User: Password:
|
|
Subscribe / Log in / New account

OOXML loses a battle

Did you know...?

LWN.net is a subscriber-supported publication; we rely on subscribers to keep the entire operation going. Please help out by buying a subscription and keeping LWN on the net.

By Jake Edge
September 5, 2007

The latest round in the battle for office document formats has gone against Microsoft's Office Open XML (OOXML) submission. It certainly won't be the last we hear about it, as there is another vote in February, but it does, at least, slow down the fast-track proposal for making the format an international standard. The process has been anything but regular, with allegations of ballot box stuffing in Sweden and last minute voting class changes by eleven countries. These kinds of shenanigans do very little to enhance the reputation of the International Organization for Standardization (ISO) nor do they promote confidence in their standards.

The vote, which closed 2 September, was made by members of the Joint Technical Committee, Information Technology (JTC1). Each country which is a member of ISO and wishes to join, can be either a Participating ("P") or Observing ("O") member of the committee. In order to pass, the proposal must get two-thirds support of the P members and no more than one-quarter "no" votes amongst both P and O members. In both cases, abstentions are removed before calculating the ratios.

The results, announced on 4 September, were 53% "yes" votes by P members and 74% "yes" votes by P and O members, which fails both tests, though either failing is all that is required to defeat the measure. Many of the votes, on both sides, were made "with comments". The comments specify portions of the OOXML spec that need clarification or change before it can be ratified.

Those comments will be passed on to the Ecma International, sponsor of the OOXML standardization proposal, to propose resolutions to the comments. Ecma is also the organization that rubber-stamped OOXML as a standard last year. They have until mid-January 2008 to submit the proposed changes and the committee members have until the Ballot Resolution Meeting (BRM) in late February to review and discuss them. Microsoft's Brian Jones estimated there would be in the neighborhood of 10,000 comments, many probably duplicates. How a committee is supposed to analyze and handle that many resolutions, in a week-long meeting, is unclear.

If enough of the "no" votes are satisfied that their comments have been addressed at the meeting, they can change their votes to yes. If that vote takes place, it will be the P members in February who get to vote. It is quite possible that, similar to the run-up to this vote, O members will suddenly decide to switch to P members. At this point, OOXML proponents know roughly how many votes they need.

Andy Updegrove has been following the approval process closely in his Standards Blog and reported on eleven countries upgrading from O to P status in the two weeks before the voting closed. Whether Microsoft is behind this sudden interest by these countries can only be speculated upon, but nine of them voted yes, one no, and one abstained. Regardless of why they felt the need to jump into the voting at the last minute, it certainly seems fishy.

If the vote fails at the BRM, we still may not have seen the last of OOXML as a proposed international standard. All of this effort has been to "fast-track" the proposal. Ecma and Microsoft can still submit it for approval under the regular, lengthier, standardization process. That could easily take several years, which is why there is a big push to fast-track it.

OOXML is a complicated, 6000+ page specification, requiring a great deal of study and consideration before a sensible decision on it can be made. By upgrading at the last minute, it certainly appears that some of that review may have been skipped. If a country was interested in the process and wanted to have more input, it seems that they might have found time to do it in the nine months of review. If it is going to be an international standard, it should, at least, be a well scrutinized standard.

Predictably, Microsoft is proclaiming the voting result as a victory, of sorts, just a step along the way to ISO acceptance. In the Microsoft view, the no voters will reverse course "once their comments are resolved." Their confidence is palpable and, to opponents, galling. There is some indication that the pressure applied to national bodies resulted in a backlash, with at least one switching to a no vote because of it. It will certainly be interesting to see how some of the comments will be "resolved."

Microsoft has admitted that an employee offered "marketing assistance" to offset the $2500 entrance fee for Swedish companies to join their national voting committee. More than twenty showed up, just before the vote, to vote yes. Eventually, the vote was thrown out, not because of the blatant ballot-box stuffing, but because somehow one company voted twice. Sweden ended up abstaining, which was a win for OOXML, as it clearly would have been a no vote otherwise.

Microsoft has made various noises about the "inadequacies" of the Open Document Format (ODF) standard – ISO passed it with so few comments that a BRM was not required – and there is some truth buried deeply in the rhetoric. The proper response is not to propose another standard, but to improve the one that exists. ODF is implemented by multiple projects, with open source reference implementations. It is very unlikely that anyone, other than Microsoft, will be able to fully implement OOXML.

