An attendee, Lee Welburn, at the meeting yesterday in Australia regarding whether or not to approve OOXML as an ISO standard has sent me his notes. I am publishing them as is, without any editing except for making urls clickable, so keep in mind that these are notes taken live.
I'd say things look grim in Australia, but it's not too late to express yourself to Australia Standards. The public is encouraged to continue to send comments until August 21.
Update: I have obtained the slides that IBM's Rob Weir prepared for the occasion, so you can follow along. Page 5 lists some technical issues that need a satisfying answer.
The main points I get from the notes are:
1. Australia will probably go Yes with Comments, which isn't the same as No with comments. No with comments is what you choose for "conditional approval", if you have problems with issues you think need to be resolved before it should be accepted. Yes with comments is, as one person told the room in Australia, Yes with a Christmas card.
2. Australia Standards is still asking for comments from the public until August 21st, so it's not too late to express yourself if you wish to.
3. Microsoft has a well-organized PR push at these events, with the ever-present Rick Jelliffe and in this case a representative from CompTIA claiming its members want a Yes vote, which will not surprise you. But in this case, CompTIA apparently was called out by the IBM representative, who pointed out that IBM is a member of CompTIA and does *not* approve of OOXML. And if Jelliffe is going to be introduced an an "independent" expert, I suggest asking if he or his company receives money from Microsoft which is used to send him to all these OOXML meetings.
4. Google was there. Sun was not. Microsoft representatives claimed that Microsoft's patent pledge is as good as Sun's on ODF. I would disagree with that heartily, but no one was there to express disagreement from Sun.
5. There were empty seats. Where were you, for that matter? Remember that it's vital to make a record of all the tricks and weaves in this process.
6. Microsoft is pushing the idea that multiple standards are fine. They are not fine for anyone but Microsoft, who can afford whatever happens. But what about the disabled? Do applications now have to be done for both standards for them? Yes. What about Microsoft's competition? The issues are not limited to accessibility vendors and disabled users. All application vendors face similar issues. Who pays for all this extra work and effort? Not Microsoft, I'm sure. Do you see the anticompetitive effect here?
Adding support is a tremendous undertaking Look how long it is taking to get it added to the Mac Office, and they're a Microsoft partner. And adding support for OOXML does not make your product more competitive. It just restores a compatibility you already had, compatibility with Microsoft Office. So the effort spent to add OOXML support will come at the expense of effort that would have otherwise been spent developing new features or otherwise competing with Microsoft Office. The net effect is Microsoft will have the competition spending a couple of years taking a "time out" to restore the level of compatibility with Microsoft Office they already had with the binary formats. This retooling effort becomes a big expense on the competition just to maintain the status quo and gives them no net benefit in the end.
Jelliffe is said to have told the room that OOXML will be the de facto standard no matter what anyone says or does. See, that is precisely the problem. Monopolies do what they do, but should standards be a monopoly enabler? We're not talking standards on DVDs now. We're talking about documents, an item that everyone in the world uses and one that has to be interoperable no matter what operating system you use if there is to be competition in the marketplace. It has to be as easy as sending each other email. We can do that no matter what operating system we use. With documents, it has to be even better, so no features are dropped or distorted. Unless you *want* only one company controlling the world's communications and owning your documents. Is that desirable? I see why Microsoft would want that, but why would you? The end result of two standards in this fact pattern as I see it is that true interoperability will be impossible for anyone who does not use Microsoft products.
7. This was yet another meeting on this issue opened with comments that sound very weird to me. In this case, the room was told that the meeting wasn't for the purpose of discussing technical issues. The technical shortcomings of this proposed standard are the main reason not to vote for it. So who writes the opening speeches, one has to ask?
That's the bottom line to me, actually, that Microsoft refuses to join with ODF and make something the whole world can use to really interoperate; instead it wishes to push its own standard out there to try to destroy ODF. That, to me, is saying that Microsoft doesn't wish to be interoperable. It wishes to own the market. Nothing new there, but since all its PR recently has been about interoperability, I wanted to point out the truth as I see it. Oh, and yesterday at LinuxWorld, Microsoft said no virtualization on Linux of XP or Vista:
Microsoft will not allow Windows Vista or Windows XP to be virtualized on top of Linux, Sam Ramji, the director of Microsoft's open-source software lab, said at the annual LinuxWorld Conference and Expo here Aug. 7.
Update 2: The story has changed, and the headline now reads Microsoft Has No Plans to Optimize Windows for Virtualization on Linux:
Microsoft is not working on optimizing the virtualization of Windows Vista or Windows XP on top of Linux, Sam Ramji, the director of Microsoft's open-source software lab, said at the annual LinuxWorld Conference and Expo here Aug. 7.
While Microsoft does provide customers with licenses to run the two Windows operating systems in a virtual machine regardless of the underlying host operating system, including Linux and the Mac OS, its technical efforts around Windows and Linux interoperability on the desktop are primarily focused on .Net and Java interoperability, he said in an address titled "Linux and Windows Interoperability: On the Metal and on the Wire."
