Doc No:   SC22/WG21/N2110
                                                  J16/06-0180
                                        Date:     2006-11-04
                                        Project:  JTC1.22.32
                                        Reply to: Robert Klarer
                                                  IBM Canada, Ltd.
                                                  klarer@ca.ibm.com

Minutes of J16 Meeting No. 43/WG21 Meeting No. 38, October 16-20, 2006

1. Opening activities

Clamage called the meeting to order at 09:03 (GMT -08:00) on Monday, October 16, 2006

1.1 Opening comments

Clark Nelson, speaking on behalf of Intel, welcomed everyone to Portland, and described facilities for the meeting.

1.2 Introductions

Clamage had the attendees introduce themselves.

1.3 Membership, voting rights, and procedures for the meeting

Nelson reviewed membership rules. Nelson then circulated the attendance list and membership list. Clamage voting rules and reviewed procedures for the meeting.

1.4 Agenda review and approval

Clamage presented the agenda (document J16/06-0133 = WG21/N2063).

Motion to approve the agenda:

1.5 Distribution of position papers, WG progress reports, WG work plans for the week, and other documents that were not distributed before the meeting.

Each of the Working Group chairs presented their plans for the coming week.

Core Working Group (CWG)

Adamczyk indicated that CWG would continue to process issues, and would place particular emphasis on proposals to add rvalue references, generalized constant expressions, template aliases, and a reformulation of the definition of POD to the language.

Library Working Group (LWG)

Hinnant reported that LWG has 44 new papers that have been submitted since Berlin. No new proposals to TR2 will be accepted after this meeting.

Evolution Working Group (EWG)

Stroustrup reminded the committee that he had posted an agenda for this week's EWG technical sessions to the mailing list, noting that the overall approach will be to first handle a the small number of items that are close to completion, and then to handle the "big lumps." The discussion of concepts will probably take place on Tuesday in the morning or the afternoon.

Threading

Boehm indicated that a subgroup of EWG will be holding sessions on threading, and will possibly participate in a joint session with LWG.

1.6 Approval of the minutes of the previous meeting

Motion to approve the minutes (document J16/05-0063 = WG21/N1993)

1.7 Report on the WG21 Sunday meeting

Sutter reports:

• 5 countries are represented at this meeting; all 5 have voting status
• there is a move on in SC22 to publish all type 2 TRs for free on the ISO website
• the drafting committee will be composed of Adamczyk, Clamage, Dos Reis, and Hinnant

Sutter reviewed two possible timelines for the completion of a standard:

Stage	Description	Timeline A accreditation in 2009	Timeline B accreditation in 2010
SC22 Registration Ballot	Ideally all major features present. Usually few comments.	Oct 2006	Oct 2006
SC22 FCD Ballot	All major features in near-final form. Need time for disposition of comments.	Oct 2007	Apr 2008
JTC1 FDIS Ballot	Final text	Oct 2008	Oct 2009

Sutter noted that the last two stages cannot be combined; they must be separate steps. Each of these stages will take 12 months or longer (including disposition of comments after FCD ballot).

Stroustrup spoke in favour of aiming to have an acredited standard before 2010, (i.e. Timeline A, above).

Plum suggested that the WG should decide in this meeting what the scope of the standard should be. What items aren't going to make it?

Vollmann asked whether it is formally necessary to do the registration this time.

Sutter responded by noting that there are two things to consider. First of all the content: for registration, we need to work at a coarser level of granularity than that at which the committee usually deals. Secondly, it's possible to combine registration and FCD, but if there are too many comments, then the WG has made no progress, and doesn't even have a registered ballot.

Benito confirmed this.

Dawes noted that any timeline may be delayed unexpectedly due to unexpected procedural obstacles that are not directly related to the work of any J16 or WG21 member.

Plum observed that both timelines are the same until October 2007, and then asked: is that the correct time to decide which timeline we'll follow?

Stroustrup indicated that, as a working group chair, he needs a deadline to get technical discussions to converge. A shorter timeline might produce a better language.

