jsh@usenix.org (Jeffrey S. Haemer) (09/29/90)
Submitted-by: jsh@usenix.org (Jeffrey S. Haemer) An Update on UNIX*-Related Standards Activities September 27, 1990 USENIX Standards Watchdog Committee Jeffrey S. Haemer, Report Editor U.S. TAG to ISO/IEC/JTC1/SC22 WG15 Susanne Smith <sws@calvin.wa.com> reports on the July 19, 1990 meeting in Danvers, MA: 1. Overview Before you ask, ISO/IEC JTC1 SC22 WG15 is ISO POSIX. The U.S. TAG is the United States Technical Advisory Group, which formulates the U.S. position on WG15 issues, and chooses the members of the U.S. delegation to the international WG15 meetings. This meeting began at 8:00 A.M. and ended before noon. This must be a record -- not just for the TAG, but for any standards group meeting. There were three major business items: - language independence, - document circulation procedures (yawn), and - rapporteurs. This short agenda, coupled with a determination to avoid an extra meeting, like the Denver meeting we were forced into in June, kept the discussion on track all morning. ISO POSIX: Winners and Losers The vote for 9945-1.2 (1003.1a draft 5) was unanimously in favor without substantive comments. If all goes well there just may be an IEEE version of 9945-1 available in Seattle. Let's all cross our fingers. Now that it's September I think we need to cross our toes as well. My last report mentioned the formatting problems with the 9945-1 document. The TAG had decided to request the formation of an ad hoc committee in Paris to try to resolve these problems. WG15 resolved to __________ * UNIXTM is a Registered Trademark of UNIX System Laboratories in the United States and other countries. September 27, 1990 Standards Update U.S. TAG to ISO/IEC/JTC1/SC22 WG15 - 2 - instruct the WG15 convener, Jim Isaak, to request written editorial requirements from the ITTF (formerly the Central Secretariat) and IEEE, and forward these to SC22. The emphasis here should be on written requirements. WG15 refused to register 1003.4, real-time extensions, as a CD (committee document, formerly DP, draft proposal) because it is not a language-independent specification. They were also concerned that the standard might have to change once there is a language independent version of 1003.1. 1003.5, Ada binding, and 1003.9, FORTRAN binding, suffered a similar fate for different reasons. 1003.5 and 1003.9 were held off until at least the October WG15 meeting because G15 had not seen the 1003.5 and 1003.9 documents, and were reluctant to register something they hadn't seen before. And again, they were concerned that these standards might have to be re-written once there is a language independent version of 1003.1. Administrivia Skip to the next section if you're easily bored or just not interested in bureaucracy. Why, you ask, was WG15 being asked to register something they had not seen before? Here are the steps that have to complete before a document gets circulated: 1. The committee and SEC approve its release. 2. The TAG approves its circulation. 3. The committee chair delivers the document to the TAG chair, Donn Terry. 4. The TAG chair forwards the document to the WG15 convener, Jim Isaak. 5. The WG15 convener distributes the document. 1003.5 and 1003.9 were approved by the TAG for circulation to WG15 during the April meeting in Snowbird. This left six weeks for for the documents to be circulated and read. The problem was that the TAG chair did not receive the documents in time to have them circulated before the meeting. To avoid this problem in the future, the TAG is going to ask the SEC to assign an action item to the committee chair so that there is a method to track this task. In other news: September 27, 1990 Standards Update U.S. TAG to ISO/IEC/JTC1/SC22 WG15 - 3 - - The TAG procedures were entered and marked up, and will be included in the next mailing. - The meeting in Seattle will start our new meeting schedule of Sunday from 6 to 10 P.M., all Thursday, and again Friday if necessary. Are You Ready for UNIX in VDM? We cannot delay language independence for 1003.1 any longer. It's now really holding up international progress on important POSIX efforts. But what format or technique should we use? ISO rules seem to require an ISO-standard method, which could restrict us to VDM (Vienna Definition Method), but no one thinks VDM will work. Paul Rabin and Steve Walli have been working on a method, but the TAG worries that a non-standard method will create problems like those we've suffered through with document formats (see last TAG report). In order to avoid rejection later we will circulate the new method in SC22 and WG15 for review and comment. To make