Microsoft Office Open XML Fails to Win ISO Vote

by Kurt Cagle

In the first few days of September, just as kids were beginning to head back to school, something rather remarkable happened: Microsoft lost its hegemony. In a vote by the various members of the ISO standards committee, managing only to come up 53% of the votes it needed to fast track the Office Open XML format for consideration as an ISO standard.

Now, Microsoft and its partisans will, most likely spin this as a temporary defeat, pointing to the number of No with comments votes that indicate that it may be possible, with some extensive work, to make OOXML workable in the near term future if some rough edges are smoothed out. In the end of the day Microsoft will have won, and those countries that were too narrow-minded focused on insuring that the standard that emerged was close to being workable with efforts will in the end change their minds, once the spec is cleaned up (and perhaps a few more people were added to the appropriate committees with positions financed by Microsoft money) that surely the OOXML standard will in fact become the de facto one very soon now.


Simon Hibbs
2007-09-05 01:14:42
Interesting post (long though).

Word became dominant through merit, not leverage. At the time WordPerfect was absolutely dominant and there were several other word processors that competed head to head but Microsoft realy got the GUI thanks to their years of Mac development. They won because theirs was hads down the best GUI word processor. The same for Excel which also beat out a dominant incumbent and viable competitors by simply being better. Bundling them into the Office Suite was the final marketing coup de grace for the competition, but let's be fair they won on merit and to this day there are no word processor, spreadsheet or desktop databases that surpass the Office versions at what they do best.

Re. your commment about software vendors making more money being ok. Yes, but open formats create a much more valuable ecosystem than any one company can exploit. Closed formats may make MS more money, but open formats create opportunities for document processing ecosystems that create far more total economic value. Monopolies - which is what dominant closed formats create - restrict freedom and therefore reduce opportunioties for economic activity.

Asbjørn Ulsberg
2007-09-05 03:21:48
What an excellent summary of the whole OOXML kerfluffle! I think your analysis is correct and that this really does draw a line in the sand. It will indeed be very interesting to see how Microsoft choses to move forward from here and how the uptake of ODF will be increasing over the next months and years.
2007-09-05 03:56:00
Simon, I disagree. WordPerfect 7 was the most usable word processor I've ever used, up to and including Word 2003. (Haven't used 2007 enough to have an opinion.) WP was killed more by dirty tricks (such as suddenly changing the APIs in Win95 just before the release date) and bundled software than anything else. Microsofts complete success against WP is the root of the many other questionable deeds over the years.

As for Office Not-so-open XML, I hope that this partial defeat will motivate Microsoft will sit down with OASIS and China and create a totally open, universal, internationalized, extensible format (hopefully based on ODF 1.2) which reuses a number of existing standards and which is adequately specified and no more. It has to be IP-safe for all implementers and users, so that ten or twenty years from now, someone finding an old document can implement tools to read and display it. (Assuming that the software vendors will have moved on to something even better by that point.)

Simon Hibbs
2007-09-05 04:19:21
I didn't even know there was a WordPerfect 7, they'd become so irrelevent by then that it's not funny. For a company completely oriented around producing printed pages, to completely miss the GUI revolution and produce such a horrible mess of an application as WP for Windows was just inexcusable. They killed themselves through incompetence. MS did play some dirty tricks, but in the end that's realy just excuses.

MS won't base anything on ODF and I don't think they need to. If the radicaly different architecture of their spec genuinely reflects the needs of their application I think it's fair for them to get it standardised. Of course they need to produce a much better thought out, cleaner and more 'implementable' format. I honestly don't care which standard becomes dominant, as long as it's of high quality. If ODF just forces MS to work to a high quality, truly open standard of their own devising which then dominates, that's fine by me.

M. David Peterson
2007-09-05 07:27:23

This was a known outcome, and one that should be well understood for what it is: An opportunity for the voting members to speak up as to what they want fixed before the final vote in five months. Why are you playing it off as if something much more grandiose has taken place?

