Yesterday, I sent out the latest issue of my eJournal, Standards Today. Not surprisingly, it focused on the OOXML process, and what can be learned from it. Below is the Editorial, and you can find the complete issue here. You can sign up for a free subscription here.
Updated: ISO has now issued its confirmatory press release. The full text (less biolerplate) is appended at the end of this entry. I note with some interest that the press release includes the following language:
Subject to there being no formal appeals from ISO/IEC national bodies in the next two months, the International Standard will accordingly proceed to publication.
The last issue of Standards Today was titled, ODF vs. OOXML on the Eve of the BRM. That issue focused on the Ballot Resolution Meeting (BRM) about to be held in Geneva, Switzerland as the penultimate act in the Fast Track approval process of DIS 29500, the specification submitted by Ecma and based upon Microsoft's OfficeOpen XML document formats (OOXML).
My editorial in that issue was prophetically titled The Overwhelming of ISO/IEC JTC1, due to the fact that only one week had been allocated to resolving more than 1,100 separate comments (some 900 of them substantive) that had been registered by National Bodies from around the world during the voting period that failed to approve OOXML during the initial balloting period in mid-2007.
Without exception, every fear that I raised in that editorial was realized, and worse. Here is a sampling:
Prior concern: "Due to the 6,000 page length of OOXML, not all problems are likely to have been identified during the formal review period. But any deficiencies in OOXML discovered after September 2, according to the JTC1 Directives as cited by Brown, are "out of scope," and may not be addressed at the BRM. Instead, they must await resolution in the next review cycle (i.e., years in the future)."
Reality: Far from worrying about addressing new concerns, there was (as expected) insufficient sufficient time to interactively discuss and, as necessary, revise the vast majority of old comments. One consequence was that addressing even many of the concerns submitted in 2007 were deferred to resolution during a future "maintenance phase" of the specification.
Prior concern: "It does not appear at this time as if the resolutions proposed by Ecma will be made available at a public Web site before the BRM, if ever. Consequently, the 500 million users of Office and the legions of independent software vendors whose software must be used in conjunction with Office will have no opportunity to convey their opinions to the delegates that will nominally represent their interests at the BRM."
Reality: Not only did those who were not involved have the opportunity to access the proposed resolutions for review, but one official delegation complained that it had not been able to consider any proposed resolutions other than those offered in response to their own comments.
Prior Concern: "The final vote on OOXML will follow the conclusion of the BRM, whether or not all comment resolutions have been resolved. It appears that if the vote is in favor of adoption, unresolved comments will not be dealt with, if ever, until the next review cycle."
Reality: Only a small percentage of the c. 900 substantive resolutions were interactively discussed and, as necessary, revised. The remainder was disposed of in a process that allowed each delegation to vote upon each resolution that it wished to weigh in on, and to make a blanket choice of "approve," "disapprove" or "abstain" as to the balance, if desired. The time permitted for voting on all c. 900 proposed resolutions (comprising well over 1,000 pages of text) was less than 24 hours.
Confronted with this impossible task, only six delegations chose to approve and four chose to disapprove. Of the remainder, 18 chose "abstain" — and four chose not to register a position at all. As one delegate stated in the meeting, "If this was all that would be permitted, I would have preferred to have stayed at home and had two weeks to consider how to vote." Despite the fact that in the ordinary course all resolutions would be discussed, during as many meetings as needed, OOXML proponents announced that the BRM "was an unqualified success."
Prior Concern: "No outsiders will be allowed to attend the BRM, nor will any transcript be prepared and made available."
Reality: Only a skeletal summary of the actions discussed and resolutions adopted was made available. Moreover, those in attendance were requested not to discuss anything that transpired during the BRM with anyone outside the meeting, either during the course of the meeting or afterwards. Not surprisingly, the result is that widely different accounts were posted even by those who had attended the BRM, ranging from the pronouncement by one delegate that the BRM had been "complete, utter, unadulterated bullshit," and the statement by another that "The process really worked (it was very cool)." (I have provided links and excerpts from the accounts of nine delegates from 7 countries here, and much more original source material here, so that those that are interested can form their own judgment.)
In the absence of a detailed official record or the admission of the press or any other neutral third party, those around the world whose lives will be impacted by the final result can only scratch their heads and wonder what just happened, and who to believe.
While every first hand account applauded the efforts of Convenor Alex Brown and of the delegates at making the best of the situation and achieving the greatest degree of improvement in DIS 29500 possible, the result by anyone’s account was the submission for final voting of a specification that had received less attention and collaborative effort to improve its quality than would typically be the case under any other circumstance.