"What we have heard predominantly from customers is that they want to be able to mix and match their data center applications, but we haven't seen significant demand for Linux applications on the desktop or for desktop virtualization on top of Linux," Ramji said.
Microsoft's customers have evidently become confused, then, or have lost their craving for virtualization since Novell and Microsoft entered into their agreement back in November. Either that or it's only in the server space that they care about virtualization. May I please direct you to this article about Novell's SEC filing, which includes this as one announced goal of the two companies' technical collaboration:
* Novell and Microsoft will develop technologies to optimize SLES and Windows running as guests on each other's operating systems.
Here's the 8K so you can verify that is indeed what the technical collaboration agreement says. You can read the joint press release here, which also included a paragraph about virtualization:
“As a result of this collaboration, customers will now be able to run virtualized Linux on Windows or virtualized Windows on Linux,” said Jeff Jaffe, executive vice president and chief technology officer at Novell. “Customers continually ask us how they can consolidate servers with multiple operating systems through virtualization. By working together, Novell and Microsoft enable customers to choose the operating system that best fits their application and business needs.”
And they said something similar in their Joint Letter to the Open Source Community still on Novell's website:
* Microsoft and Novell will collaborate in enhancing and developing the functionality required to efficiently virtualize Windows on Linux and Linux on Windows.
* Both will now be first class citizens in data centers, addressing the needs of mixed environments. They will both enjoy optimized, supported and tuned device drivers to maximize their potential.
So customers demanded virtualization and now they don't? Or Microsoft sees no demand for virtualization on the desktop? How is that even possible? Anyway, that's the latest on this story.[end update]
That is how much they *don't* want to be interoperable, so don't let anyone tell you how much they do or how important it is to reach out to Microsoft in a mixed environment. You can't be pals with sharks. Sharks have no friends, only feeding opportunities.
And that is why, in my view, ODF matters so much, if a person actually does care about interoperability and the right for people to get away from a single vendor monopoly if they so desire. The fact that Microsoft makes it impossible to interoperate 100% even with translators is the reason some will be determined to get away, frankly, and I believe many governmental entities around the world will feel that way. I feel that way. I don't want anyone forcing me to use Microsoft's software if I don't feel like it. And I am sure, 100% sure, that over time, enough people will fall in love with GNU/Linux systems, if only because they are becoming disgusted with Microsoft's overbearing tactics, and that they will choose to escape, no matter what the initial cost.
With that, here are his notes:
Standards Australia Forum on DIS 29500/OOXML, by Lee Welburn
Note the following is from notes and memory as it was impossible to write and keep up with all of it, recording was not allowed but SA have stated that some sort of transcript of the forum will be available in the future. I only a software engineer and am not a member of any standards setting organisation and any errors or omissions are my own.
It was not a full house they stated a maximum of 30 could attend and there was a few seats spare.
The forum was opened by Alistair Tegart of SA.
There was a short discussion about the various ways SA can vote namely
• Approve with comment
• Disapprove with comment
There was some discussion about the point of “Approve with comments” as there is no guarantee that the comments will be considered or implemented, there was a comment that it may as well be “Approve with Christmas card”.
Alistar went on to say that the forum was not about debating technical aspects of OOXML which I thought was a little strange.
First speaker up was Rick Jelliffe(check spelling) Rick was introduced as a XML expert and his association with Microsoft was not mentioned he was put forward as a independent expert.
Rick briefly outlined his background and went to on say that OOXML had created quite a “bun fight” in SC34 . He restated some of what he has posted
here about contradicting ISO standards like programming languages ISO POSIX and ISO Linux. Stating that he believes there is no issue with overlapping standards.
Also did not see the issue with OOXML becoming a standard as standards are voluntary and no one has to use them. The issue I think with this is that if it gets pushed by a desktop monopoly the likes of Microsoft and the standard is not up to scratch, it has the potential to distort the market. Rick later said that OOXML would become the "de-facto standard" anyhow, with or without ISO approval, and that ISO was threatened with irrelevance if it did not approve it.
Rick believes that OOXML has had more review than ODF and is more complete than ODF and that the first paragraph of the ODF spec and OOXML spec are very similar and it is in the second paragraph that they deviate.
Rick made it clear that OOXML is application specific in its features. Which I think fails the basic idea of an ISO standard, because Microsoft and its dependent partners hardly constitute a group in that sense. He also stated that in regard to standards "small is beautiful".
How is 6000+ pages of poorly written specification small or beautiful?
Next up was Kaaren Koomen for IBM.
Karen was there alone as Paul Robinson was unable to attend as he had a back injury that prevented him from attending. Kaaren played a slide show and audio prepared by Rob Weir.