Kurt Cagle
2007-09-05 10:26:21

Word became dominant in great part because of the power of Windows; having used Word in its pre-windows incarnation, I found it cumbersome, awkward, and in general a far inferior product to Word Perfect in nearly every way (and at the time I was actually a big Microsoft fan).

Word for Windows was revolutionary, with Microsoft taking advantage of an in-depth understand of their operating system and the ability to prepare products for new versions of Windows before competitors could. It was also a logical buy for enterprises who were upgrading their Windows version, a position that significantly hobbled WordPerfect. The switch to a dedicated GUI for Windows put WordPerfect into a tailspin from which it never really recovered, because they could never get to market first when the platform they were working against was controlled by Microsoft.

This is not dismissing the sophistication of Word and Office in general. Microsoft has had nearly twenty years to perfect it, to more closely attune it to their system, and to build up the corresponding infrastructure around it. It's becoming highly specialized to its environment. The problem of course is that when you shift it out of its native environment, Office is considerably less impressive as an application, as the Mac version amply illustrates- the mean time in getting Mac Office products out is actually growing as the differences between the two platforms become more pronounced, and many Mac people find that its metaphors feel clunky and awkward when expressed in the Apple paradigm.

Personally, I think the biggest mistake that Microsoft made was in not choosing to create MS Office for UNIX. I say this not because I see the Linux market as shooting up radically (I don't), but because I think it would have forced the designers and developers there to think about the problem of writing a real world application that didn't have the MS core capabilities underneath it.

I absolutely agree with your assertion that open standards create a more valuable ecosystem. I would just question whether OOXML really is an open standard.

Kurt Cagle
2007-09-05 10:45:03
M. David,

I don't think this was a foregone conclusion. It is not the end of the ISO process, but it's a lot like a parliamentary defeat that opens up a vote of no confidence. Microsoft may in fact be able to pull a rabbit out of the hat here, but I don't personally think they will, in great part because their efforts will be scrutinized much more closely now. ODF went through the process pretty much unscathed; the fact that MS did not now means that even delegates who may have been mildly supportive are now going to be considerably less inclined to give MS the benefit of the doubt this time around, because it has gone from being a technical decision to being a political one.

The way I see it, this vote was disastrous for Microsoft. The problems inherent with the specifications are not cosmetic ones, or they would have been resolved by now - they are deeply technical and intrinsic to the product itself. At a minimum, this necessitates that Microsoft change not just their spec but also the underlying product, and this in turn pushes the time to get this spec out from months to years, at a cost of losing more and more potential contracts to governments going toward open standards.

If this vote had gone through, the vote in five months would have been anticlimactic - it would have passed easily. It didn't, though, and in failing there's now blood in the water. It is still possible for Microsoft to make this happen, but its now far from certain that it will, and the hubristic attitude that Microsoft has brought to the table will need to be tempered considerably before that happens.

M. David Peterson
2007-09-05 11:56:58

Every indication that I have seen from folks such as Rick Jelliffe has been one of "expect the September vote to contain a lot of "No with comments", and that it was the content of those comments that mattered most. In other words if the content of those comments were items that were impossible to implement then No meant no, but if they were possible then No means yes.

In this regard, has a general consensus been gathered as to whether or not the "No with comments" are in any way realistically feasible?

M. David Peterson
2007-09-05 11:59:57
Just to provide context, @

An vote by a national body of “No with comments” is a “Maybe”, and not an absolute “No”. Looking at it more, I wouldn’t now go as far as Job Bosak’s comment that “No with comments” is the same as “Conditional approval”, however. What really matters is the particular comments: if they are doable or reasonable and inline with goals of the standard and the proposer’s conception of the standard, (and if no-one’s hair is on fire) then No means Yes. But if the comments are undoable or unreasonable or out-of-scope for the standard’s goals or depart from what is acceptable to the proposer, the No means No.

2007-09-05 17:03:53
It's feasible. It will take a lot of work to dig out details of code possibly unvisited in some time.

The real test will be what happens if the review comments are settled and the business competitors are faced with passage after the resolution of comments. Then we will know what open and fair mean to whom. In other words, who actually respects process?

I repeat what I've said elsewhere: the problem of unleashing attack dogs is they tend to circle back to the hands that held the leash.