And there was more to come. During the thirty day voting period that immediately followed, accusations of abuse of process at the National Body level once again abounded. And once again, accounts of what actually happened varied widely. Here is a sample, posted by Geir Isene at his blog on March 30. In it, he gives his version of what happened when the appropriate committee met in Norway to decide whether or not to change its vote on OOXML:
March 28th: Meeting in the Norwegian Standards Institute (Standard Norge).
Purpose: To decide the final vote for Norway on whether the document format OOXML should become an international standard.
The meeting: 27 people in the room, 4 of which were administrative staff from Standard Norge.
The outcome: Of the 24 members attending, 19 disapproved, 5 approved.
The result: The administrative staff decided that Norway wants to approve OOXML as an ISO standard.
Their justification: "Standard Norge puts emphasis on that if this [OOXML] becomes an ISO/IEC standard, it will be improved to better accommodate the users’ needs."
This translates to: "Yes, we know the standard is broken, 79% of our technical committee have told us. But we hope that it someday will be repaired by someone. And we’ll be happy to help if someone can give us the resources."
Alright, the Norwegian Standards Institute is moving away from adopting quality standards to promoting a repair shop philosophy.
Needless to say, such accounts do not inspire confidence in those that must live with the decisions made by those with the authority to make them. Nor did the news on March 31 that the Chairman of the same committee had just sent a formal protest to ISO, that included the following language:
Because of this irregularity, a call has been made for an investigation by the Norwegian Ministry of Trade and Industry with a view to changing the vote.
I hereby request that the Norwegian decision be suspended pending the results of this investigation.
The denouement of this ongoing drama is that at the end of the thirty day voting period, a sufficient number of National Bodies — including Norway — appear to have changed their votes to secure the final adoption of OOXML (the formal announcement may not be made until just after this issue is delivered). [Which was confirmed, of course, only hours later]
Without assigning blame to either the proponents or opponents of OOXML, two questions that demand answers must be posed: Is this any way to conduct the process whereby the global standards upon which governments and society rely are developed and adopted? And if not, what will be done about it?
It is impossible to avoid the conclusion that the credibility and integrity of the formal standards development process has suffered serious damage as a result of what has just transpired. While that process may serve perfectly well under less contentious circumstances, reforms are obviously needed to address those exceptional circumstances in which greater protections are needed.
In order for the credibility of the traditional system to be restored, a thorough review of the just completed DIS 29500 Fast Track process should be immediately commissioned. That review should include recommendations for reform that would include, but not be limited to, suggesting revisions to the rules relating to Fast Track and PAS submissions, new National Body and ISO/IEC JTC1 rules relating to transparency and conflicts of interest, and providing for circuit breakers and corrective actions that could be invoked the next time such a process has clearly run off the rails.
For further blog entries on here , click
Text of ISO confirmatory press release (biolerplate at end eliminated):
ISO/IEC DIS 29500 receives necessary votes for approval as an International Standard
ISO/IEC DIS 29500, Information technology – Office Open XML file formats, has received the necessary number of votes for approval as an ISO/IEC International Standard.
Approval required at least 2/3 (i.e. 66.66 %) of the votes cast by national bodies participating in the joint technical committee ISO/IEC JTC 1, Information technology, to be positive; and no more than 1/4 (i.e. 25 %) of the total number of ISO/IEC national body votes cast to be negative. These criteria have now been met with 75 % of the JTC 1 participating member votes cast positive and 14 % of the total of national member body votes cast negative.
The 30-day period during which ISO/IEC national bodies had the opportunity to reconsider their votes on the draft ISO/IEC DIS 29500 closed at midnight on Saturday, 29 March 2008, with the result that the criteria for approval of the document as an ISO/IEC International Standard have now been met.
ISO/IEC DIS 29500 was originally disapproved in the "fast-track vote"which ended in September 2007, when 3 500 comments were received. However, under the rules of ISO/IEC JTC 1, the DIS vote was followed by a ballot resolution meeting (BRM) at which the comments were addressed. After the meeting, the ISO/IEC national bodies had 30 days to modify their votes if they wished.
The BRM was held in Geneva during the week 25-29 February 2008. By eliminating redundancies, the comments had been reduced to just over 1 000 individual issues to be considered. Issues considered as priorities by national members (such as accessibility, date formats, conformance issues) were discussed, and the other comments were addressed through a voting process on the remaining items, a system agreed by the BRM participants.