this circulation useful, Donn Terry is listing specific questions for SC22 and WG15 to answer. [Editor: I believe that ISO rules only restrict us to VDM if we produce a formal definition, i.e., something from which one could do correctness proofs. Of course, rules and politics are not always the same thing and using VDM might help grease the skids. Still, we should stop and ask if not using VDM would hold us up any more than using VDM.] The TAG will also ask the WG15 convener to schedule an ad hoc meeting on language independence, during the October WG15 meeting, to help move it along. ``Rap, a-rap, a-rap, they call me the rapporteur.'' Rapporteurs are technical experts on specialized aspects of a particular standards effort. Their scope is usually broader than an individual standard, and they usually coordinate efforts in several standards bodies. WG15 has three rapporteur groups, one each for conformance, internationalization, and security. We send a representative to each. The conformance-testing rapporteur group will be looking at 1003.3 draft 12 (conformance testing), and the OSF-UI-X/Open Phoenix project as potential base documents for the ISO 9945-series documents. The Phoenix project is developing a conformance-testing platform. We will not have to decide whether we want to submit 1003.3 as a new work item in this area until 1991. Ralph Barker asked that UniForum be allowed to send him and one UniForum Internationalization Technical Committee member to the next internationalization rapporteur group meeting. This person would be subject to subcommittee approval but selected by UniForum. Worry September 27, 1990 Standards Update U.S. TAG to ISO/IEC/JTC1/SC22 WG15 - 4 - about the fact that the TAG would not choose this person evaporated when it became clear that Donn Terry would continue as internationalization rapporteur and that the UniForum members would just be an addition. The TAG appointed Al Weaver security rapporteur to fill the vacancy Terry Dowling left when he resigned in January. Summary The most important development is that the synchronization proposal discussed in the last report is already dead. This proposal was to have fed balloting responses from IEEE into WG15, and vice-versa, allowing WG15 approval to follow on the heels of IEEE approval. Now, while the IEEE is advancing, everything in WG15 is blocked by 1003.1 language independence. September 27, 1990 Standards Update U.S. TAG to ISO/IEC/JTC1/SC22 WG15 Volume-Number: Volume 21, Number 147
ok@goanna.cs.rmit.OZ.AU (Richard A. O'Keefe) (10/01/90)
Submitted-by: ok@goanna.cs.rmit.OZ.AU (Richard A. O'Keefe) In article <559@usenix.ORG> in comp.std.unix, jsh@usenix.org (Jeffrey S. Haemer) wrote: > We cannot delay language independence for 1003.1 any longer. It's now > really holding up international progress on important POSIX efforts. > But what format or technique should we use? ISO rules seem to require ************************* > an ISO-standard method, which could restrict us to VDM (Vienna **************************************************************** > Definition Method), but no one thinks VDM will work. Paul Rabin and ******************** > Steve Walli have been working on a method, but the TAG worries that a > non-standard method will create problems like those we've suffered > through with document formats (see last TAG report). In order to > avoid rejection later we will circulate the new method in SC22 and > WG15 for review and comment. To make this circulation useful, Donn > Terry is listing specific questions for SC22 and WG15 to answer. > [Editor: I believe that ISO rules only restrict us to VDM if we > produce a formal definition, i.e., something from which one could do > correctness proofs. Of course, rules and politics are not always the > same thing and using VDM might help grease the skids. Still, we > should stop and ask if not using VDM would hold us up any more than > using VDM.] My main interest here is in the ISO Prolog standard. I am confused by this extract from comp.std.unix, because the ISO Prolog standard contains a formal specification of Prolog. Personally, I would find it easier to read if it _were_ in VDM. Instead it's in a variant of first-order logic (exact semantics unknown) with a new syntax. This definition was developed with the explicit intention of permitting correctness proofs. Does this mean that ISO _will_ accept "make up your own formal specification language", or does it mean that the Prolog specification in the ISO Prolog draft is forbidden by ISO rules? Can someone who really knows clear this up? -- Fixed in the next release. Volume-Number: Volume 21, Number 156