It's also not clear that anyone should want to implement OOXML as international standard. Besides being complex, the proposed standard contradicts other ISO standards. It also has the kind of bug-for-bug compatibility that is one of Microsoft's calling cards. An international standard should not have to implement a sloppy collection of bugs and compatibility hacks. It should be noted that OOXML contains some very important features – some not available in ODF – but that does not make it a good standard. It should not be adopted just to appease the world's largest software maker.

Microsoft is behaving like a company that is terrified of losing their near-monopoly in the office software market. If they, instead, embraced the standard – leaving behind extend and extinguish – and competed on the feature set of their office suite, their much touted "innovation" could shine. Unfortunately, for anyone with a historical perspective on Microsoft's tactics, this OOXML standardization move looks like the first act of some kind of customer lock-in scheme.

There will be close scrutiny on the participants between now and the vote in February. Hopefully, we will see no more gaming of the standards process, by anyone; the committee will judge the resolutions on their technical merit, coming to a sensible decision. From what we have seen so far, that seems unlikely, but one can hope.


(Log in to post comments)

OOXML loses a battle

Posted Sep 6, 2007 7:28 UTC (Thu) by ekj (guest, #1524) [Link]

Actually, if all the criticisms where fairly resolved, I wouldn't have much problem with the standard being accepted. There's no snowballs chance in hell of that though.

The proposal is 6000 pages long, there are over 10.000 comments to it. Many of them duplicate, but even so sorting trough and adressing these, even minimally, will take *much* longer than from now until february. Inside of a one-week meeting is beyond laughable, that's literally not even enough time to read the proposal once.

(Well, if you can read and comprehend technical standards at a pace of 50 pages an hour, 7 days a week, 17 hours a day, then you'll have managed reading it, you won't even have started *correcting* it or even *glancing* at any of the comments though)

I predict the following happens:

A few smaller issues are resolved.

Another dozen countries upgrade their membership to P

A handful of high-profile P-members change their vote from no to yes.

The proposal is accepted as a "standard".

OOXML loses a battle

Posted Sep 6, 2007 11:33 UTC (Thu) by smitty_one_each (subscriber, #28989) [Link]

>The proposal is accepted as a "standard".

And so, what? Who has the time, interest, or resources to deal with such a Melvillian disaster?

OOXML loses a battle

Posted Sep 7, 2007 0:48 UTC (Fri) by brouhaha (subscriber, #1698) [Link]

And so, what?
The "so what" is that if ISO adopts it as a standard, when citizens of various states and countries demand that their governments use open standards for government documents, Microsoft will be able to say that their software does so, even though it doesn't actually provide most of the benefits an open standard should provide.

I wouldn't have any problem with OOXML if it was a reasonable standard, but it's not. It's just a way of translating their old binary blob data format into nearly equally impenetrable XML. Many reviewers have said that the spec is full of ambiguities. As noted in the article, the chance of anyone else really supporting OOXML is nil.

There's no reason that an open, standardized document format needs to support every random, bizarre feature that Microsoft has invented over the last 25 years. It should be quite possible and manageable for a working group to define a set of extensions to ODF to support the things that Microsoft has complained are missing from it. (Most of them, anyhow; it's not clear that all of the things Microsoft wants are reasonable.)

What's the importance of the ISO standard?

Posted Sep 6, 2007 9:38 UTC (Thu) by jschrod (subscriber, #1646) [Link]

Yes, it is quite possible that OOXML gets to be an ISO standard.

But IMHO the importance of ISO standards in general is over-emphasized, and the anxiety about ISO's good reputation is a shenanigan -- as everyone can attest, who remembers the struggle between the ISO/OSI network model and IP network architecture. Don't you remember that there's a reason for the proverb »the nice thing about standards is that there are so many to choose from«?

We should tell it like it is: MS's Office formats already are de-facto standards in business, and also in government in many places. People try to change the latter with the goal to get rid of proprietary file formats. (They hope that companies will follow suit, especially when they do business with government.) Rubber-stamping OOXML as an ISO standard hinders that goal.

That's all. There are no other real disadvantages -- business doesn't care if Office is an ISO standard, they are already satisfied with the available de-facto standard as it is. If you do business beyond the restricted IT realm, you know how ubiquitous Office is. Even more so in small companies than in big companies. Staff from big companies often knows that there are different file formats -- employees from small companies take it simply for granted that everybody has Office. I have received emails from potential customers in doc format, and just in doc format, without plain text; and after an inquiry from me they complained »what? you don't want work from me? I'm the customer and I do as I want!« (paraphrased).

