OOXML Approval Vote Fails in INCITS

As I reported on July 23, INCITS, the US balloting body on the OOXML vote, put out a ballot to see whether the US should vote to approve OOXML, with the ballot to close on August 9.  That ballot has now closed on schedule, and there is a public link that shows the vote - which failed, with 8 in favor, 7 opposed, and one abstaining.  As I noted previously, a vote of 9 in favor would have been required for passage.  That number is a simple majority of the 16 INCITS Executive Board members that have voting privileges on this ballot (in fact, the Board has 18 members, but due to attendance rules, only 16 of the 18 had voting priviliges on this ballot).

There is a second leg of the vote, which also failed:  out of the total number responding (in this case, all 16), the abstentions (one) are subtracted, yielding a number (fifteen) of which two-thirds (in this case ten) would need to be in the affirmative.

The link above includes links to the individual comments filed by eleven Executive Board members.

As you will recall from my prior post, a schedule had been pre-agreed upon that could accommodate a further vote of "no with comments."  I have no further information on that at this time, but will report when and as I do.

Here are the votes:

  • Yes votes:  Apple, Department of Homeland Security, the Electronic Industries Allliance, EMC, Hewlett Packard, Intel, Microsoft and Sony Electronics.
  • No votes:  Farance, Incorporated, GS1 US, IBM, Lexmark International, NIST, Oracle, and the Department of Defense.
  • Abstention:  IEEE

There are a number of interesting things to note about the voting.  First, EIA voted for, while the IEEE abstained, although both are industry organizations.  Why the split?  IEEE is, of course, an organization in which all of the large corporate members referred to above are active members.  Not surprisingly, the brief comment that accompanied IEEE’s abstention reads: "IEEE abstains due to the divergent viewpoints of key IEEE members and stakeholders." EIA is an umbrella organization of six other industry organizations.  It is also currently influx, and undergoing a reorganzition.  EIA did not provide a comment with its vote.

The government members of the Executive Board also split, with two agencies voting against, and one voting in favor.  DoD and NIST, of course, are old agencies, with long involvement in the standards arena, while Homeland Security is new. 

[The following is an update]

The NIST vote, however, requires a special explanation, especially in light of a seemingly contradictory press release issued by NIST earlier today.  In that press release, it states that NIST "has voted for conditional approval of a proposed international standard for open documents."  The INCITS voting page for NIST and its comments, contains the explanation, which reads in part as follows:

Based upon the technical comments identified, NIST believes that the US National Body should be voting for conditional approval to DIS 29500. The JTC 1 procedures in clause 9.8, Votes on Fast-track DISs, contain the note: “[Note: Conditional approval should be submitted as a disapproval vote.]” While this is advisory (i.e., should versus shall), it is the best way to ensure that the comments submitted by the US National Body are given careful consideration.

As noted earlier, a "yes" vote with comments does not require consideration of the comments, so a "yes with comments" is an unconditional approval.  If someone voting thinks that OOXML, if properly documented, is entitled to become a standard, but should not be a standard in its current form, then the approved procedure, as noted in the NIST comment, is to "disapprove."  Where NIST has muddied the water is by not being more explanatory in its press release – in effect, it opted to base its press release on its desired outcome, rather than its actual vote.  Presumably, they threw up their hands at trying to explain the complexities of the ISO/IEC process in a press release.

[End of update]

Among the corporate members, I’m somewhat surprised at Hewlett Packard’s vote.  While they do not have a direct stake in the OOXML v. ODF competition, they are a stalwart participant in the standards infrastructure, and well experienced. 

The overall vote is not, to me, a surprise.  Leaving all politics or opinions aside on whether OOXML should become an ISO/IEC standard when ODF already exists, I do not believe that OOXML is in good enough shape to recommend a yes vote, with or without comments.

The brief comments supplied by eleven of the Executive Board members are of particular interest in this regard.  Here is the comment that accompanied IBM’s "no" vote, for example:

IBM Comments – http://www.incits.org/ref-docs/in071208.zip IBM is voting "NO" because they are substantial technical issues/comments(min 476), IPR issues, accessibility comments and the JTC 1 Directives are clear that in order to be assured that your comments will be addressed in a satisfactory manner you should vote "NO"..IBM is willing to change the vote to a YES if the US changes its position to a "NO" with comments

Several of the comments relate to what external comments should, or should not, be included with a US vote.  Here is the comment supplied by Microsoft along with it’s "yes" vote:

Microsoft further believes that (a) the 205 letters of general support/opposition/caution received from the public are not actionable at the BRM and should not be submitted; and (b) that the remaining 186 ‘unprocessed/unapproved’ comments should not be submitted as supported US comments as they were never adopted by INCITS/V1, and should be simply transmitted to the Submitter for their review and appropriate treatment.

