IEEE 754R notes from February 15, 2006

The February meeting of the 754 group was hosted by David Hough in the Pulgus Water Temple Room in building 16 at Sun in Menlo Park on Wednesday 2/15 from 1:00 to 5:00.

The list of those who did or did not attend the February meeting & are or are not eligible to vote on February issues may be found in http://nonabelian.com/754/roster0602.

IEEE 754 Revisions Committee

Wednesday February 15, 2006 from 1:00 PM to 5:00 PM
Pulgas Water Temple Room, Building 16, Sun, Menlo Park
Host: David Hough

Proposed Agenda

  1. Roll Call & Introductions: 27 + 3 have quorum
  2. Show Patent Slides: done.
  3. Call Meeting to Order: at 1:12
  4. Approval of Agenda: (http://nonabelian.com/754/agenda) approved
  5. Approval of Minutes: (http://nonabelian.com/754/minutes/754/060118) approved
  6. Principles and Procedures Vote Announced: (24-0-1 of 31). Passed.
  7. Ned Present something on intervals.
    Kahan commented on infinities.
    Jeff on whether or not to support intervals.
    Jim on forward compatibility.
    Ned's reply.
  8. Presentation of Phoenix Draft Principles & Procedures
    (http://nonabelian.com/754/draft/slides.pdf)
    (http://nonabelian.com/754/draft/754p.pdf)

    Dave observed that Dan has changed both the content and the form & feels that the change is too large to complete in the time available. Therefore, he suggests that we consider going back to the 1985 text.

    Dick asked about coordination about the language committees.

    John Harrison observed that these are the issues not appropriate for discussion.

    Prof Kahan thinks we are being forced to decide in haste & repent at leasure.

    Pamela asked if its easy to decide what's new & what's old.

    Jim thinks we have to decide on what the new base document is.

    Mike generally agrees with David's position & offers a third option of continuing with the current draft.

    Dan responded to those issues.

    Dave said there are a number of issues that remain to be decided. The real issue is whether we have to be finished by June. Why does this draft go to the MSC if we're deadlocked.

    Prof Kahan agrees that the deadlines being imposed are unwise.

    Jim is in sympathy with Prof Kahan's remarks. He asked if its feasable to take David's draft & pair it down to that.

    John Harrison pointed out that I am not proposing this document.

    Dan replied about the draft going to the MSC.

    Dave doesn't believe we have time to modify the September draft.

    Pamela supports Prof Kahan's recommendation that it be a well thought out document & is this group comfortable with abandoning the current draft in favor of this document.

    Jeff was skeptical on first reading of the draft but after reading it through thought there were some valuable approaches in it. Characterizing the document as a redistillation rather than abandoning the old.

    Jim asked Mike to comment on the feasibility of adding decimal as an annex to 754.

    Mike replied yes. (?)

    Dan asked for clarification of Jim & Mike's comments.

    Jim was talking about the 1985 document.

    Mike believes its feasible but not as clean as with the October draft.

    Dave clarified that adding small annexes is not appropriate for anything so involved as decimal.

    Jim wanted Prof Kahan to speak to bringing the standard to conclusion quickly.

    Prof Kahan observed that Jim feels there is a core about which we agree and with which we could continue to change to finish the document. Prof Kahan elaborated & basically agreed.

    John Harrison has only looked at the decimal part in detail & believes that it comes close to what was agreed.

    Dan replied that that was what I was trying to do.

    Jim asked what is the next step?

    Fred Zemke is not clear what is our base document & thinks we should adopt one & recommends that.

    Jeff agrees with the tactics but believes that strategy should be decided first.

    Prof Kahan: strategy. The worst defect with the 1985 standard has been that is was regarded as a hardware standard. He believes that an important part of our strategy is that we have to describe the programming ambiance & we need greater participation from the languages.

    Eric S thinks we are divide into two groups: those that believe we have done all we can & those that believe we should do more. Suggests we base with 1985.

    Peter Tang asked Eric S if FMA & quad should be included to 1985.

    Dan said that the languages under consideration were C, Fortran, Java, & Cobol.

    Nobu elaborated on this point. And pointed out that C++ doesn't meet before June.

    Dan clarified the decimal format issue raised by Nobu.

    Fred Zemke also wants to see only one interchange format for decimal.

    John Harrison said that this should be easy in text for decimal.

    Prof Kahan has his doubts that this can be made to work. He pointed out that the value of the 1985 standard was to the arithmetic rather than the encoding & that the common attribution of this to the encoding is misplaced.

    Dick mentioned that big endian/little endian is not a counter example to standard formats.

    Mike commented that getting into the decimal is not what we want at the moment.

    Dan asked what is the next step?

    Fred Zemke suggests that we adopt the former base document & ask who wants to change what.

    Jim repeated that we should decide which document we want to continue with & points out that 754/1985 is the base document.

    Nobu made a suggestion about mutiple decimal formats & feels we need more data.

    John Crawford commented that we need to move away from the deep trouble we're in. Three options (1) 1985, (2) 2005, (3) 2006.

    Pamela pointed out that Eric S suggested starting with 1985.

    John Harrison asked for clarification.

    Dave mentioned we can have both normative & informative annexes.

    Dan clarified.

    Peter Tang said we should read the Phoenix draft.

    Roger said we should at least adopt the principles of the new draft.

    Jim are we at the point that we can go off, study 754p, talk about it, and vote on it at the next meeting?

    Dan pointed out that we have

    Dave

    Nobu said it is not easy to follow the discussion over the phone.

    Prof Kahan moved that we cancel tomorrow's meeting & Jim seconded. motion is passed.

