion with Alain LaBonte, whether the CSA ordering standard requires the substitute operation. (See also RIN9105-8 above.) RIN9105-21 Patric Dempster: Propose a measurement unit questions which Donn Terry could incorporate in the questionnaire. RIN9105-22 Ralph Barker, Yashusi Nakahara: Synchronize the national profile-related aspects, including definitions, of 1003.0 with the content of Draft guideline for national profile. RIN9105-23 Ralph Barker: Circulate copies of the internationalization section of 1003.0. RIN9105-24 Ralph Barker: Circulate materials from UniForum/X/Open internationalization group, including problem statements relating to distributed locale and multi-byte support, as they become available. RIN9105-25 Ralph Barker: Coordinate participation (item 5 above). RIN9105-26 Ralph Barker: Request that a representative of X/Open Company Ltd. provides a report to RIN either in written form, or in person, at the next meeting of RIN. (Nobuo Saito's former action RIN9010-10) RIN9105-27 Yashusi Nakahara: Add section discussing short character names to Draft guideline for national profile. RIN9105-28 Yashusi Nakahara: Add wording to Draft guideline for national profile stating that locales should be written so as to be independent of code set and that use of the codeset part of XPG's LANG specification should be avoided pending a more general solution to the problem of data announcement. RIN9105-29 Yashusi Nakahara: Check allowed format of charmap comments. RIN9105-30 Yashusi Nakahara: If acceptable to national member body, distribute revisions to Japanese profile to mail list both in nroff output and in troff -mm source form. 6.4. Review of standing items The dates preceding items identify the meeting at which they were added to the list. 1. (9105) Messaging. 2. (9105) Data interchange. 3. (9105) Improvement of POSIX internationalization support facilities. 4. (9105) Harmonization of C binding multibyte support with 9899 as amended. 6.5. Thanks to host The meeting having temporarily adjourned at 12:45 and reconvened to complete its business at 20:00 on 91-05-14, Keld Simonsen adjourned the meeting at 23:00 after thanking the host organization, support staff and officers. ISO/IEC JTC1/SC22/WG15 RIN Attendee list for May, 1991 meeting Attachment 1 to minutes Canada Patric Dempster (rapporteur) AT&T Canada Inc. 3650 Victoria Park Avenue Suite 800 Willowdate Ontario M2H 3P7 Canada Phone: +1 416 756 5116 Fax: +1 416 499 2422 Email: uunet!attcan!patric Denmark Keld Simonsen (Rapporteur) DKUUG Studiestraede 6 DK-1455 Copenhagen K Denmark Phone: +45 33 13 0023 Fax: +45 33 91 1828 Email: keld@dkuug.dk Germany Christian Brack (observer) SNI AG Otto-Hahn-Ring 6 P O Box 830951 D-8000 Munich 83 Germany Phone: +49 89 636 44922 Fax: +49 89 636 48976 Japan Akio Kido (observer) IBM Japan (LAB-S16-A) 1623-14 Shimotsuruma Yamato-shi Kanagawa-ken 242 Japan Phone: +81 462 73 5436 Fax: +81 462 74 6192 Email: jl01376@ymtvm8.iinus1.ibm.com Yashui Nakahara (observer) Toshiba Corp. 5-20-7 SORD Bldg. MASAGO SHIBA-City CHIBA 260 Japan Phone: +81 472 77 8670 Fax: +81 472 79 2628 Email: ynk@ome.toshiba.co.jp Nobuo Saito (rapporteur) Keio University School of Environmental Information 5322 Endo Fujisawa 252 Japan Phone: +81 466 47 5111 x3256 Fax: +81 3 461 6493 Email: ns@sfc.keio.ac.jp Erik van der Poel (observer) Software Research Associates Inc. 1-1-1 Hirakawa-cho Chiyoda-ku Tokyo 102 Japan Phone: +81 3 3234 2692 Fax: +81 3 3262 9719 Email: erik@sra.co.jp Netherlands Johan van Wingen (SC2 liaison to SC22) Postbus 486 2300 AL Leiden Netherlands Phone: +31 71 143739 Email: butpaa@hlerul2.bitnet UK Dominic Dunlop (rapporteur) Oxford University Computing Service 13 Banbury Road Oxford OX2 6NN UK Phone: +44 865 273280 Fax: +44 865 273275 Email: domo@ws.ox.ac.uk Greger Leijonhufvud (invited expert) Interactive Systems Corporation St JohnUs Court, Easton Street High Wycombe Buckinghamshire HP11 1JX England Phone: +44 494 472900 Fax: +44 494 472383 Email: greger@isc.com USA Ralph Barker (invited expert) UniForum 2910 Tasman Drive Suite 201 Santa Clara CA 95054 U.S.A. Phone: +1 408 986 8840 Fax: +1 408 986 1645 Email: rbarker@uniforum.org Donn Terry (rapporteur) Hewlett Packard 3404 East Harmony Road Fort Collins CO 80525 U.S.A. Phone: +1 303 229 2367 Fax: +1 303 229 2838 Email: donn@hpfcla.hp.com