I haven’t had time to review the other comments in detail, but will update this post later today (I’m in a meeting at the moment) when I have the opportunity to do so.

Update:  This article from Ars Technica is instructive and useful.  Be sure to also read the comment below by Frank Farance, an INCITS Executive Board member, and my response.  And also this article, in which Eric Lai interviews Farance and receives further intormation.

For further blog entries on ODF and OOXML, click here

subscribe to  Standards Today – a free eJournal of News, Ideas and Analysis

Comments (21)

  1. Seems like you guys have some additional educating to do (either that or DHS is being disingenuous).  I read in the DHS comments "Both standards seem to be the XML-ization of legacy software binary file formats."  Err… while that’s certainly true for OOXML, I don’t believe that’s true for ODF.   Funny, that after all this time, DHS would get such a basic fact wrong.

    — Swashbuckler

  2. About 30 years ago, the equivalent question would have been "Shall we stay with punch-cards (EBCDIC) or shall we move to paper-tape (ASCII)"

    You moved to ASCII, mostly. Except for IBM mainframes, which are EBCDIC to this day.

    Was the shift good, bad, or inevitable ?

    • I disagree that this is the same as moving from EBCDIC to ASCII.  ASCII is clearly defined.

  3. Thank you for the timely information….Wow, by the skin of our teeth!

    As to motives for the DHS, HP, and IEEE votes, my speculation starts with a mentor’s best advice:  If you don’t understand the behavior check the compensation plan. 

    DHS?…of the members here, possibly the USGov Executive Branch agency with closest ties to the current Administration?  The administration that "lobbied EC over Microsoft fines".  As for HP, well, the old HP, the  engineering company would get it – that’s Agilent.  The current HP sells a lot of MS products.  IEEE?  Who knows.

    THANK YOU to all the SANE members!
    Tom, a Citizen, without affiliation to any corporate (or free software) organization.

    p.s. to MS:  You could be first and best with ODF tools for desktop and the enterprise.  1/2 a market 10 times bigger then the current little ms-office market is billions of new dollars.  You are failing your shareholders and burning good will world wide.

    • If you don’t understand the behavior check the compensation plan.

      Your mentor’s good advice went straight to my personal quotes file…

  4. out of the total number responding (in this case, all 16), the abstentions (one) are subtracted, yielding a number (ten) of which two-thirds (in this case 10) would need to be in the affirmative.

    that first "ten" should be "fifteen" ("16" subtract "one").

    • It’s not meant to; it’s an update which is meant to be applying the prior rules consistently to the new information (including figuring out what the disconnect was between 16 and 18 members on the EB).  But if you’re seeing a problem, let me know what you’re seeing, and I’ll try and address it.

        –  Andy

      • The previous commenter meant that the news.com story he linked seems to be in conflict, as it indicates that NIST approved OOXML.

        The news.com reporter was misled by NIST’s comments.   NIST wished to have "conditional approval" in which the outstanding comments against OOXML were addressed, which in this case means that they should (and did) vote No on this ballot.  I’m thoroughly shocked, of course, that the media got this wrong.

      • Sorry – I didn’t see your link.  I thought you were talking about my prior blog entry.  Based on the INCITS link, it looks like Martin misunderstood something.

          –  Admin

  5. HP is not the old company it used to be with lots of history in standards.  The new HP is a computer and printer company that is heavily dependent on Microsoft.

  6. Apologies Andy for Off Topic but just to spread the word:

    SCO have lost their claims on UNIX to Novell – the latter officially own the UNIX copyrights so IBM case is dead in the water too. Watch the fallout over the next few days.

    PJ as usual has all the details at Groklaw


    • Absolutely.  I’ve already added it to the News Picks.

        –  Admin

    • I don’t, unfortunately, know whether they are available on line, and if so, where (although Googling them using the designiations given) should answer that question, I should think).  I expect that the bulk of them are ones that have been aired elsewhere already (e.g., Rob Weir’s list), so you should be able to have a pretty good handle on them throught a general investigation.

        –  Admin

  7. I’m Frank Farance and my company is a member of the INCITS Executive Board.  We’ve been a member of INCITS committees for over 20 years.

    It seems that there is a misunderstanding about the process here wth "OOXML Approval Vote Fails In INCITS", as if INCITS EB members don’t want the document approved as an ISO/IEC standard.  My personal impression is that the document will become a standard after some changes have been made through the Ballot Resolution Meeting (BRM), which will start in early 2008 and, depending upon the number of comments received from National Bodies, might take 3-6 months to complete.  It will be at the end of the BRM process  (i.e., where we’ve processed all the comments and we’ll know what changes will be necessary) that we will determine whether or not the document will become an ISO/IEC standard.

