SC22/WG21/N1213

J16/99-0037

November 9, 1999

Sean A Corfield

sean.corfield@ocsltd.com

J16 Meeting No. 29

WG21 Meeting No. 24

October 19-26, 1999

Aston Keauhou Beach Resort Hotel

Keauhou-Kona, HI

Plum called the meeting to order at 18:04 HST, 19 October, 1999.

Plum welcomed everyone to Hawaii and explained the facilities.

In attendance were:

Plum was acclaimed.

Corfield acted as secretary.

Plum nominated English. Approved by acclamation.

Motion by Clamage / O'Riordan:

Move we adopt N1200 = 99-0024 as the agenda.

No drafting committee was deemed necessary as WGs were drafting their own wording for Defect Resolutions.

Corfield noted that Andy Sawyer was omitted from the attendance list but that otherwise there were no corrections to the previous minutes. Plum asked that we accept N1194 = 99-0018 as a true record of the minutes of the previous meeting with the correction that Sawyer be added to the attendance list.

There were no action items outstanding.

There were no documents to recognize.

Mostly deferred to the joint meeting.

Plum reported on the SC22 plenary in Berlin which took place since our last meeting.

Plum reported that working groups were suffering declining participation and it seems difficult to get support to continue work. There will be further debate on how standards bodies obtain funding from the sale of published standards. There was discussion about the recent change of rules (that now allows 'O' status members to participate) and how this will affect SC22-level participation.

Deferred to J16. Plum feels the procedures are in place and working. Koenig raised two issues:

• so far we have said how we want to deal with issues but not when,
• we have not decided how we should actually publish the TRs etc.

Kühl felt we had agreed that we would approve issues at this meeting. Clamage agreed but we still haven't said what we will actually do with them. Clamage said we can have more than two TCs. Plum agreed - JTC1 have different procedures to ISO (who only allow two TCs). Kühl feels we need to produce a TC soon because it then puts an official stamp on decisions we have made and helps implementations conform. Plum said that government procurement is no longer such a driving force behind standards compliance - now compliance can be used as a 'measuring stick', e.g., "93% compliant".

O'Riordan expressed concern about resolutions that invalidate existing implementations (for DRs covering ambiguous issues) and more importantly the issues for which no implementation can conform (because the DR covers an unimplementable issue). Plum agreed it was very bad politics to change the standard in ways that break existing implementations. Glassborow felt that ambiguity was actually more important - for users. Sutter observed that we seem to have consensus on the urgency of publishing something fairly soon. Stroustrup felt one of the most damaging perceptions is that we change our minds - once we make a decision, we should be sure that we have final agreement. There was discussion about classification of issues. Koenig feels we have three options:

• publish one TC mid-course (early 2001),
• publish one in 2000 (jokingly called "C++ 2000"),
• decide at the end of some future meeting that we have enough important issues to publish something.

Plum says we've been very public about our process so perhaps we should stick to our originally agreed statement, that we will publish a mid-course correction.

Glassborow asked Plum to clarify the voting process on the TC. Glassborow feels the end of 2000 would be a good milestone to aim for. Sutter agrees with Stroustrup that we should publish when we are sure of the decisions - we should adopt a schedule and stick to it. Stroustrup said that publishing either more often or less often than around 18 months would be difficult, either for us as a committee or for users (or both). O'Riordan expressed concern that we are not looking forward at possible resolution of issues that are as yet unexamined. Koenig said it is important that we attempt to resolve related clusters of issues in a consistent manner - according to an implied strategy. Some consensus that we should aim for a TC completely agreed by October 2000. That would mean that the majority of the TC was agreed by April 2000 so that the October meeting would be taken up with proof-reading the TC (as we did for the two CDs). Sutter felt that this timescale would mean we would need to draft a TC, starting at this meeting. Plum said there were no particular issues of formatting. Koenig said he has not yet begun work on folding agreed issues into the standard because we have not yet agreed any issues formally. Sutter feels we should agree something and let J16 know what our consensus was. Stroustrup feels that the schedule is being unduly accelerated. Kühl thinks we will continue to get new issues so we will be working on those in Toronto anyway.

Plum summarised: we propose the goal that we shall vote in Toronto on adopting TC1. There was no opposition to this as the sense of the meeting. Sutter feels we should also decide on the format of the TC. Koenig agreed that it is reasonable for the committee (as a whole) to decide whether the TC should be just a list of corrections or whether it should be a revision of the standard or both. Plum feels we are obliged to itemise all the changes regardless of whether we publish a revised standard. Koenig recalled that revised standards are required to have a change history so he may be able to incorporate the list of issue resolutions as the change history in some format.

O'Riordan felt we should begin to think about the next standard as well. In particular, we should begin to think about when we should make a decision on when to aim for a new project for this.

Denmark has an official invite out for Copenhagen in April 2001 but this cannot be co-located with WG14 (due to space requirements). It was agreed that it would be nice to have a West Coast, USA meeting in October 2001.

See also 5.3 below.

There were no resolutions to approve nor any outstanding issues to review.

Plum agreed to chair the Tokyo meeting. Plum noted he was the only contender for the convenor post.

Plum asked for any additional issues, otherwise it will be the same as this meeting.

The next meeting will be in Tokyo, Japan, April 2000. The exact dates were no available at this meeting.

Deferred to joint session.

None to assign.

Plum to research and clarify the exact ISO voting process for a TC.

Corfield notified his intention to stand down as secretary after this meeting due to his impending relocation to the USA.

Plum moved that the committee thank Corfield for his hard work as secretary over the last four years. Accepted by acclamation.

WG21 thanked Plum for organizing this meeting.

Plum moved to adjourn at 19:40HST.