Sutter reminded the group that when we vote in text, that doesn't mean that we are done; subsequent implementation experience will identify problems that will need to be addressed by the committee.

Meredith asked: what do we need to have for each stage in the timeline?

Sutter replied that the committee can't add a major new feature after FCD ballot.

Benito added that the real issue is that, if country X says "we don't like this because Feature Y needs to be changed," then the committee has an opportunity to change X's NO vote to a YES by changing Y.

Nelson noted there was talk of the risk of combining Registration with FCD, and the asked: what is the risk of calling it a FCD and not a CD?

Sutter explained that FCD just means that the committee is not obliged to issue another one if it succeeds.

Dos Reis reported that the French NB is concerned about the complexity of the language, and is worried that there are too many new features that are expert friendly. One example is variadic templates, which appears to be a feature to make it easier to implement TR1, and is therefore only useful to library implementers. France wants to see support for network libraries, and intervals in C++0x, not TR2.

Meredith asked: if the French NB is concerned about having intervals and sockets in C++0x, how long will they be willing to stretch the timeline to get them?

Dos Reis responded by explaining that the French NB are more concerned about the quality of the standard than the timeline.

Sutter proposes a Straw Poll: (worded by Stroustrup)
"Do you think it (i.e. Timeline A) can be done?"

Nelson asked: who thinks it's important enough to publish in this decade that they will be willing to make a sacrifice to achieve that? (where sacrifice means more meetings, fewer features, or whatever).

Myers noted that it sounds like we are voting (in the straw poll, above) on whether to ruthlessly cut in order to make the deadline.

Stroustrup agreed.

Becker argued that politically, in order to avoid surprises, there should be as much in the registration document as possible, and the WD isn't there yet.

Plum suggested that whatever is in the working paper now is quite fine for a registration ballot, while noting that the committee will never get by with voting out a document for registration if the paper lacks a feature that the whole community agrees is required.

Hedquist indicated that "you've got to have some words (in the registration document) to the effect of 'this feature is going to be in the FCD.'"

Plauger agreed with Hedquist, noting that if the committee adds a big chunk after registration, a NB can shoot down a document on procedural grounds. The committee can get away with being sketchy, but it can't get away with large omissions. The registration ballot document doesn't have to specify concepts in precise standardese, for example, but it must contain a description of scope.

Spicer asked whether the committee can issue the concepts paper as an addendum to the registration ballot document?

Maurer observed that, for proposals moving from EWG to CWG, it is useful to have at least one member of CWG review the proposal before giving it to core as a whole.

1.8 Liaison reports

WG14 Liaison

Benito reported that WG14 currently has three projects underway:

• TR 24731 "Extensions to the C Library, Part 1: Bounds-checking interfaces"
  This item has passed its final ballot, and there is a proposal for a part two.
• TR 24732 "Extension for the programming language C to support decimal floating-point arithmetic"
  This item has passed registration ballot with no comments.
• TR 24747 "Special math functions"
  In registration ballot right now.

The next meeting of WG14 will be the week of October 22, in Portland (the location of the present WG21 meeting). The following meeting will be located in London (not Oxford, as is the case with the Spring 2007 meeting of WG21). The meeting after that will be in Kona. There are no plans as yet for the Spring 2008 meeting. For the Fall 2008 meeting, WG21 has received an invitation from Cisco in Denver.

OWG Vulnerabilities Liaison

Benito explained that this OWG gets renewed each year by SC22. Vulnerabilities is a working group that was started in Mt. Tremblant in October 2005 and renewed in London in 2006. They will be meeting the week after the WG14 meeting in London. They have liaison to every WG under SC22 except WG21. Benito requested that an individual volunteer to act as liaison.

Sutter reminded the group that there had previously been concern about this OWG because of the potential for its work to become a kind of "Use Ada instead of C or C++" report.

Benito explained that he got involved because of that concern, and that he doesn't think there's anything to fear. The attitude of this committee is positive.