    Pamela pointed out that this meditating is costing us a month.

    Michel Hack said we can use email.

    Dave moved that we re-adopt 754-1985 as our base text & that people should use the time between now & the next meeting to write specific proposals for changes. Ian seconded.

    Prof Kahan doesn't believe we have the time for this.

    Jeff thinks some is easy (quad).

    John Harrison points out that there were substantial disagreement & that just moving it to an annex doesn't solve that.

    Jim would not expect a proposal except that this is what the committee will do.

    Michel Hack is surprised we want to throw out so much to go back to 1985.

    Mark agreed that going back to 1985 is too far & would rather go back to 2005.

    Mark porposed amended that we go to the 2005 document as the base document. Eric S seconded.

    Dave refered to our old ballot when we couldn't solve this problem on the web past year.

    Pamela asked what the relationship is to the base document & voting procedures.

    Dan asked what the vote was when we asked this question (of going back to the 1985 document) last year (0-10).

    Jim thinks we can start with 754 & add to it whereas doesn't see what we do with the 2005 document.

    Mark thinks we can start with the 2005 draft & get people to figure out what they want to change by next month.

    Jeff asked for a break at 3:20 & the chair called one.

    Prof Kahan asked what is the reason behind the amendment?

    Mark clarified that he believes it is closer to what we want than the 1985 document.

    Roger said something about 2/3s votes. It appears to be a misunderstanding.

    Peter Tang asked if Mark's proposal is to start with the 2005 document rather than the 1985 standard.

    Mark basically said yes.

    Jim discussed the text color convention.

    John Crawford pointed out that an implication of going back to the October draft predates the decimal agreement.

    Mike

    Roger parlementary procedure.

    Mark

    John

    Peter Tang is lost in the parlementary procedure.

    Jeff

    Jim pointed out the obvious.

    The vote on Mark's amendment was called for. It was 11-12-2 & the amendment is defeated.

    The vote on Dave's original proposal was called for. 10-12-3 the motion is defeated.

    Dave moved that we adopt the Phoenix draft. Peter Tang seconded.

    Prof Kahan moved to table the motion for at least two weeks.

    Dick seconded.

    Pamela pointed out that should Dave's motion fail that we are back at the beginning.

    Roger wants closure today & this completes the symmetry of the votes.

    Peter Tang points out that people may vote against because they haven't had a chance to read it.

    Eric P agreed.

    Jeff agreed with Roger.

    John Harrison feared the reverse.

    Jim suggested a 3 way vote.

    Prof Kahan withdrew his motion to table & amended Dave's motion to vote on it electronically in two weeks. Peter M seconded it.

    Jim asked if there are rules of order that prevents

    Dave called the question. Amendment passes. We read P & vote by email.

    Dick

    Ian

    Peter M thinks we should identify points of broad agreement & start from there. He moved we have electronic ballots seeking guidance on inclusion in the new standard of each of :

    • quad precision,
    • FMA,
    • decimal arithmetic (?),
    • 2005 conditionals (?)

    Jeff seconded.

    Jeff opined that this motion seeks guidance.

    Jim has a problem with the content of the list.

    Pamela asked what we were voting on.

    Peter Tang answered guidance rather than any specific text.

    Dick thinks those that pass should be on the agenda next time.

    Dave will vote against anything that is not specific.

    Prof Kahan said relagating decimal to a separate document is a possibility. Discussions in principle are possible.

    Dave said its not 2001 anymore.

    Jeff agrees with Dave if we have a base document which sounds like he doesn't agree.

    Roger agreed with Jeff.

    The chair called for the vote. 23-2-0 the motion is passed.

    Dave moved that among our electronic ballots we seek guidance on the new standard for inclusion each of:

    • predictable & controllable expression evaluation,
    • alternate exception handling mechanisms,
    • no dynamic global modes & flags.

    Peter M seconded.

    The vote was called for 16-1-0 the motion passes.

    Prof Kahan would be willing to provide an explanation of Dave's proposal.

    Dave believes we all knows what the issues are.

  9. Adjournment

(After the meeting adjourned, Prof Kahan asked me to include the following text by reference: RNDGMODE.TXT)

Next Meetings

Wednesday 3/15 full meeting, Joe Darcy at Sun in Santa Clara
Thursday 3/16 draft review, Joe Darcy at Sun in Santa Clara

Wednesday 4/19 full meeting, Jeff Kidder at Intel in Santa
Thursday 4/20 draft review

Wednesday 5/17 full meeting, Roger Golliver at Intel in Santa
Thursday 5/18 draft review

Wednesday 6/21 full meeting, Jim Thomas at HP in Cupertino
Thursday 6/22 draft review

Monday 7/10 6:00 MSC

Wednesday 7/19 full meeting, David Hough at Sun in Menlo Park
Thursday 7/20 draft review

Wednesday 8/16 full meeting, Eric Postpischil at Apple in Ten Forward
Thursday 8/17 draft review, Conference Room at 2 Infinite Loop, Cupertino

Wednesday 9/20 full meeting, TBA
Thursday 9/21 draft review

Monday 10/9 6:00 MSC

Wednesday 10/18 full meeting, TBA
Thursday 10/19 draft review

Wednesday 11/15 full meeting, Eric Postpischil at Apple in Ten Forward
Thursday 11/16 draft review, Conference Room at 2 Infinite Loop, Cupertino

Wednesday 12/20 full meeting, TBA
Thursday 12/21 draft review

754's PAR expires 12/06

Monday 1/8 6:00 MSC

Monday 4/9 6:00 MSC

Monday 7/9 6:00 MSC

Monday 10/8 6:00 MSC

754 | revision | FAQ | references