Open XML Still Likely to Be Certified

Last Friday's news that Microsoft's Office Open XML failed to pass a letter ballot for recommendation by the Executive Board of INCITS to the International Organization for Standardization (ISO) without comments, was interpreted yet again by press sources as an outright failure of the standard, and the end of the road for global acceptance. But a check of comments posted by voting members, and a re-read of what rules the INCITS group has posted, says otherwise.

The test which OOXML failed last week was whether it could be recommended by the Board without any concerns being raised by members, regardless of whether they approve or disapprove. BetaNews has been told conflicting accounts of how the rules of INCITS actually work.

On Saturday, BetaNews heard from IBM software architect Rob Weir, who represents his company to the Executive Board. Weir explained that the final letter ballot would have required a two-thirds majority of voting members, regardless of whether any abstentions were cast. "So the ballot to approve OOXML failed on both counts," Weir wrote.

"Part of the problem is that many of these rules, in INCITS and in JTC1 [Joint Technical Committee 1], have not really been exercised in all of their details," he wrote in a later follow-up to BetaNews. "The vast majority of standards ballots are unanimous or at least with clear and wide consensus. It is only rare cases like this, where there is a lot of contention, that we need to figure out what the details are. The scary part is when we find that the details aren't... well...very detailed.

"So think Florida 2000 and hanging chads," Weir went on. "A close election causes every rule, every appeal, every avenue to be used. When that happens you often find out that the rules weren't thought out so carefully."

But the chairman of another INCITS Executive Board member company wrote an open letter to one of our sources on this topic -- Linux Foundation board member and open source attorney Andrew Updegrove -- saying that the current voting procedure is just part of the standard procedure a method or process or format goes through to eventually get accepted as an international standard - a procedure that, from his point of view, is already well thought-out. This from a company that, like IBM, voted no on the letter ballot.

Farance, Inc. chairman Frank Farance wrote for ConsortiumInfo.org, "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."

Indeed, Farance's impressions are indicated by the company's formal comments on last week's vote, which have since been publicly posted.

"The reason for our NO vote is as follows," begins Farance, Inc.'s official comments: "The JTC1 Directives subclause 9.8 states, 'Conditional approval should be submitted as a disapproval vote.' It is clearly the case that the processed comments from INCITS/V1 are the consensus of technical experts, many of the proposed changes are substantive in nature, and the proposed changes need to be adopted by the DIS 29500 [Ballot Resolution Meeting]."

The company formally recommended that the US position on OOXML -- for now -- should be "Disapprove with comments." This is stage 5 of a six-stage relative approval scale, the bottom of which (#6) means permanent rejection. But stage 5 allows for "conditional approval," which means that rules are spelled out regarding how the submitters of a prospective standard address the comments and redeem themselves.

As Farance, Inc. suggested, INCITS should acquire all comments from Executive Board members, and group them into two batches: "Batch #1" for important matters that cannot be overlooked, and "Batch #2" for "for your information" matters that it would be nice if someone addressed, but not critical. "The US will change its vote from DISAPPROVE to APPROVE upon the adoption and acceptance of the proposed changes in Batch #1 comments," reads Farance's comments.

Based on that suggestion, BetaNews examined all the comments made available by Executive Board members (not all members posted comments, though the abstaining IEEE was required to do so). Most no voters sided with Farance, saying if comments and concerns were adequately addressed, they would be willing to change their vote during a later ballot to yes. Assuming that date of reckoning did arrive and all was satisfactory, we believe the final vote for that ballot could be 14-1-1.

Even IBM would vote in the affirmative, we were surprised to learn. "Directives are clear that in order to be assured that your comments will be addressed in a satisfactory manner you should vote 'NO,"' its official comments read. "IBM is willing to change the vote to a YES if the US changes its position to a 'NO' with comments."

The IEEE would continue to abstain, citing irreconcilable differences within its own membership. The lone no voter, we believe, could be the US Dept. of Defense, whose concerns may supersede their ability to be addressed with fixes or workarounds. Among them was this: "The use of proprietary file formats within the open standard appear to cause potential intellectual property ownership concerns."

Of course, all this presumes Microsoft's willingness to address the Executive Board's concerns. Thus, we may rightly be left scratching our heads with regard to its apparently dismissive attitude in its comments accompanying its fully anticipated yes vote, concerning recommendations that INCITS received from the outside:

"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."

At any rate, this is clearly not the end for OOXML by any means.