TG5 Liaison

There will be a ballot resolution in April.

1.9 Editor's report and WP approval

The current draft of the WP is J16/06-0079 = WG21/N2009. The editor's report corresponding to this draft is J16/06-0078 = WG21/N2008

Becker reports:

Section numbers change as things move around, so please refer to section labels in proposals.

Mike Miller has proposed an index that correlates section labels to section numbers. This will appear in the next WD.

All of the labels in the Library clauses start with lib.*, and this makes it difficult to scan an alphabetized list of labels. Becker has removed the lib.* prefix from each of these. This produced only one conflict, and that has been addressed.

Plum invited Crowl to volunteer for the job of backup editor.

Crowl agreed to do this.

Motion to accept the working paper (J16/06-0079 = WG21/N2009 ):

1.10 New business requiring actions by the committee

2. Organize subgroups, establish working procedures.

We have three subgroups: Core, Library, and Evolution. There will be a subgroup of Evolution to deal with issues relating to concurrency.

The committee broke into subgroups at 11:30 (GMT -08:00).

3. WG sessions (Core, Library, Performance, Evolution).

4. WG sessions continue.

5. WG sessions continue.

6. WG sessions continue.

7. General session.

7.1 WG status and progress reports.

Registration Ballot Paper Motions

Motion 1.

Move to direct the project editor to create a registration ballot paper by taking a copy of the current working draft (N2009), applying the motions adopted at this meeting, and inserting the following additional sections with actual references to the indicated WG21 papers:

• <#.#> Concepts
   This section is a placeholder. The next C++ standard is intended to include support for concepts. This feature is intended to provide language support for describing features of types, for example to express the container requirements tables in the C++ Standard Library as code that can be checked by the compiter. For more information and snapshots of current draft proposals still under discussion and development, see: .
• Between <1.9> and <1.10> Memory Model
   This section is a placeholder. The next C++ standard is intended to include support for a memory model. This feature is intended to provide foundational support for concurrency by defining rules for the way programs' memory reads and writes may be transformed, optimized, and executed. For more information and snapshots of current draft proposals still under discussion and development, see: N2052 and N2075.
• Between <3.7.1> and <3.7.2> Thread Local Storage
   This section is a placeholder. The next C++ standard is intended to include support for thread local storage. This allows declaration of storage unique to a particular thread and with lifetime of that thread. This will be modeled on similar facilities found in many current implementations, using for example an __thread storage class specifier. A more detailed proposal for such a facility is given in N1966.
• Atomic Operations API
   This section is a placeholder. The next C++ standard is intended to include support for atomic operations, which may be used to concurrently access data from multiple threads without introducing undefined behavior. The rationale for much of our current approach and an early proposal along these lines are described in N2047.
• Threading API
   This section is a placeholder. The next C++ standard is intended to include support for a threading API. This feature is intended to provide support for synchronization facilities and thread lauching and joining. For more information and snapshots of current draft proposals still under discussion and development, see: N1907, N2090.
• <3.5> Minimal Support for Dynamic Libraries
   Modify section 3.5 "Program and Linkage" to provide minimal support for dynamic libraries along the line of N2117.
• <21.5> Unicode Character Support
   Add support for 16-bit and 32-bit characters in support of Unicode, along the lines of N2018 and N2035.
• <#.#> Optional Garbage Collection
   This section is a placeholder. The next C++ standard is intended to include support for optional garbage collection. This feature is intended to provide automatic memory management, whereby that explicit delete/free is not required for all objects (used if and only if a programmer requires it). For snapshots of current draft proposals still under discussion and development, see: N1943.
• <#.#> Modules
   This section is a placeholder. The next C++ standard is intended to include support for a module system. This feature is intended to provide an alternative to the #include source organization model to enable better code organization and avoid repeated compilation of headers. For more information and snapshots of current draft proposals still under discussion and development, see: .
• <#.#> Dynamic Libraries
   This section is a placeholder. The next C++ standard is intended to include support for dynamic libraries. This feature is intended to standardize support for organizing object files in libraries that are not statically linked. For more information and snapshots of current draft proposals still under discussion and development, see: .