The issues addressed and revised have resulted in sufficient national bodies withdrawing their earlier disapproval votes, or transforming them into positive votes, so that the criteria for approval of the document as an International Standard have now been met. Subject to there being no formal appeals from ISO/IEC national bodies in the next two months, the International Standard will accordingly proceed to publication.
ISO/IEC 29500 is a standard for word-processing documents, presentations and spreadsheets that is intended to be implemented by multiple applications on multiple platforms. According to the submitters of the document, one of its objectives is to ensure the long-term preservation of documents created over the last two decades using programmes that are becoming incompatible with continuing advances in the field of information technology.
ISO/IEC DIS 29500 was originally developed as the Office Open XML Specification by Microsoft Corporation which submitted it to Ecma International, an information technology industry association, for transposing into an ECMA standard. Following a process in which other IT industry players participated, Ecma International subsequently published the document as ECMA standard 376.
Ecma International then submitted the standard in December 2006 to ISO/IEC JTC 1, with whom it has category A liaison status, for adoption as an International Standard under the JTC 1 "fast track" procedure. This allows a standard developed within the IT industry to be presented to JTC 1 as a draft international standard (DIS) that can be adopted after a process of review and balloting. This process has now been concluded with the end of the 30-day period following the ballot resolution meeting.
The process was open to the IEC and ISO national member bodies from 104 countries, including 41 that are participating members of the joint ISO/IEC JTC 1.
Andy, I’m not much of an optimist as far as human nature is concerned (inciting friends to dub me the proverbial curmudgeon), but I do believe some people and organisations would go through great lengths to protect their reputation by not allowing the fruits of their labours to be (ab)used for the wrong purposes.
Let’s assume the worst: the ISO does NOT evaluate what has just happened, therefore nothing regarding her processes will be changed. I think it goes without saying the reputation of ISO in this particular field would become quite tarnished (taking a nose-dive would be another way of putting it), if it pans out that way. Now I can imagine some (consortia of) organisations that offered proposals of standards in the past and are still involved in maintenance of those standards may wish to retract them, in order to avoid becoming collateral damage of an ISO not capable of evaluating its own processes (is ISO certified according to the ISO 9000 series, btw? If so, who audits them?).
What recourse would those (consortia of) organisations have? Is it possible at all to retract a standard from the ISO, or would one need to fork the standard?
Moreover, is it possible for organisations outside the ISO with a vested interest in OOXML meeting certain quality levels (like the W3C controlling XML, that language getting perverted due to OOXMLs ‘extensions’ according to http://www.robweir.com/blog/2008/03/ooxmls-out-of-control-characters.html) to lodge a protest with the ISO?
PS I’m not a native English speaker, so if I don’t make myself clear enough don’t hesitate to ask what I mean!
Here’s a thought in answer to the previous post.
Microsoft has now opened the flood doors on how to game the ISO process, so you can expect more abuse by other companies in the future.
That being said, most companies *do not* game the system, either in ISO or the ECMA. The credibilities of *those* standards are at stake, so we already have some allies. Call them the "allies of quality standards".
So what can be done? I propose a three pronged attack:
a) Use the ISO process to launch an appeal to suspend the OOXML "standard" and either reject it or allow a proper re-vote with proper supervision from our "allies of quality standards" (sort of like UN supervision of the validity of a country’s election).
b) Together with "allies of quality standards", force ISO to adopt standards of standardization. ISO is a quality standards organization. Why on earth don’t they have a quality auditor on the whole process? Every company I’ve been to has one to ensure that keep our ISO/CMMI/whatever creditation, so why not ISO itself?
c) Once (b) is in place, it should be possible for our "allies of quality standards" to force an audit of OOXML which would bring it back into stategy (a).
d) Whether or not ISO complies, ECMA and a few other orgs have been bought too, so we have "allies of quality standards" in those fields too.
So there’s an opportunity for a separate organization, call it "The auditor general of standards" to audit *all* standards from all standards organizations.
If we can get governments to require both a standard and clearance from "The auditor general of standards", OOXML and other fake standards would be useless.
e) Once (d) happens, ISO and other orgs can be force to go to do (b) so that they always pass the "The auditor general of standards". That is, of course, if they haven’t already done so.
Unfortunately, I don’t have the pull to do any of these things, but there are people who can and I’m hoping they’re listening.
The point is, we shouldn’t become cynical and just say "Well, there’s nothing we can do. That’s just the way things are". That’s BS.
That’s not the way things were as many in the field for 30 years can testify, and that’s not the way things things should or will be in the future.
