January 20-21, 1992
Attending:
1 Opening of Meeting
Don Folland UK Roy McKean UK observer Pete Meier US observer Yasushi Nakahara Japan (for Prof. Saito) Arnie Powell CA Vilhelm Rosenqvist DK observer Don Terry US Willem Wakker EWOS (NL) Andrew Walker UK observer & WG15 to SGFS Liaison
The meeting was called to order at approximately 9:45, Monday, January 20th at the San Francisco Marriott, San Francisco, CA, USA.2 Administration, Status, Liaison and Action Item Reports
1.1 Welcome from Host
The host, Ralph Barker representing UniForum, was required to be elsewhere so the acting Rapporteur, Jim Isaak welcomed the participants for him.1.2 Introductions and roll call of Rapporteurs and Observers
The participants introduced themselves. The list above indicates the participants and their status.1.3 Secretary
Don Terry volunteered to be secretary.The drafting committee volunteers were Roy McKean and Willem Wakker. They were asked to be responsible for both the resolutions and the Terms of Reference (ToR) document.
1.4 Adoption of agenda
The agenda was modified (as reflected in these minutes) and adopted.In the process it was determined that several items were unnecessary. Since this was the first meeting of this group, there were no previous minutes to approve.
1.5 Need for Lead Rapporteur. Jim Isaak observed that as WG15 Convener, he was required to act as (lead) Rapporteur until one could be appointed. He urgently requested nominations from national bodies for individuals to fill that role.
2.1 Issues List It was agreed to carry the Issues List forward from the Ad-hoc held in Amsterdam last spring, with those numbers. The actual text of the issues appears in the final issues list in 5.1.
Issue 1. ... influence JTC1 ... Closed by resolution of ad-hoc. Issue 2 ... process for ... user. Still open. Don Folland brought forward a paper from the UK (DISC) as a basis for discussion. (RGCPA 10)
Willem Wakker suggested that the topic be left to the SGFS context: that we should think about whether this issue is appropriate for RGCPA. (He noted that it's the same set of people, just a different focus.) He suggested that RGCPA should concentrate on coordination, not investigation.Don Folland indicated that feedback from RGCPA goes upward anyway, so it may not be critical exactly where the work is done.
See AI 9201-01
Issue 3 ... gaps ... Closed. Some discussion of usefulness of the issues and the issues list led to a suggestion that the issues be re-reviewed once the Terms of Reference are done.
Issue 4 ... building upon ... View in context of ToR. See AI 9201-02
Issue 5 ... View in context of ToR. Issue 6 ... View in context of ToR. Issue 7 ... Possible new wording was suggested: Should the use of named options in POSIX standards be encouraged. View in context of ToR.
Issue 8 ... Closed. The content of the issue has been sent to SGFS. These were re-reviewed at the conclusion of the meeting, with some further changes and additions.
2.2.1 Action Item review The following action items were carried over from the Amsterdam ad-hoc meeting. 1. Jim Isaak to request X/Open, UNIX International and OSF each to provide a list of their interface specifications not in 9945, which they feel are required for a formal standard, together with an indication of priorities. Closed.
See also AI 9201-03.2. IEEE to provide a report on the options in 9945-1. Donn Terry reported that IEEE 1003.1 is working on the issue, pending an "options annex" format. The Drafting Committee was asked to look as issues of profilability and profile review in the ToR. Open, but progress. Closure would be by having 1003.1 provide an Options Annex for review.
3. X/Open to submit a report on its Xtra programme for capturing user requirements, and to suggest how the output could be used in setting priorities for profile development. X/Open responded with: if any standards committee would welcome input they can request information from X/Open, and X/Open will provide it. Closed.
4. Donn Terry to raise with .16 and .18 the question of whether or not there is a requirement for profiling C, including addenda in progress. Closed, pending .16 or .18 taking a relevant action.
2.2 JTC1 or SC22 actions affecting the group. The TSG-1 report was approved by JTC-1. It will affect us as the implications propagate through the standards system. It was observed that the work within SC22 on means for interpretations might affect us, but that was not yet clear.
2.3 Liaison Reports SGFS: The SGFS meeting in June asked for WG15 input, which became impetus for the authorized subgroup meeting in Brussels. We have minutes and various notes, as reflected in N013.
Walked through Andrew Walker's summary notes in N013, which summarize the isssues relevant to this meeting. The items found in that list were assigned to agenda items for this meeting.
It was determined that SGFS authorized subgroup issues 1, 5, 6, 10 were not relevant at this time. The drafting committee was asked to include in the ToR any aspects of issue 11 that applied to this group.