2007-09-06 13:33:57
The thing is, a LOT of the problems identified in this round were known for a long time. They have been repeatedly pointed out to MS and to the ECMA TC putting together the spec. They for the most part ignored those comments, believing in their hubris that they could jam the spec through the process. This vote clearly said they cannot.

I see two choices for MS here: go back to the table and adapt a totally different tack (e.g. be willing to compromise, actually address the comments, etc.) or risk an even worse vote in the next round.

And those that argue the resistance to OOXML can be attributed to IBM are delusional. To understand why people were upset about OOXML, just imagine what would happen if a vendor submitted a proposal to ISO for a new web document language; one that had little relation to HTML, which had its own presentation system (no CSS), etc.

2007-09-06 15:01:53
Isn't that what Flash is? Isn't that what a Second Life client is? Is it bad for IBM to be financially supporting Linden Labs while there is an ISO standard for 3D on the web? Is it bad for the Web3DC to support both X3D and Collada?

MS has multiple options, but the ISO process is what one should be paying attention to. At this point, comment resolution is the next step. Two things are important:

1. Follow the process cleanly and thoroughly.
2. Don't try to add extra burdens beyond what is ISO policy for this process.

The hysteria and attacks just make it harder for the serious people to do their jobs and that will ultimately make it hard for any other standard to be submitted or improved much less adopted and implemented.

Be professional and clear (even if it hurts).

Rick Jelliffe
2007-09-07 03:46:53
Bruce: Once the draft was put into play at ISO, ISO's rules prevent the draft from being altered. There is only one place in the process that changes can be made: at the end as the result of the Ballot Resolution Meeting, which has always been expected.

So talk of Ecma "ignoring" comments as if they could go ahead and create a new draft mid-way through the process is either ill-informed or malicious.

Rick Jelliffe
2007-09-07 20:56:38
Kurt: For a realistic view on the process, see Alex Brown's blog.

On the issue that there has only been limited review, the fast-track will have taken 13 months, with three formal opportunities for more than 50 independent national bodies to make various kinds of comments, with over 2200 people actively involved (for example in committees) and with tens of thousands of other people interested. (How many did more than cut-and-paste is another matter.) "Limited" seems an inappropriate word.

Finally you casually mention people being added to committees because of Microsoft money as if everyone knows this is what happens. You are aware that there is actually no known example of this? (AFAIK MS rectified the Swedish mix-up before any harm occurred.)

2007-09-08 13:56:04
See my comments:

As for their future direction, I think any large-scale change will have to wait until Gates and Ballmer are completely out of the picture. I don't expect that anyone else has enough "pull" inside the company.

Kurt Cagle
2007-09-08 23:05:26

I am trying to understand the issues involved from the standpoint of a non-participant, and thus am heavily reliant upon the reportage of the various activities in order to try to piece together some semblance of what's going on.

My opinion that I do not feel that OOXML should be an ISO standard is my own, just as I feel that there are other ISO standards which should likely never have made it that far (PDF and Flash come to mind, and even ODF could conceivably fall into that category, though I think that the OASIS process helped to remove some of the obvious corporate taint).

What I have heard is that there are a large number of organizations, most of whom are Microsoft business partners, that have shifted from O to P status or joined the ISO process in the last year. It is this, rather than the Swedish vote snafu (which sounds to me like an inadvertent mistake rather than a deliberate ploy) that I think has raised concerns that a lot of people have about this process. Am I mistaken in this (if so, please let me know, as I'd prefer not to be spreading misinformation)?

Note that the upgrading from O to P for many of these companies is certainly LEGAL within the ISO charter, and certainly any of those companies may be joining or upgrading legitimately because they feel there may be an advantage to seeing OOXML as an ISO standard to them, but it smacks of stacking the vote when you see an obvious bias in that conversion.