We’re not talking about just a document format. We’re talking about standards that affect every aspect of your life. Lives are at stake, and even if people get complacent and let things slide, there will be a backlash. After all, ISO wouldn’t even exist if it didn’t have a reason. The main reason for acting now is that if no-one does, it will get worse before it gets better and will be paid in lives (either actual lives or people’s quality of life). This has to be nipped in the bud.
— Robert Devi
Unfortunately, I’m not an expert at the detail level regarding the questions you are asking, so I can only address some of them, and at a high level.
First, as to the reputation of existing standards: i don’t think that already adopted standards will be tainted, except to the extent people start revealing how this or that prior standard also had an unhappy and unduly influenced path through approval. That’s the exception rather than the rule in any event. Future standards might be more affected, as now that people have seen one process get politicized, they may want to know more about how things went before they decide how they feel about a given new standard.
As to whether a standard can be forked: that depends on the agreement entered into between the originating standards body and ISO, and there is a certain amount of leeway in how that agreement is struck (for example, some standards – like ODF – are available as free downloads even at ISO (which is not the norm) although if you want to order a paper copy, you would still have to pay). That said, one does not give awaay all rights when one submits to ISO, and I believe that forking (with both ISO and the originator using different names for their work product) is possible, altnough not desirable for the usual reasons. Unlike open source, forking of a standard where the two versions differ as to old elements of the specification, rather than simply extensions of it, automatically negativelly impacts the value of having a standard at all as it would split apart an existing ecosystem of already compliant products, while separate open source projects, since they are code rather than specifications, could each head in different useful directions without necessarily dividing and confusing the market.
Regarding appeals: there are specific rules laid down for appeals (there is a recent blog entry on this at Groklaw), but I have not had time to look at them in detail to see (for example) who would be acknowledged by ISO as the source of an appeal.
The appeals process is going to give us a fast read on what’s going on.
I say this because the ISO Secretariat has a perfectly solid excuse to avoid the whole matter: they only accept appeals from the authorized and registered agents of the NBs themselves — and those are the people who filed the votes notionally being appealed. In other words, the only party with standing to appeal the vote of Standards Norway is the person who filed it in the first place. The wording on the appeals process can certainly be read this way.
Infernal snowballs, anyone?
On the (IMHO very remote) chance that ISO accepts the appeal, we can read this as the ISO Directorate starting to look for a way out of this mess.
Why am I not surprised?
Overshoot, can one NB appeal the vote of another? If not, then one legal argument would be that the reading you suggest (only an NB can appeal its own vote) would become nonsense, and a court would be likely to assume that this was therefore not the intended meaning.
Of course, if ISO said that only the registered NB could appeal its own vote, then someone would have to sue ISO first, and that puts a huge burden on anyone who feels like an appeal is justified.
Overshoot, can one NB appeal the vote of another?
I think that the answer there is along the lines of "apparently yes, if they have a death wish." Sort of like lawyers attacking each other in court, it’s theoretically possible but the long-term consequences for the attacker are not attractive.
Mostly I suspect that the appeals process is intended to deal with issues such as Cuba’s vote misregistration and maybe the voting irregularities in the BRM.
Here’s a question: Does (did) the ISO Secretariat have the ability to discard this vote result on account of the obvious deficiencies and irregularities (not to say blatant corruption, vote-buying and deck-stacking) that have taken place? Or are they required to honor the ballot result unless/until an NB files a formal appeal? This fact should give us something of a read on whether to expect any sensible response from them. (I shan’t use the terms "corrupt," "bought off," or "hopwelessly inept" with regard to the ISO Secretariat, as I don’t know whether they had the discretion to act sensibly or not.
The more critical question…what are the legal steps involved in evaluating a appeal?
Suppose that a NB appeal on the grounds that DIS29500 can not be accepted as standard since the editor could not finish the text before the 30 day deadline after the BRM. Must the ISO Secretariat in that situation explain why they ignore the appeal if they chose to ignore it?
I suppose that they must explain themselves if the matter is taken to the court…but taking the matter to a court is not your average thing.
It’s rather amusing to see comments here asking how other companies can repeal their standards to avoid the ISO taint.
I think it is Microsoft who likely will want to avail themselves of the repeal process.
Microsoft will want to approve all their products as standards, and repeal all software standards that are not Microsoft products.
Not necessarily in that order.
Now that Microsoft has all its people in place, its work can move to the next phase.
All Microsoft products can be approved as standards.
All software standards that are not Microsoft products can be repealed.
Not necessarily in that order.
Apologies if already asked and answered… Does anyone know of any past appeals of fast track or other "approved" "standards"?