Dr AJG’s Commentary

The ODF Alliance has released a  Fact Sheet on Microsoft’s ODF Support, a document that I find to be a very interesting piece of fiction. Before I address the various issues raised in that “Fact” Sheet, I probably should point out a few issues regarding my relationship with Microsoft:

• I don’t like them. I don’t like their attitude, and I don’t like their products in general.
• The last time I willingly used a Microsoft software product, was in 1990: I unfortunately used Microsoft Word to write my dissertation. (Ocasionally I have to use MS WIndows since the classrooms I teach in have instructor computers running MS Windows, but I don’t consider that “willingly”.)
• Since 1993, the department I worked in has moved from NextStep, to Openstep toGNU/ Linux. (We did inherit an MS WIndows lab for a while from another department but have changed it to GNU/Linux since then.)
• If I am bias (and I probably am) then I am bias against Microsoft.

None of this of course means that I like unfounded  accusations or one-sided statements. Free software can and will stand on its own. We do not need to stoop to inappropriate behaviour.

In my last post, I mentioned that more than half of the “wedding files” do not validate using Sun’s convenient ODF validator.

When I recently worked on Gnumeric’s ability to write ODF files, I found myself  using that validator frequently to ensure I didn’t accidentally had Gnumeric create invalid files. Often I had to go back and reorder xml elements I was creating since the ODF schema is quite particular.

Of course to see that my files wre not just syntactically correct but in fact would transfer the desired meaning I regularly opened the files in OpenOffice 3.0.1, since I had viewed that program as a good sample implementation of ODF.

Imagine my surprise when suddenly a certain feature did not load correctly although I was quite sure that my file was correct: it had alidated and the required elements all seemed to be in place. Some longer digging showed that it was in fact the sample implementaion that I had relied upon that

• created invalid ODF files, and
• failed to import some simple feature

simply because of a spelling error in their element names. A mistake that would have been detected by validation against the schema.

(For those interested, it should have been <style:diagonal-tl-br-widths> rather than <style:diagonal-tl-br-width> and <diagonal-bl-tr-widths> instead of <diagonal-bl-tr-width>.)

Recently Rob Weir posted some commentary on Microsoft’s ODF support.

I am anything but a Microsoft friend, but there are many misleading points in Rob’s presentation:

• It has been clear for a along time that the ODF standard primarily tries to codify various implementations and as such does not guarantee interoperability. Why blame Microsoft if their implementation does not further interoperability either? The writers of the ODF standard are to blame.
• Why didn’t anybody notice that more than half of the files Rob mentions are not valid ODF? Microsoft’s ODF file does in fact validate. (I have used Sun’s validator for these tests.)
• Gnumeric has been supporting ODF for a while. While we don’t do very well, even we can read Microsoft’s ODF files, including their formulas. (They appear to be simply Excel’s formulas that every spreadsheet implementation needs to be able to parse anyways.)

If everything would be as easy to install as Wordpress, I would spend so much less time on the computer…