Microsoft Tech Ed 2007: OpenXML

Doug Mahugh gave a presentation on the Microsoft Office OpenXML (MSOOXML) file format and its one and only application which implements it, Microsoft Office 2007 at the Microsoft TechEd 2007 held in Kuala Lumpur Convention Center yesterday.
A clear and versatile speaker, he demonstrated the file format as a zip file, deftly moving between his PowerPoint presentation and opening the files by renaming it in Windows Explorer, and viewing the XML files in IE or editing it in Notepad. Its obvious that he has done this many times already, and its great to see an expert at work.
He was straightforward, honest and very frank during the question time. The only snide remark during the presentation made was when he apologized for his slow laptop loading Visual Studio "... its taking a long time to load because I have Linux running in a VPC here..." He probably gets a few laughs in other countries, but hey, we're in Malaysia, ... and it was late: 6:30pm.

He talked about VML and DrawingML a lot because of the nice eye-candy MSOffice2007 can render, talked about the "separation of Presentation and Data" extensively, too, as though only Microsoft does it this way. Some interesting things which was said was:

• "OpenXML is very flat; its just a series of paragraphs" - probably because it reflects the BIFFiness of its lineage. Stephane Rodriguez has an extremely detailed account of the similarities of MSOOXML and its binary counterpart BIFF12.
• "It is easier to transform hierarchical XML to flat XML, but harder to go from flat to hierarchical. A lot of inference of the file structure would have to be done" - this means that converting from ODF to MSOOXML would be easy, but the MSOOXML to ODF conversion would be trickier.
• "Part 4 of the OpenXML spec, the 5000 page section is not really required: Nobody would read it" - this trivializes the work done / ongoing work by ISO and the National Bodies who actually have to vet through this beast of a spec.

The Linux running in the Microsoft Virtual PC was a sample application which generated a document from a Apache / Java web application. The document had a header, watermark, text and some graphics. Nothing particularly spectacular. It could have been any read-only text based markup document, be it PDF, HTML or ODF.
He saved the file on his desktop and opened it in MSOffice2007. Voila it worked. Thus it proves that MSOOXML is interoperable on multiple platforms.

Not so fast. Try creating an application which implements MSOOXML as its native file format running on multiple platforms providing all the "rich" functionalities and Macros, THEN people will agree that  MSOOXML is platform independent.

He then showed MindJet's MindManager 7. The application has a nice GUI which behaves similarly to FreeMind. He demonstrated the "interoperability" with MSOffice2007 via MSOOXML. You can export the MindMap as a MSOOXML file and MSOffice 2007 can open the file and view the bubbles as Header 1, Header 2, etc. nested text. Adding a new Header, Saving the file, and Reloading it in MindManager will reveal a new bubble. Additionally, MindJet has its own free Ribbon Add-In to view MindMaps within MSOffice2007. That's Microsoft's definition of interoperability!

All I see are just primitive import and export filters. I'd be impressed if MindJet actually uses MSOOXML as its default file format! I mean, if MSOOXML so powerful and extensible, why not?

Doug accidently loaded the file generated by the Apache / Java web-app into the MindJet application. Now the generated file by Java is supposed to be a well defined MSOOXML document. However the primitive MSOOXML import filter by MindJet failed to handle a valid MSOOXML file and froze onscreen. The progress bar just remained at 80%. Interoperable?

When asked whether Novell or Apple, who "contributed" to Ecma's TC45 in defining MSOOXML, would strive towards making MSOOXML their native and default file format, Doug said:

"Apple have OpenXML support in their iPhone. Novell is working on the OpenXML plugins for OpenOffice"

When clarified that the OpenXML support were read-only features in the iPhone and iWork, he subsequently elaborated:

"Well, its too early for other vendors to commit to this file format. After the BRM (Ballot Resolution Meeting - in February 2008) there may be changes to it, so it is risky, and may not make commercial sense to implement OpenXML as it is at the moment."

Contrast this when ODF was being defined, there was active and definite migrations for two independent implementations of ODF. OpenOffice.org being the main one, and KOffice. The developers at KOffice had more to risk as ODF is completely different to their previous default file format. However even during the drafting stage, there were already plans to move towards this vendor neutral file format.

The reason Novell and Apple would want to avoid MSOOXML, is not because it is in flux at the moment, but more likely because MSOOXML is tailor made to address one and only one vendor's product. And that is the Microsoft Office Suite.