EWOS (EG-CAE): It appears that it has been formally approved to expand the scope of EG-CAE to OSE. TR10000.1-3 is being reviewed. They have presentaed a detailed work plan to the EWOS Technical Assembly. OIW: They did formally expand their charter for OSE. There are reports on the topic available (no reference numbers given). They appear to have changed their name from "OSI..." to "OS[E]" "...Implementors Workshop". (It was unclear whether it was "Open Systems" or "Open Systems Environment"...) See AI 9201-04
AOW: MITI thinking about posssible OSE focal points: the choices include AOW, POSIX WG (reorganized), Sigma. MITI has not determined the appropriate organization at this time.
2.4 Creation of Liaisons Liaisons were determined to be according to the following table. "To" indicates a liaison to the organization, "From" means from that organization. <Open> indicates that a liaison could be created, but that there is no-one to fill it. <Encourage> is a stronger veriosn of <Open>. -- indicates that a formal liaison is not possible. In all cases informal liaisons are encouraged, in addition to the formal ones. Liaisons (formally recognized bodies):
To Liaison From Liaison SGFS: Andrew Walker (already WG15) <Open> EWOS: -- Willem Wakker OIW: -- <encourage> AOW: -- <encourage> See AI 9102-05.
Coordination (other bodies): From Liaison PSC: <Open> (However, Folland, Terry are covering) X/Open: Roy McKean OSF: Fritz Schulz UI: Nakahara-san, Shane McCarron COSA: <Open> PortSoft: Eric van der Poel CCTA: Don Folland NIST: <Open> See AI 9102-06, 9106-07.
2.5 Participant reports Many had already been distributed, and more are expected. There was no discussion.
3 Project Activities
2.6 Rapporteur Group organization No issues discussed.
2.7 Status of Development in the US Member Body. 2.7.1 Expected approval of base standards. A chart (N017) was presented indicating that a large number of standards would be starting IEEE ballot during this year. 2.7.2 Profile project and plans The status of the various profiles in the US Member body (IEEE TCOS) was summarized. N015 and N016 do much of this. 1003.0 is expected to ballot in the spring to summer timeframe, as may 1003.10, 1003.12, 1003.14, 1003.18.
See AI 9201.09
2.8 TOR and Programme of Work An initial discussion for the guidance of the Drafting Committee was help. The outline was developed by identifying the specific work to do.
4 New Business
3.1 Review of SGFS ad hoc recommendations 3.0.1 issues raised by WG15 Discussion of TR10K-1.3. Willem Wakker provided a high level view of the issues of EWOS concerning it. Much of it concerned conformance clauses and the asymmetric nature of the interface. A topic variously referred to as horizontal vs. vertical, or serial/parallel was discussed. The key element is that the relationship between a nconforming implementation and conforming application is not peer to peer, but rather some form of up to down relationship. It was also noted that the "Finite State Automaton" on the conforming application end of this relationship is really a human (programming is a human activity), and thus not subject to any form of simple description.
It was noted that OSI stack profiles do not take well to having gaps in the middle (such as layers 2 and 4, but not 3), but that OSE profiles work just fine with (for example) database or a specific language missing.
See AI 9201-10, 9201-11.
The issue of taxonomy was discussed. Specifically it was asked (but not answered) if the proposed P profiles from EWOS are truly peers of the current F, A, B, T, U and R profiles, or whether some or all of those groups might belong at a lower level of classification.
It was unclear whether P profiles should reference OSI elements. It was noted that the F element appears to apply to OSE as well as OSI.
See AI 9201-12.
A number of issues were discussed as examples. The discussion was useful and educational, but no specific conclusions except that there was nothing to do at this time, although there might be in the future.
This discussion also covered Resolutions 2, 3 and 4 from the SGFS authorized subgroup. Issues 2, 3, 4, 8, 9 from the SGFS authorized subgroup were also discussed under this agenda item.
3.0.2 TR10000-1.3 proposal The topic included SGFS Authorized Subgroup resolution 7 (part 1). 3.2 9945-1 Profile activities Coordination of option selection. Several Action items resulted from the discussion: 9201-13, 9201-14, 9201-15. 3.3 9945-2 Profile activities No new ones beyond those mentioned for 3.2. 3.4 9945-3 Profile activities ... ? 3.5 Language Binding Profile activities
No activity to discuss at this time.3.6 Technical Report: Guide to POSIX OSE (a.k.a. IEEE P1003.0) Roy McKean asked the question: what is the difference between a base profile, platform profile, and vertical profile. The point hasn't been addressed. This question came up elsewhere on the agenda, and was discussed a bit here. It seemed closely related to the horizontal/vertical issue, and was informally subsumed under that, pending a resolution to that issue. The uses and applicability of that document in the ISO arena were discussed.