Some highlights was a graph that shows that a average standard is 50 pages and takes two years but OOXML is 6000 pages and has taken less that a year.
It reiterated the well known deficiencies of OOXML like the bit maps, the copy word 97 type behaviours etc. Also mentioned some features only support IE which is fine for MSDN but not for a ISO standard.
Recommended SA vote No with comments and take out of fast track and maybe split into multiple parts.
Rick Jelliffe disputed the graph because ODF includes various other standards that each
had their own process. Which I did not really get as it is a ratio so even if it was taken into consideration it would still be similar. ODF does use many other standards, instead of trying to
undermine them by defining conflicting methods for the same purpose, which bloats OOXML. The forum was moved on because of time constraints.
Next was Microsoft.
Greg Stone spoke for Microsoft there was also 2 other Microsoft staff there.
Stated did not want to get into technical aspects, said that ODF and OOXML are not competing standards but fit for different purposes. A goal was for a clearer spec believe that trying to accommodate requests is why it has blown out from a initial size of 2000 pages to 6000 pages.
Talked a lot about legacy support and that is why the spec needs this and that, seemed to imply it allows support of the old binary formats even though it does not describe them .
Noticed they called it OpenXML the same as Rick Jelliffe did in his presentation.
Microsoft mentioned Rick several times in his presentation.
Talked more about requirements to work with backend systems rather than document fidelity, also mentioned a Linux Java app that can use OOXML acting on the structured data. Believes requirements are different now than they were in the past years.
Had a slide with many examples of support of OOXML including the ODF converters etc, did not explain how ODF converters can work when one of the reasons that they do not like ODF is that it has deficiencies that meant it is not suitable for Microsoft to use.
Believe that any vendor can operate on a equal footing with OOXML. Talked about support from Apple IPhone (did not mention the issues they are having with OOXML support on OSX).
Believes that Microsoft's Open Specification Promise is as good as Sun's patent promise.
CompTIA was represented by Danaka Bakailch and were pro OOXML mentioned they have 25000 corporate members and gave the impression that she was speaking on behalf of the member companies. Kaaren from IBM asked how this position was arrived at as IBM is a member and does not agree with that position(Sun is also a member), I don't believe CompTIA adequately explained how they arrived at this position. Claimed it was feedback from members but this was disputed.
James Dalziel spoke as a member of E-Learning Australian Standards Committee (IT19-1) stated that they want a no vote as they believe 2 ISO documents standards would not be good for students, James also stated he is a member of IT19 (Standardisation of practices relating to libraries, documentation and information centres, indexing and abstracting services, archives, information science and publishing, particularly to produce standards to address the use of computer applications in information and documentation) and said two of the members of this committee wanted it stated that they want a yes vote on OOXML. The stated position of IT19-1 was heavily debated and disagreed with by Rick Jelliffe and several other OOXML supporters debated on this point, the general argument was that even different implementations of a single standard may not guarantee fidelity between different documents. One of IT19-1 worries was student's been forced to buy Microsoft Office to submit assignments, once again the forum was moved on.
In open questions comments they tried to give whoever wanted to say something a opportunity but they did not want to get into technical details or back and forward discussions.
There was quite a few pro OOXML companies represented but most of their points were more to do with standards and/or XML in general, and were not OOXML specific, but were represented in that way. The focus was more that transactional systems could use a defined schema to extract data, they did not focus on document fidelity between applications.
There was no real arguments for why it needed to be a ISO standard to achieve this or why it has to be fast tracked.
Archives Australia representative, (I think this was the organisation) stated that they worry about long term preservation of documents and were against OOXML becoming a ISO standard as it was not suitable in their opinion for long term document storage.
There was several people that spoke up against OOXML the fact it was been fast tracked,overly long did not reuse standards and was of poor quality were the recurring themes. The bit map issues, dates, impossible for anyone but Microsoft to implement, the problems with some of the mathematical functions showing a lack of technical review was also mentioned.
Google was represented by Lars Rasmussen and Georg Greve and they also spoke against OOXML becoming a ISO standard.
During discussions Microsoft's Greg Stone stated that he believed that Microsoft's Open Specification Promise was similar in wording to Sun's Patent promise and that it was at least as good as Sun's and that it is not ambiguous.
I asked Greg Stone about this and read out during the forum part of the promise namely “Microsoft Necessary Claims” are those claims of Microsoft-owned or Microsoft-controlled patents that are necessary to implement only the required portions of the Covered Specification that are described in detail and not merely referenced in such Specification.”
After the forum Microsoft gave me a Document from Baker & McKenzie titled Standardisation and Licensing of Microsoft's Office Open XML Reference Schema dated June 2006 (here).
They believe it should allay any fears about the Open Specification Promise.
At the end of the meeting Alistair stated that Standards Australia will accept comments unit the 21st of August but recommend they be submitted as soon as possible.
My gut feeling is that Standards Australia are leading to vote Yes with comments but this is just my opinion.