When asked to comment on ODF vs MSOOXML, he had this to say:

"Microsoft was careful not to compare the two, but the last few weeks before the vote (2nd September), the gloves came off"

Well, the gloves came off a long time ago in Malaysia when Microsoft Malaysia actively worked against ODF as a National Standard. His comparison was that "OpenXML is less consistent" than ODF and ODF "is easier to develop with". Of course he said that MSOOXML was more "feature rich".

He was asked "Why did Microsoft push OOXML through the "Fast Track" process instead of the standard ISO process? Wouldn't they get less resistance than faced now?"

His response was very frank:

"Office is a USD$10 billion revenue generator for the company. When ODF was made an ISO standard, Microsoft had to react quickly as certain governments have procurement policies which prefer ISO standards.
Ecma and OASIS are "international standards", but ISO is the international "Gold Standard".
Microsoft therefore had to rush this standard through. It's a simple matter of commercial interests!"

You cannot get more honest than that! Well done Doug. Please come again next year.

yk.

Comments

[Yoon Kit]

Hi Sammy,

I attended the TechEd with Ditesh in Kuala Lumpur. After the talk, Doug had a small discussion with about 10 people where we had a chance to ask him these questions. Most of the people hanging on were just looking for a freebie book.

I explicitly remember him saying those words, and jotting them down as accurately as I could. I was pleasantly surprised with his frankness, and respected his corporate position.

I'm not sure if Microsoft recorded the event, let alone recorded the small session after his talk. I myself did not record the event, so unfortunately you will just have to rely on my first hand reporting.

I can't make this stuff up: He did go on about the $10b revenue generator, equated ISO as the gold standard, and summed up that MSOOXML initiative was driven by commercial interests.

I put it down to his honesty, or maybe jet lag.

Ditesh would be the other witness, but the Microsofties would just say he too is biased ... no?

yk.

[Sammy]

Dough Mahugh has since denied making the "commercial interests" statement:

http://blogs.msdn.com/brian_jones/archive/2007/09/17/a-new-faq-page-now-available.aspx#4991719

What do you say? Do you have other witnesses? Perhaps a recording?

There has also been some wikipedia discussion on this: http://en.wikipedia.org/wiki/Office_Open_XML#Protection_of_Commercial_Interests

(and see talk).

[Yoon Kit]

> Since 2003, Mindjet MindManager's file format is zipped
> XML using a publicly available schema. Very similar in spirit to Open XML.

Hi Michael,

Thanks for clarifying. Just because a file format is in XML and zipped up doesnt mean its MSOOXML.

Here are some questions which NBs are asking third party developers such as yourself:

1) Did you offer the same "Export to Word" in binary formats?

2) Did you try to do so with MSOffice 2003 XML format? Or MSOffice 2000 HTML+ format? How does it compare with MSOOXML?

3) Will MindJet extend MSOOXML for it to officially cater to the requirements of MindJet?

4) Did third parties (e.g. Apple and Novell and MindJet) ever have any say in the direction of MSOOXML development?

5) Is Stephane's allegation that you are a Microsoft employee, paid to evangelise the benefits of MSOOXML to the public true?

Thanks for responding.

yk.

[Stephane Rodriguez]

Keep in mind, when Microsoft says anything about Mindjet, that :

- Michael Scherotter, the guy above, the MindJet guy, is actually a Microsoft employee. He was actually hired as a Microsoft evangelist by Doug Mahugh. Announcement here : http://blogs.msdn.com/dmahugh/archive/2007/06/22/san-francisco-workshop.aspx
- Also, and without much surprise, MindJet was one of the "neutral, innocent, almost naive" third-parties that votes Yes in the first US INCITS V1 ballot in July.

So anything you here about those guys, just make sure to read it as follows : MindJet = Microsoft.

If Microsoft's Doug Mahugh is making a reference to MindJet as an example of wonderful external implementation of whatever he fancies, I guess it's time to pack your things and move on.

[Michael S. Scherotter]

Since 2003, Mindjet MindManager's file format is zipped XML using a publicly available schema. Very similar in spirit to Open XML.

[Wu MingShi]

"Part 4 of the OpenXML spec, the 5000 page section is not really required: Nobody would read it"

Then,... why write it?

I think this is an insult to ISO/NB/many others who acted in good faith and to take the time to evaluate "Part 4".

He just gave ammo to one of the reason not to approve OOXML: Bloated beyond believe

[carlos]

"Microsoft therefore had to rush this standard through. Its a simple matter of commercial interests!"

ok, let's go, National Bodies: hurry up this +6000 review and give us a "yes"!! we have USD$10 billion at stake!

Pathethic