The schedule was discussed... it was expected that it would be done sometime in 1994, possibly somewhat sooner by pushing the process.
Several members raised concerns about the timelyness of the document because of the amount of time balloting will take, and the volatility of data it contains.
This was also related to the issue of what kind of TR (2 or 3) it should be, and the mechanisms for amending it.
If its a TR type 3: what process to update it? (Is amendment possible?) There don't seem to be any rules for updating type 3 TRs, but on the other hand TR 10K is a Type 3, and is being updated.
It was stated that there is a definite need to maintain the information at a relatively current level. It was anticipated that an amendment might be developed in parallel to update it.
The secretary was specifically requested to minute the above point for WG15 awareness of the issue.
See action items 9201-16, 9201-17.
3.7 Test Method issues This topic reflected Resolution 6 from the SGFS Authorized Subgroup, and SGFS Recommendation 6, Issue 7. There was a discussion of the considerations of test methods that apply to profiles. AI 9201-10 was also noted to apply to this topic, and it was at this point that RGCT was added to the discussion.
It was noted that there are two aspects to the concept of "using extensions": should conforming applications be allowed to use facilities outside of the profile (or standard) where there is overlapping functionality within the profile. Ther was concern that this might weaken the profile concept.
It was strongly noted that application conformance is not getting much attention: in fact it was unclear whether 1003.3.1 has any application conformance elements to it at all.
5 Review / Approval of Resolutions, Action Items and Issues List
4.1 1003.18 At the request of the US Member, the nature of 1003.18 was discussed. It was suggested that .18 might be simplified. Specifically, it was suggested that the material on development/execution platform ... (?) 4.2 Attributes of profiles An informal discussion of a list of possible attributes for profiles was held, and it was noted that the horizontal/vertical was related. It was observed that 1003.0 is a "structured component list" for forming profiles. It was also noted that the "characteristics annex" requested by TSG-1 would help understand what the set might be. The attributes were also noted to have a relationship to user requirements, specifically that if the attributes and user requirements did not mesh nicely that it might indicate a problem in the base standard or profile. A list of standards/profiles such as that provided by 1003.0 appeared to be a useful tool.
See also AI 9201-18.
5 Closing Process
Action Items:
5.1 Future Meeting considerations, request for invitations. There was a discussion of possible meeting dates, considering both the variability of the date of the WG15 meeting, and constraints such as EWOS week. (The IEEE meeting was known to be a constant.) It was tentatively decided to meet the Friday and Saturday of the week immediately preceding the WG15 meeting, whenever that may be. This would (currently) be either Oct 9-10 or Oct 16-17. Once the Wg15 date is determined, the meeting will be formally announced. The locations would either be: for the 9-10, Brussels (because of EWOS) or wherever WG15 is; for the 16-17, Netherlands (with IEEE).
The invitations to be requested would be a consequence of the desired meeting location and date.
5.2 Document number assignments:
N010 Framework for User Requirements (Folland) Temp No. N011 Proposed Terms of Reference N012 TR10K-1.3 SFT1 N013 SGFS Report (Walker), SGFS ad-hoc Issues list, SGFS ad-hoc resolutions SFT2, 3, 6 N014 These minutes, resolutions, attendance. N015 From PSC: summary of profile usage of base standards N016 From PSC: summary description of each POSIX SP activity (excerpt from 1003.0) SFT 4, 5, 7, 8: not retained. 5.3 Next meeting Chair A request for a meeting chair was made. Don Folland indicated a degree of willingness, but was unable to commit to it at this time. 5.4 Agenda for Next Meeting Not discussed. 5.5 Thanks to host The host was thanked when he was able to visit the meeting for a few minutes Tuesday morning. 5.6 Adjournment The meeting was adjourned at 2:06, Jan 21, 1992.
Annex A: Resolutions
9105-02 IEEE to provide a report on the options in 9945-1 Open, but in progress.
9201-01 Don Folland: Forward RGCPA N010 to IEEE PSC.
(See also 9201-04. There was some implication in the meeting that this AI and that one were related.)Also notify OIW of our work, pass the same document to them, and ask them to determine how they want to participate.