Sutter asked whether the registration document needs to specifically mention Unicode support? Is it big enough?

Nelson replied that we can add it later, but people might like to know that we intend to add support for Unicode.

Stroustrup suggested that we might want to add Unicode so that people know that we didn't forget it. However, if we start naming small items, we might have a very long list.

Benito indicated that, if the committee issues for registration a document that does not mention Unicode support, and a country sends a No vote because of a perceived lack of support for Unicode, then the committee has an opportunity to turn that No vote into a Yes vote simply by telling that country that we intend to support Unicode. On the other hand, if you have a list that is excessively long, then you are sending a message that we are very far from completion.

Plauger suggested that the list is safer if it is shorter.

Discussion ensued concerning which items should appear on the list.

Maurer inquired about the consequences of putting something on the list and then failing to fulfill that promise.

Sutter replied that we should seek to avoid that. As well, it is bad to fail to list a major item that we intend to include.

Discussion continued. Sutter then proceded to invite discussion on each bullet item in the motion. Where unanimous consent to the inclusion of a particular bullet item did not exist, a straw poll was held (details below).

Bullet item: Concepts.

no objection to unanimous consent.

Bullet item: Optional Garbage Collection.

Stroustrup spoke in defense of Optional Garbage Collection, noting that there has been 10 years of existing practice in the field.

Nelson asked for an explanation of the meaning of the term "Optional" in this context.

Stroustrup explained that "Optional" means that there is a directive that tells the system that you want Garbage Collection (GC) in a particular region of the code.

Hinnant asked Stroustrup whether he anticipated that this item would have any impact on the C++ Standard Library.

Stroustrup responded by indicating that he expected no impact to the standard library, since library entities should not depend on the existence of GC. They should continue to work in the absence of GC.

Dawes suggested that, instead of "Optional Garbage Collection," a clearer title might be "Programmer-Controlled Garbage Collection." The text of the motion was changed to reflect this suggestion.

Straw poll: Include the bullet item "Programmer-Controlled Garbage Collection" in the motion?

Bullet item: Modules

Caves expressed his belief that it had not yet been demonstrated that this feature solves the problems for which it was intended.

Maurer asked whether the proposers have a specification of this feature that is suitable for integration into the Working Draft.

Vandevoorde explained that there is currently no written working paper text.

Discussion ensued.

Hinnant reported that the LWG had expressed concern about the impact of this feature upon the standard library, and the LWG's schedule of work.

Meredith asked whether the committee could pursue the modules proposal as a possible topic for a TR?

Straw poll: Include the bullet item "Modules" in the motion?

The bullet item pertaining to "Modules" in the text of the motion (above) was removed.

Bullet item: Dynamic Libraries

Straw poll: Include the bullet item "Dynamic Libraries" in the motion?

As convenor and meeting chair, Sutter decided that this result did not constitute consensus, and the bullet item was removed from the text of the motion.

Bullet item: Memory Model

no objection to unanimous consent.

Bullet item: Thread Local Storage

no objection to unanimous consent.

Bullet item: Atomic Operations

no objection to unanimous consent.

Bullet item: Threading API

Dawes noted that this item is not meant to represent the whole threading package, including futures, read/write locks, and thread pools. Those items will likely be directed by LWG to TR2.

Stroustrup opined that, by failing to provide higher-level features like futures, the standard leaves programmers with no choice but to struggle with locks.

Vollmann expressed the opinion that low-level threading facilities are hard to use correctly, and that higher-level facilities are needed.

Discussion ensued.

Straw poll: Include the bullet item "Dynamic Libraries" in the motion?

As chair, Sutter decided that this vote constitued consensus.

Bullet item: Unicode Support

Marcus asked whether it was possible that someone will see the promise of Unicode support, and assuming full, deep support for Unicode (in string library, etc), be disappointed.

