The Standards Blog

The *Complete* Story on the US INCITS Vote

OpenDocument and OOXML

I headed in to town from the desert this afternoon to gas up and get groceries, and to catch up on all things ODF/OOXML.  In scanning my Google Alerts, I ran into this posting by Microsoft's Jason Matusow, himself just in from vacation.  In that post, Jason writes as follows:

Even though there were early predictions of doom for Open XML from Andy Updegrove and Rob Weir (and others), the US vote is likely to be either a “Yes with comments” or “Abstain” – not a ”No” vote. While the parties opposed to ISO adoption of Open XML have gone quiet on the US vote in the blogosphere, I think it is worth taking a close look at this key vote. In order to clarify my opinion – here are the details as I understand them.

Well, it's hard to take a vacation, isn't it?  Not only is it styled as "going quiet," but it offers an opportunity for others to present only part of the story.  While much of what Jason writes is accurate, it's curious what he leaves out - including the fact that not one ballot, but two, have been circulated to the INCITS Executive Board for simultaneous voting.  According to Jason's blog entry again:

By the end of the meeting enough of those who originally cast a “No” vote indicated likely support for a second “Yes with Comments” ballot to begin on Thursday August 16. Thus, the ballot will move to the next phase as “Yes  with Comments” heading into a Resolution Meeting on August 29. At that meeting, if Open XML gets 10 supporting votes, the US position on Open XML will be “Yes with Comments.” If it does not get the 10 needed votes, the EB is being asked to consider “Abstain with Comments” as its fall-back position.   At this point, it seems a “No with Comments” is off the table.

To read that, you would assume that there is a single ballot under consideration.  Curiously enough, there are two ballots that have been distributed, on an equal voting, and either - or neither - may be approved. One  ballot is to approve with comments, and one to abstain with comments.  Here, then, is the whole story, as given to me during a lengthy phone call with someone who attended the meeting, as well as the schedule and possible outcomes during the time remaining before the opportunity to submit a US position expires. 