    Right now, the INCITS EB is formulating the US position and that position will be submitted to ISO/IEC on 2007-09-02.  There are several possible votes: (1) Approve as Presented, (2) Approve with Comments, (3) Abstain, (4) Abstention with Comments, (5) Disapprove with Comments (conditional approval), and (6) Disapprove with Comments ("No, and hell no!").  Our first choice was to try to build consensus around choice #2, but that failed to achieve the 2/3 consensus.  For the INCITS EB, consensus is a 2/3 vote, so given that one member abstains with 15 others voting, that would require 10 Yes votes (not a majority but a 2/3 majority).

    In anticipation of the contentious nature of this vote, the INCITS EB had already scheduled meetings to meet the international deadline.  We’ll have a ballot resolution meeting on 2007-08-15 and we’ll hopefully build consensus around a different ballot question, which will also go through the formal balloting process with 2/3 approval of INCITS EB members.

    What is misleading about the title of the article that I’m responding to ("OOXML Approval Vote Fails In INCITS") is that it conveys the idea that INCITS EB members don’t want the document to become a standard, and I don’t think that is true.  Many of the No votes (including my company’s) cited the JTC1 Directives, subclause 9.8 that explains how to convey "Conditional Approval".  Thus, if the US position is "Disapprove with Comments (Conditional Approval)", then if the US’ technical comments are accepted and adopted by the BRM (assuming everything else is the same), then the US will change its vote from DISAPPROVE to APPROVE at the conclusion of the BRM.  Assuming other National Bodies (NBs) have cast Conditional Approval votes, and assuming that their comments have been accepted and adopted by the BRM, then they too would change their votes from DISAPPROVE to APPROVE.

    This process of Conditional Approval (Disapprove with Comments) is how we build consensus in ISO/IEC standardization.  Assuming the BRM (which consists of delegates from the NBs voting on the document) adopts and accepts all the changes and assuming that the 2/3 majority of NBs vote APPROVE at the end of the BRM (and we don’t have a large number of abstains and DISAPPROVE votes), then the document with the approved changes will become an ISO/IEC standard.

    For the INCITS EB members that voted No and gave the rationale that they wanted "Disapprove with Comments (Conditional Approval)", these members (Farance, GS1, IBM, NIST, Oracle) want the document approved as a standard, but want the US’ vote registered as Disapprove with Comments (Conditional Approval).

    In the article I’m responding to, it states that one can ignore the comments of an Approve with Comments vote ("As noted earlier, a "yes" vote with comments does not require consideration of the comment …").  That is not true.  All comments are addressed within a BRM.  The distinction between Approve with Comments vs. Disapprove with Comments (Conditional Approval) is that in the Conditional Approval, one indicates which of the comments (one, some, all) are the conditions for changing the DISAPPROVE vote to an APPROVE vote.  In short, the distinction is: "Approve with Comments" means the NB would still be accepting of the document (i.e., voting APPROVE at the end of the BRM) even if none of the comments were accepted/adopted, while "Disapprove with Comments (Conditional Approval)" means that some or all of the comments are "must-have" and the NB would not switch to APPROVE unless these comments were addressed satisfactorily (e.g., having them accepted and adopted).

    The issue of "can-ignore" vs. "must-have" is the essence of concern for INCITS EB members that voted No and cited the JTC1 Directives subclause 9.8.  In short, we believe this ("Disapprove with Comments, Conditional Approval") is the normal way of expressing our need for certain changes to be made before the document is acceptable as an ISO/IEC standard.  (Of course, these suggested changes don’t make the document perfect and I’d guess that it will take years to discover all the errors in the 6000+ page document.)

  8. Hi Frank,

    First, thanks very much for taking the time to give this very detailed explanation.  I think that we’re actually saying the same thing, although if I wasn’t sufficiently clear (or perhaps better stated, if I was more conclusory and predictive), I appreciate your providing a more thorough and nuanced explanation.

    What I was trying to convey in a more pragmatic fashion in the difference between a "yes" and "no" vote final vote by the EB on behalf of the US was more of an "in the trenches," pragmatic observation that comments that are required to turn nay votes into yay votes are going to be given more serious attention, as well as more willingness on the part of Microsoft and Ecma to accommodate the change requests.  This would be particularly true give the current 1700 or so comments that have been lodged globally with National Bodies (many, of course, being duplicative or simply statements of support or non-support).

    On the title, I agree.  "Latest Vote" might have been a better choice.  I do state later that it is not clear to me at this point what the next decision will be, coming out of the resolution meeting this Wednesday, and that the final US position remains unknown at this point.

    Thanks again for your willingness to share your personal, first-hand views and facts here.  Unfortunately, there’s far more second hand (or worse) than first hand information available in the marketplace, and everyone would be better off if there was more of the latter.

      –  Andy

Comments are closed.