Plauger indicated that the LWG is committed to full support of Unicode, but does not intend to duplicate the library with Unicode character variants of existing library facilities.

Meredith then asked whether -- in that case -- we are setting expectations too high by mentioning Unicode support.

Straw poll: Drop the bullet item "Unicode Support" from the motion?

Additional items: Lambda Functions and Contract Programming

In addition to the bullet items identified above, two other features were discussed for possible inclusion (as placeholders) in the registration document, Lambda Functions and Contract Programming.

Sutter inquired of the Working Group chairs whether they objected to the exclusion of any mention of Contract Programming from the text of the motion. Hearing no objection, Contract Programming was excluded.

Stroustrup suggested that Lambda Functions are too small a feature to be mentioned in the text of the motion.

Jaarvi noted that there are currently two slightly different proposals before the committee to introduce Lambda Functions to C++.

Nelsom asked whether either of the proposals been implemented.

Jaarvi replied that there is a partial implementation of at least one of the proposals.

Straw poll: Exclude Lambda Functions from the motion?

Straw poll: Support for the motion, as amended above, overall?

Motion 2.

Move to direct the convener to submit the foregoing registration paper for ISO registration ballot.

Motion 3.

Apply to ISO for registration of a Technical Report on modules for C++. N2073 is the current draft proposal for modules.

As chair, Sutter decided that this constituted consensus.

Core Working Group

Adamczyk presented Core Working Group status and reviewed formal motions to be made Friday (for formal motions, see 9.1, below).

Adamczyk then asked a procedural question. Can the committee vote on a document that was not quite yet in its final form prior to the beginning of the Thursday plenary meeting?

Plauger observed that we must agree on a cutoff time.

Sutter recalled that, in Berlin, the rule was that the wording had to be available by 6:00pm Thursday.

Discussion ensued.

Sutter suggested that the cutoff should be be end of session Thursday.

Adamczyk presented Core Working Group status and reviewed formal motions to be made Friday (for formal motions, see 9.1, below).

Motion 1. Move all issues in Ready status in J16/06-0125=WG21/N2055 to DR status and into the Working Draft. Issue numbers are 58, 301, 305, 367, 370, 372, 414, 415, 454, 468, 477, 480, 484, 494, 500, 505, 506, 514, 515, 516, 517, 521, 522, 524, 526, 534, and 557.

Straw Poll: support for Motion 1

Motion 2. Adopt J16/06-0058=WG21/N1988, "Adding extended integer types to C++ (Revision 1)".

Straw Poll: support for Motion 2

Motion 3. Adopt J16/06-0188=WG21/N2118, "A Proposal to Add an Rvalue Reference to the C++ Language (Revision 3)".

Straw Poll: support for Motion 3

Library Working Group

Hinnant presented Library Working Group status and reviewed formal motions to be made Friday (for formal motions, see 9.1, below).

Motion 1. Move Ready issues into C++0X Working Paper: 520, 521, 530, 531, 535, 537, 538, 540, 541

Straw Poll: support for Motion 1

Motion 2. Adopt N2114: "long long Goes to the Library, Revision 1" into the C++0x Working Paper.

Straw Poll: support for Motion 2

Motion 3. Adopt N2111: "Random Number Generation in C++0X: A Comprehensive Proposal, version 4" into the Working Paper.

Straw Poll: support for Motion 3

Motion 4. Adopt N2072: "Iostream manipulators for convenient extraction and insertion of monetary values" into the C++0X Working Paper.

Straw Poll: support for Motion 3

Motion 5. Adopt N2058: "Proposed Text for Proposal to add Date-Time to the Standard Library 1.0" into the second Library Technical Report (TR2).

Myers requested that this motion be withdrawn so that it can be advanced at the next meeting. N2058 doesn't have full wording.

Hinnant agreed to withdraw the motion.

Motion 6. Adopt N2071: "Iosteam manipulators for convenient extraction and insertion of struct tm objects" int the second Library Technical Report (TR2).