The August 15 Meeting
As I previously reported, the Executive Board (EB) of INCITS issued a written ballot to "Approve OOXML – With Comments" in July, which failed to achieve the necessary majority. The EB had already scheduled a "reconciliation" meeting to determine what to do next in anticipation of the rapidly approaching September 2 deadline, depending on how the ballot came out. 
It should be noted that the National Body rules are directed towards perfecting and approving a submitted standard, rather than simply voting upon them on an "up or down" basis. Perhaps this is because there is a presumption of validity and utility for any specification that has made it this far (i.e., it has either already been adopted by a national member, been developed through the ISO/IEC JTC1 process, or been adopted by a standards organization through the PAS process, ostensibly indicating that it is a useful and technically complete standard). Accordingly, the order of the August 15 meeting, chaired by a Lexmark representative, was to see whether objections could be resolved sufficiently to allow OOXML to be approved, and virtually the entire day was spent in pursuit of that objective.;
15 out of the 16 eligible EB members participated in the August 15 meeting. Early in the day, a series of straw votes was held to establish starting positions. The first vote was the same as the written ballot that had just closed – to approve, with the final list of comments yet to be determined by the EB. The Lexmark representative announced that he had now seen enough of his comments satisfied that he would switch his vote to Approve. All others present cast their votes as before.   The vote of 8 yes, 7 no and 1 abstaining from the written ballot now stood at 9 in favor, which would still not be sufficient to approve with comments.
The next straw vote was to Disapprove – With Comments. It came in as follows: 8 yes, 5 no, 2 abstaining (1 absent).
The third and final straw vote was to Abstain – With Comments. This time the votes were 9 yes, 5 no, 1 abstaining (one absent).
In order of consensus, then, there was more support for approval or abstention with comments than for non-approval with comments, but not enough support for any of the three to be successful.
The meeting then turned to making further progress on which comments would be submitted along with the eventual vote, whatever that vote might be. It was agreed that all letters of support of one position or another that were not accompanied by specific criticisms would be eliminated, reducing the total number of potential comments from c. 480 to about 300.
At the end of the day, it was agreed to issue two ballots rather than one, with those two ballots to be the two votes that had polled the greatest support earlier in the day – one ballot to approve, with comments, and one to abstain, with comments (both would be considered to be desirable by Microsoft, and undesirable by those opposing OOXML, at least in its current form). The comments to be referenced in these votes would be the updated comments reflecting the results of the current meeting. As before, the voting period would be seven days.  The new ballots were distributed on August 16.
What Happens Next
A final reconciliation meeting will be held on August 29. If one of the two balloted votes achieves the necessary majority, then presumably that will be the US vote. If neither vote achieves the requisite majority, then any result, including a no vote, or no result at all, is possible, although it appears most likely from what I can tell that one of the two balloted votes is, at this point, most likely to prevail.
What could upset the result one way or another? On the one hand, under the prevailing rules the discussions are focusing not on whether two standards are needed or other concerns, but over whether OOXML in its current form is in proper form technically to achieve status as an ISO/IEC standard. At the same time, tensions and emotions continue to run high, both within and outside the process. I am told that during the meeting it was noted that Steve Ballmer personally called Secretary of Commerce Carlos M. Gutierrez to lobby for approval (NIST has a vote on the EB). 
Turning to Jason's August 19  blog entry again, he stresses once again that the gloves are off:
Everyone with a vested interest in this – both pro and con – are working with all of the tools available to them. No matter what people argue about, though – at the heart of this remains the idea that making document formats more open is a good thing.
What exactly does he mean by "using all of the tools?"  Here is an example, taken from The Economic Times of India, in a story bylined today.  The article is an interview with Vijay Kapoor, national technology officer of Microsoft India, and the subject is whether or not India should approve OOXML.  This is a copyrighted article, so I will limit myself to including one question and answer, with the actual facts added by me [in italicized bracketed text].  I do not expect anyone at Microsoft to challenge my corrections: 
Question:  Why does Microsoft want another standard, what's the rationale?

Answer:  There are at least 4 good reasons why:

*ODF started out and was completed as an XML format, specifically supporting OpenOffice with a tight scope around that product.

[ODF did start as a description of OpenOffice.  This is not uncommon in standards setting - to use a not-over specific specification for an existing product as a starting.  But the committee had no "scope" to stay there.  Through a multi-year process, it was developed into a specification with many changes. [Update:  Te original Call for Participation is too long to include here, but clearly states that while OpenOffice would be used as a starting point during the first phase of the TC's work, the second phase would move beyond it.]
In contrast, OOXML not only started as a description of Office, but with a mandate at Ecma to describe that product and no other.  There are many implementations of ODF today in addition to OpenOffice.]

*It wasn't until 2005 that the spec was offered up as a general XML office document format and consequently renamed to ODF.

[Blatantly false.  ODF had been under development within OASIS for several years at that point.[Update: since December 16, 2002, the first meeting of the ODF Technical Committee.]  It was formally adopted by OASIS in May of 2005]

*No opportunity existed for Microsoft to actually participate in this full process - given the original scope, the 6 months between the re-naming of the spec to ODF, and its subsequent approval by OASIS as a standard.

[Blatantly false.  Microsoft was a member of OASIS throughout the entire development process, and at any time could have joined the Technical Committee and influenced the result.  If memory serves, Microsoft was even represented on the Board during part of this period.  [Update:  Microsoft's Chris Kurt was on the OASIS board since 2001 or 2002, and remained on the board until last year.]

*The scope of the ODF spec never included even the basic requirements that Microsoft required to support a fully open format, and nor did the OASIS technical committee want to include these requirements.[Blatantly false.  In September of 2005, I interviewed Mary McRae, OASIS Manager of Technical Committee Administration and other key OASIS, Microsoft and Massachusetts participants in the ODF debate.  Mary stated that every effort was made to make ODF as compatible as it could, absent the cooperation of Microsoft.]