Again reiterating, I'm not part of this process, and I don't have hard numbers of how many of those O to P conversions were Microsoft allies vs. the supposed anti-Microsoft crowd. This has become something of a hot button issue, and it may very well turn out that the number of anti-OOXML organizations balance things out - again, information that is hard to pick up in all of the rhetoric. However, there's enough of a question here about the propriety of the process as to at least raise the spectre that OOXML is not being promoted because of merit but because of external influence - fiscal, marketing agreements, whatever - on the part of Microsoft.

Rick, I understand that you are trying to be fair and rational about this, and that you have come to the belief that OOXML should in fact be an ISO standard. Given your superior experience with standards and the standardization process, I have to respect your position even if I do not agree with it. OOXML has been standardized through the ECMA process.

I'd personally be a lot more comfortable with OOXML if it had gone through OASIS, which I believe has in general a considerably more stringent set of requirements with regard to standardization, before it had been pushed up to ISO. Not only would many of the issues that plague OOXML now have been raised and resolved prior to submission had it gone that route, but it would have also provided a basic imprimature on the standard that would readily have satisfied open standards advocates - that the process for developing the standard was open to public review, that there are no legal liabilities that may arise due to the adoption of such a standard, and that there are other entities that participated in the process that hold no direct loyalty to Microsoft and would, as a matter of course, have provided a "devil's advocate" viewpoint for improving the specification itself.

As with Microsoft's commitment to "open source", the fact that they pushed a proprietary format expressed in XML through a friendly standards body with comparatively minimal review and then pushed to fast-track that standard raises serious questions about the degree to which this "open standard" is anything of the sort.

This is not an academic distinction - as more and more governmental organizations embrace an open standards ethic in procurement contracts, the "open standards" label has considerable economic clout. It is in the obvious best interest for Microsoft to gain that label - the question is whether it is deserved, and indeed whether granting this to a company that has gone out of its way to bypass as much formal review as possible may in fact be subverting the value of the label.

Those who would push to see OOXML become an ISO standard obviously feel it does not, those who don't who oppose it may just as fervently believe that such criteria do matter. Microsoft has submitted OOXML and as a member organization this does means that it must be acted upon, but that does not automatically grant to the specification the automatic privilege of becoming a standard. That's why I tend to wonder if the (unexpected) vote last week may in fact have far more significance than many pro-OOXML supporters prefer to acknowledge.

Rick Jelliffe
2007-09-10 03:27:14
Kurt: Yes, it may already be having an effect. Ken Holman's latest news email to SC34 (Canada provides the Secretariat until next year, which is him acting pretty much for free) comments on a large number of new national bodies who have not responded to a vote (a liaison issue) where there are supposed to.

This would not be new members from a couple of months ago, I guess, so there certainly is the danger that single issue NB voters who don't take the rest of their responsibilities seriously (even by just issuing an abstain vote for everything) could paralyze the SC34 process.

Hopefully it is just early hiccups: I am not sure how long it takes to declare that a NBs membership of an SC has lapsed. (The BRM membership is JTC1 membership not SC34 membership. I suspect that some NBs might find SC34's emphasis on how to allow more people to make more and better custom schemas for their individual requirements might be a little boring if not puzzling compared to the fixed schema world of the pre-fab OASIS and Ecma fast-track specs.)

2007-09-10 06:10:12
OASIS might not be preferred now, Kurt. As I mentioned on your other blog, the first to file patent changes mean we have to look very hard at the organizations that sponsor this work and prefer those that have participation agreements which close the conversations to non-members. ISO will also have to rethink this. It works quite well with the Web3DC and the W3C. I haven't looked at OASIS rules in some time, so I don't know if that comes up to the requirements.

There is no triumph in all of this for open source. Open source is now being used like a brand. Any act against a company dealing open source is being described as an 'attack on open source'. Note Sun's public response to the latest round of patent lawsuits. The big companies hiding behind that apron will go to the well once too often (to mix metaphors). The open source model as a business model that will come into question.

Some semblance of technical rationality may reemerge where project tools are picked for performance in the domain of application. For example, Linux with Open Layers really rocks. ODF is feature-deficient. MS Visual Studio for all its quirks is still the best rapid application development tool in combination with SQL Server. SQL Server standing alone has some peculiar quirks.

And that is the right way to go about picking tools, IMO.