Straw Poll: support for Motion 6

Evolution Working Group

Evolution Working Group (EWG) will be making no formal motions Friday.

Stroustrup reported on EWG activities and progress, noting that the list of proposals that were discussed is on the Wiki.

Concurrency Group

The Concurrency Group will be making no formal motions Friday.

Boehm reported on Concurrency group activities and progress.

Future meetings:

See 10.1, below.

7.2 Presentation and discussion of DRs ready to be voted on. Straw votes taken.

8. WG sessions continue

9. Review of the meeting

9.1 Formal motions, including DRs to be resolved.

Core Working Group motions

Motion 1.

Move all issues in Ready status in J16/06-0125=WG21/N2055 to DR status and into the Working Draft. Issue numbers are 58, 301, 305, 367, 370, 372, 414, 415, 454, 468, 477, 480, 484, 494, 500, 505, 506, 514, 515, 516, 517, 521, 522, 524, 526, 534, and 557.

Motion 2.

Adopt J16/06-0058=WG21/N1998, "Adding extended integer types to C++ (Revision 1)."

Motion 3.

Adopt J16/06-0188=WG21/N2118, "A Proposal to Add an Rvalue Reference to the C++ Language (Revision 3)."

Library Working Group motions

Motion 1.

Move Ready issues into C++0x Working Paper:
520, 521, 530, 531, 535, 537, 538, 540, 541

Motion 2.

Adopt N2114: "long long Goes to the Library, Revision 1" into the C++0X Working Paper.

Motion 3.

Adopt N2111: "Random Number Generation in C++0X: A Comprehensive Proposal, version 4" into the C++0X Working Paper."

Motion 4.

Adopt N2072: "Iostream manipulators for convenient extraction and insertion of monetary values" into the C++0X Working Paper.

Motion 5.

Adopt N2071: "Iostream manipulators for convenient extraction and insertion of struct tm objects" into the second Library Technical Report (TR2).

Registration Ballot Paper Motions

Motion 1.

Apply to ISO for registration of C++0x. The registration ballot paper shall consist of a copy of the current working draft (N2009), additionally applying the working paper changes approved at this meeting, and additionally inserting the following additional descriptions of additional major features and the project editor shall supply actual references to the indicated WG21 papers:

14.9 - Concepts
This section is a placeholder. The next C++ standard is intended to include support for concepts. This feature is intended to provide language support for describing features of types, for example to express the container requirements tables in the C++ Standard Library as code that can be checked by the compiler. For more information and snapshots of current draft proposals still under discussion and development, see: N2081, N2082, N2083, N2084, N2085, N2098.

3.7.3 - Programmer-Controlled Garbage Collection
This section is a placeholder. The next C++ standard is intended to include support for programmer-controlled garbage collection. This feature is intended to provide automatic memory management, whereby that explicit delete/free is not required for all objects (used if and only if a programmer requires it). For snapshots of current draft proposals still under discussion and development, see: N1943.

Between 1.9 and 1.10 - Memory Model
This section is a placeholder. The next C++ standard is intended to include support for a memory model. This feature is intended to provide foundational support for concurrency by defining rules for the way programs' memory reads and writes may be transformed, optimized, and executed. For more information and snapshots of current draft proposals still under discussion and development, see: N2052 and N2075.

Between 3.7.1 and 3.7.2 - Thread Local Storage
This section is a placeholder. The next C++ standard is intended to include support for thread local storage. This allows declaration of storage unique to a particular thread and with lifetime of that thread. For more information and snapshots of current draft proposals still under discussion and development, see: N1966.

29 - Atomic Operations API
This section is a placeholder. The next C++ standard is intended to include support for atomic types and operations, which may be used to concurrently access data from multiple threads without introducing undefined behavior. For more information and snapshots of current draft proposals still under discussion and development, see: N2047.

30 - Threading API
This section is a placeholder. The next C++ standard is intended to include support for a threading API. This feature is intended to provide support for synchronization facilities and thread launching and joining. For more information and snapshots of current draft proposals still under discussion and development, see: N1907, N2090.