I find myself increasingly appalled by Microsoft's behavior, which is turning the accredited standards process into a laughingstock.  It's spokespersons seem to feel that they are under no compunction to be even passingly accurate in their statements.  And despite Jason's statements, I do not find similarly missleading statements being made by anyone representing (for example) Sun or IBM.
And consider this as well:  I am also told that two countries in recent weeks have upgraded themselves in ISO to "P" status (there were c. 30 P members previously). And I have heard rumors that more countries may, at the last minute, upgrade their status. The voting rules in JTC1 are very confusing and complex (I will describe them in a future post), but suffice it to say that modest changes to the composition of the P membership can have a profound change on the final outcome. If this happens, and those countries all vote to approve, we will have seen what appears to be a pattern of stacking the vote not only within individual National Bodies through sudden increases of membership, but stacking the global vote as well.
Could such a blatant tactic be utilized? Technically, the answer is yes. Under the ISO rules, any member may upgrade at any time, with immediate effect. On the question of probabilities, I am reminded of another recent quote by Jason Matusow at his blog: There is no question that all over the world the competing interests in the Open XML standardization process are going to use all tactics available to them within the rules.
If there are reports of further upgrading of National Bodies, perhaps resistance to OOXML may stiffen within the EB, notwithstanding the technical focus of their deliberations, and a final vote may (or, in the view of some, should) revert to "No – With Comments." The reason, of course, is that notwithstanding the comment of Frank Farrance in response to an earlier post of mine, while all comments need to be "addressed," regardless of whether the vote to which they relate is to approve, disapprove or abstain, a comment can be addressed by determining that no change is in fact necessary.  Moreover the comments will be addressed in the first instance by Ecma, the same organization that passed a standard along to ISO/IEC that was met with over a thousand comments, including many serious deficiencies and incompatibilities.  IThe members of SO/IEC JTC1, in contrast, will have only a few weeks to review what Ecma gives back before its own reconciliation meeting is held in February of 2008.
No, on this record I think that it is best for comments to be appended to "no" votes.  To paraphrase Samuel Johnson, "Depend upon it, Jason, when a vendor knows that its specification may will be rejected in a fortnight, it concentrates his mind wonderfully."
It will be interesting to see what finally transpires on September 2, both at home and abroad. I have long observed that Microsoft has been playing its hand as aggressively as possible, and perhaps more dangerously than is wise for its own good, leaving aside its impact on the credibility of the accredited standard setting process. Will its bet pay off, or will it finally so overplay its hand that a backlash proves to be its undoing?
We won't have to wait long to find out, at least to learn the end of the chapter of this enduring saga that ends on September 2.

For further blog entries on ODF and OOXML, click here

subscribe to the free Consortium Standards Bulletin



In case you're not aware of him, you should keep an eye on Alex Brown, who has recently been named as the convenor of the BRM. He's stated that "the process is on rails" now, and has left a comment on a blog stating that abstainers can attend the BRM, and can change their vote to either 'yes' or 'no'. As such, INCITS' abstention would be no problem for OOXML opponents.

Also, as a regular commenter on Matusow's blog, I'd like to make a partial defence of Jason's statements. One of the few strong conclusions I've come to about this whole process is that there's a gulf of understanding between Microsoft and the wider community. My explanation for this gulf is that they've been doing their own thing for the past 20 years, and have come to very different conclusions about lots of issues. As a programmer, I normally use a programming analogy at this point, but since you're a lawyer, I'll try my hand at a legal analogy instead: the American legal system split away from the British system when America declared independence, so both systems share many fundamental similarities, but have come to opposing conclusions on many issues, and I'll bet they have incompatible definitions for many important technical terms too.