9201-02 US member Body to distribute PSC profile guidelines which result from the April PSC meeting, where at least partial approval is expected. 9201-03 Jim Isaak to distribute X/Open's response to the letter from Action item 1 from Amsterdam. 9201-04 Jim Isaak: Notify OIW of our work, pass N010 and N011 to them, and ask them to determine how they want to participate, encouraging some form of Liaison from them. 9201-05 Jim Isaak to ask IOW and AOW to send liaisons to RGCPA. 9201-06 Member body rapporteurs to invite appropriate experts from "Coordination groups" such as those listed in item 2.4 above for coordination. 9201-07 Jim Isaak to provide supporting cover letter for action 9201-06. 9201-08 US Rapporteur: RGCPA Rapporteurs to be added to invitation list for joining balloting groups for profiles. This would be a coordination review (not a counted balloting group member). 9201-09 Member bodies: provide to RGCPA any issues concerning profiles in general that they find in reviewing the IEEE ballots in addition to sending them to IEEE. (Provide other comments directly to IEEE.) 9201-10 Jim Isaak: Send RGCT the conformance clauses from 9945-1 for information (with explanatory cover). The cover will request them to package it with their work on conformance and forward as a proposed liaison statement from WG15 to SC21/WG1, SGFS, SC22. It should point out the asymmetric nature of the two parts, and the role of extensions in conformance clause. (This AI is intended to deal with the relationship with IS 9646, and point out the horizontal vs. vertical issue (asymmetry). It relates also to the differing effects of gaps between OSI and OSE and to test methods.) 9201-11 Members: Contribute on the nature of OSE profiles (horizontal vs. vertical, e.g.) if they can to help with the completion of AI 9201-10. 9201-12 Members: try to identify categorizations similar to 7.2 of TR10K1.3 suitable for AEP profiles and forward to SGFS. 9201-13 US Member Body: Tell IEEE editors that options annex should include notation about the degree of "common-ness" (frequency of use) of the option. May help avoid conflict, maximize harmonization. 9201-14 US Member Body: Request IEEE 1003.1 to fill out template for 9945-1. 9201-15 Member bodies: provide feedback on 9945-1 options annex (from AI 9201-14), both as to relevance and to selection of "core" options. 9201-16 Convener: Figure out what can be done to update a TR type 3 and recommend to US NB and WG15 as to what actions are required to provide for timely update of 1003.0 in both US and internationally. Investigate means to improve the current-ness of the document. 9201-17 US: consider how to bring about a timely amendment to 1003.0. (Maintenanace.) 9201-18 Participants: forward information on nature of profiles they are developing or of which they are aware to RGCPA, Regional Workshops, with a mind to helping them understand the issues associated with taxonomies by examples. 9201-19 Participants: encourage harmonization of a precedence list of standards (that is, standards approval bodies).
(See issue 4).9201-20 Participants: recommend whether RGCPA or WG15 should be a profile developer (or of what types). 9201-21 Participants: review WG15 N062 to determine if further changes are needed to Terms of Reference. (N062 is referenced by other WG15 Rapporteur Groups Terms of Reference.)
Resolution 1, Interaction with User Groups: Whereas JTC1, at its October 1991 meeting in Madrid accepted (amongst others) Resolution 18-B (in relation to the adoption of thee TSG-1 report), instructing SGFS to review its scope to ensure that it allows for interaction with user groups, coordination of AEP processing and development of a taxonomy for application portability, and
Whereas SGFS, during its Authorized Subgroup Meeting in Brussels, drafted an "Issue on Interaction with user groups" in reaction to the above JTC1 resolution (see SGFS N439, issue 11), asking for member bodies and other organizations to have a process to represent user requirements and to translate them into technical requirements, and
Whereas this affects all SC22/WG15 work and not just RGCPA work,
Therefore the RGCPA requests SC22/WG15 to look into this matter, and to respond to SGFS on this issue.
Resolution 2, Progression of 1003.0: The WG15 RGCPA requests ISO/IEC JTC1/SC22/WG15 to approve that IEEE 1003.0 D15 be distributed to SC22 for review and comment as soon as this document becomes available, under the assumption that the corresponding NWI is approved by JTC1.
Resolution 3, Terms of Reference of RGCPA: The RGCPA requests SC22/WG15 to approve their Proposed Terms of Reference and Program of Work, as laid down in document RGCPA N011.
Resolution 4, Thanks to Host: The WG15 RGCPA thanks Uniforum and Mr. Ralph Barker for hosting the RGCPA meeting.
Resolution 5, Thanks to Chair: The WG15 RGCPA thanks Mr. Jim Isaak for chairing the meeting.
Resolution 6, Thanks to Secretary: The WG15 RGCPA thanks Mr. Donn Terry for acting as secretary for the meeting.
Resolution 7, Thanks to Drafting Committee: The WG15 RGCPA thanks Mr. Roy McKean and Mr. Willem Wakker for participating on the Drafting Committee.