Plum noted that no reference to a paper describing concepts had been included in the motion.

Dos Reis requested that all of the current papers concerning concepts be identified in the motion. The motion was amended accordingly.

Abrahams objected to the amendment, explaining that he hadn't had time to consider the implications of this late change.

Sutter explained that the change was necessitated by an inadvertant omission in the wording of the motion, not a change to the intention of the motion.

Motion 2.

Apply to ISO for a new work item for a Technical Report of type 2 on modules for C++.

Evolution Working Group motions

None.

9.2 Review of action items, decisions made, and documents approved by the committee

There are no outstanding action items.

9.3 Issues delayed until Friday

None.

10. Plans for the future

Goldthwaite invited the committee to meet in Oxford in April 2007. The hotel is the Paramount Oxford hotel. It's not immediately downtown, but it's close. A single room including bed, breakfast, and buffet lunch is 115 pounds. There will be a Google-sponsored reception.

Abrahams announced that the inaugural Boost conference will be held May 14-18, 2007, in Aspen, Colorado.

Plum noted that, for the Fall 2007 meeting in Kona, the meeting rooms have been reserved to accomodate a 6-day meeting, if the committee wishes to meet for six days.

Sutter suggested that J16 and WG21 should have a full meeting in July 2007.

Meredith noted that the WG is planning to issue an FCD after the Kona meeting and suggested that, if ever there was a meeting that should be extended, it is the Kona meeting.

Stroustrup supported Meredith's comment.

Abrahams spoke to the need to work better between meetings using collaboration technology such as internet chat.

Plum volunteered to look into solutions along this line with Abrahams, Hinnant, and Glassborow.

Plauger indicated the LWG is planning to meet informally in Illinois for three days in January.

Hinnant expanded on this, reporting that LWG is meeting Jan 22 through 24, 2007 at FermiLab in Batavia, Illinois. The meeting facility can handle a maximum of 20 people.

Sutter asked whether anyone among the group wanted to speak in favour of a full meeting (not an ad hoc meeting) in July?

Crowl noted that many members do not control their own travel budgets, and are unable to immediately commit to a meeting.

Informal show of hands:
How many people think that it would be a good idea to meet in July?

• 19 hands were raised.

Informal show of hands:
How many people think that they might not be able to attend a meeting in July?

• 17 hands were raised.

Informal show of hands:
How many people think that it would be a good idea to meet three times in 2008?

• 14 hands were raised.

Informal show of hands:
How many people think that it would be a bad idea to meet three times in 2008?

• 7 hands were raised.

Plauger reminded Sutter that it is his responsibility to convene meetings. He can act autonomously in this capacity.

Sutter indicated that he will consult with the WG chairs. Members that don't know whether they'd be able to attend a July 2007 meeting were requested to respond to Sutter in one month's time to report whether they have learned that they are able to attend after all.

Goldthwaite indicated that meeting space will be available for small ad hoc meetings during the ACCU meeting that precedes the WG21 meeting (Spring 2007, in Oxford).

Hinnant reported that LWG is planning to meet in July, depending on how the meeting at FermiLab goes.

Meredith asked: do we have any guidance for Tom regarding a 6-day meeting in Kona?

Informal show of hands:
Who's in favour of a 6-day meeting in Kona?

• Many hands were raised.

Informal show of hands:
Who's opposes meeting for a 6th day in Kona?

• A few hands were raised.

Sutter asked: if there were a third meeting in 2008, would anyone object to it being held in the west coast area of the USA?

No objection was expressed.

10.1 Next meeting

10.2 Mailings

Nelson reported the following mailing deadlines:

post-meeting mailing	November 3 2006
midterm mailing	January 12 2006
pre-meeting mailing	March 9 2006

10.3 Following meetings

Motion to adjourn

Applause.

Meeting adjourned at 03:20 (GMT -08:00)

Attendance