Talking to Microsoft is a little like talking to a British lawyer: you think everything's going fine, until they start making nonsensical claims and you realise you've been talking about subtly different things for the past little while. I generally find that when you drill down into Microsoft's arguments, claims that seem absurd on their surface turn out to be rational claims based on different assumptions. I've not yet got fully to grips with Jason's "gloves are off" comments, but his position seems to be that IBM took their gloves off way back (e.g. by waiting until OOXML had been ratified by ECMA until they started raising technical complaints), and as such Microsoft's actions are a legitimate reaction.

- Andrew Sayers


I agree with part of what you say (my own favorite metaphor would have to do with evolution, and the divergence of species), but only up to a point.  I think it's very true that folks at Microsoft really think that they are making great products.  The problem is, that where they have been dominant for so long, they've lost touch with the marketplace and their customers.  So when they think they are giving the customer, "what the customer wants," what they are really doing is listeningn to all of the things the customers say, and then to their own self interest, and then giving the customer what they want the customer to have.  There's no substitute for competition to make you really deliver what the customer wants.  Yet when you talk to folks from Microsoft, they really don't understand that.

The next layer, though, really is pure FUD.  They're constantly talking about giving the customer what the customer wants when it's totaly in Microsoft's self interest.  Of course, all vendors play the same game.  It's the tactics here that I find over the top (see the Indian interview, for example).  I really mean it when I say that I just don't hear the same sorts of stories about IBM and Sun as I do about Microsoft.

On whether IBM held back:  OOXML was a proprietary spec.  I do know that as soon as it was made public by Ecma, IBM (and everyone else) was all over it as quickly as they could, given that it is so long.  IBM is a member of Ecma, and I don't know how much, and when they said there.  I also think it's not exactly fair to expect a competitor to expose your issues for you, especially when your whole strategy is ram it though as quickly as possible, issues be damned.  Microsoft's goal here is clearly to get it accepted by ISO/IEC as quickly as possible, which means not having to deal with any more issues than they absolutely have to - that's clearly what they did during the comment period in ISO.  They could have addressed issues then, but instead pushed ahead.

-  Andy

I would be wiling to ascribe much of Microsoft's apparent wrong-headessdness about protocols and file formats to the kind of subtle worldview clash you suggest, were it not for the internal Microsoft documents (the "Halloween papers" and others made public in recent court cases) that show Microsoft intentionally manipulating protocols, API's, and file formats for competitive advantage in just the way their detractors claim.

Right--this is what bothers me even more than MS' aggressive tactics: I don't for one second believe they are serious about producing an office suite that holds to any community-governed standard. They are fighting "with all available tools" for nothing more than a rubber stamp that says "ISO Approved" to place on their marketing brochures and government purchasing documents.

For MS, compatibility is not driven by a standard, it is driven by the fact that all the developers work at the same company. It doesn't matter what's in the standard document. This works great when the customer is willing to be indentured to one vendor forever, and many customers are willing to make the trade, but it does not produce the long-term benefits of using a true open standard.

I only hope that ISO looks beyond the politics and technical arguments to see that OOXML is not now, and never will be, a meaningful international standard.

here is what will happen

it will get approved and then afterwards we will hear of all the migrations to vista and office because it open standard from all these countries that got upgraded to P status.

I for one welcome our winders overlords.

<blockquote>It will be interesting to see what finally transpires on September 2, both at home and abroad. I have long observed that Microsoft has been playing its hand as aggressively as possible, and perhaps more dangerously than is wise for its own good, leaving aside its impact on the credibility of the accredited standard setting process. Will its bet pay off, or will it finally so overplay its hand that a backlash proves to be its undoing?</blockquote>

I wonder...could the U.N. bring Anti-Trust charges against Microsoft? It's already had U.S. and the E.U. convict it of anti-trust behavior. Wonder if the U.N. could too...wonder what that would say if it happened...

For better or worse, I'm not aware of any UN treaties that deal with antitrust type offenses.  The closest thing would be the World Trade Organization Agreement on Technical Barriers to Trade, which does address misuse of standards, but that relates to actions by nations, not companies. 

  -  Andy