Joachim

POSIX

Posted Sep 6, 2007 17:38 UTC (Thu) by cate (subscriber, #1359) [Link]

IIRC US required some standards to buy software, and it was for this reason that windows is POSIX compatible (ok, it has the interface, but to much bugs to be useful). But why they don't required standard for office documents?

BTW the new POSIX will be very similar to SuS, so nearly impossible to have POSIX on any windows version. This means that US will no more buy windows OS?

POSIX in Windows

Posted Sep 11, 2007 0:16 UTC (Tue) by pr1268 (subscriber, #24648) [Link]

Don't count on it.

<snide comment>Knowing Microsoft, they'll find a way to slap on a POSIX-compliance layer to add to the cruft of whatever version of Windows is current...</snide comment>

...Which is pretty much what they did to get existing/prior version of Windows into POSIX compliance.

"marketing assistance" offered in Sweden

Posted Sep 10, 2007 9:40 UTC (Mon) by tyhik (guest, #14747) [Link]

It should be noted that corruption is one of the lowest of the world in Sweden. Therefore, it is telling that "marketing assistance" was offered, I'd say EVEN, in such an "assistance-unfriendly" country.

OOXML loses a battle

Posted Sep 11, 2007 18:31 UTC (Tue) by hazelsct (guest, #3659) [Link]

Thanks! Just put this ISO rejection in my support statement for Massachusetts to adopt OpenDocument and not OOXML at http://devalpatrick.com/issue/opendocument .

OOXML loses a battle

Posted Sep 13, 2007 14:47 UTC (Thu) by obi (guest, #5784) [Link]

To be honest, I'm not so sure OOXML is much worse or better than ODF. Both are more or less a serialization of their respective apps' state. That makes both of them pretty hard to implement by a third party.

f.e. AbiWord is not using ODF as its standard output filter, because it doesn't match their internal representation that well.

I'd prefer to see a more abstract document format that shares as much as possible with existing standards (html, css, rdf, svg) and allows extensions (xml makes this pretty easy) so the vendors don't feel held back by the standard, and common extensions get merged into the standard where it makes sense - a bit like how the OpenGL ARB works.

However, back in the real world this would just create yet another standard that nobody implements. So I guess we're stuck with ODF and OOXML.

OOXML loses a battle

Posted Sep 14, 2007 11:26 UTC (Fri) by andrejp (guest, #47396) [Link]

>To be honest, I'm not so sure OOXML is much worse or better than ODF. Both are more or less a serialization of their respective apps' state. That makes both of them pretty hard to implement by a third party.

Pff. That has got to be the most asinine thing regarding document formats I've read in quite a while.

The document format should be the representation and description of *document structure*, not application's state. That's exactly what makes OOXML largely unimplementable by third parties: unlike ODF, which represents *document structure*, OOXML represents *application state*. A document structure can quite a bit more easily be equivalently rendered by different applications then application state which is inherently unique to the application (hence it is called 'application state' - and if you want to work on, or with, that state you essentially need to duplicate, or emulate to a sufficient degree, the application).

>f.e. AbiWord is not using ODF as its standard output filter, because it doesn't match their internal representation that well.

Gee... Let's just forget ODF then, standardize AbiWord's internal representation (*application state*), put it into a standard envelope (which is what XML is, and nothing more), prepend a catchy word of the day that seems to matter to *other* people (and cross our fingers that those open-dummies and ignorant decision-makers don't notice the difference between professed and actual) and call it.. well OpenAWXML!

That should make it easier. And standard. Not a standard description of document structure though, but STANDARD APPLICATION, which is exactly what this game is all about (and them poor open-dummies thought it's about interoperability and that our objections have to do with technicalities heh).

Ka-ching!

PS: Prepare grea$e. Just in case it gets stuck and needs some pushing at the last minute.

OOXML loses a battle

Posted Oct 12, 2007 15:37 UTC (Fri) by obi (guest, #5784) [Link]

Not sure what you're getting at - you seem to agree with me that there should be a more abstract format that represents the document structure, instead of all these different apps internal representation, but then go on to claim that ODF is it?

Is it really?

The snippets I'd seen of it made me believe otherwise, but I'm perfectly willing to revise my opinion of it.


Copyright © 2007, Eklektix, Inc.
This article may be redistributed under the terms of the Creative Commons CC BY-SA 4.0 license
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds