March 19 2008
Over the weekend, Microsoft published another public effort to show "hey, we're the good guys" on their efforts to take a heavily modified version of the Office 2007 document formats and get it rubber-stamped as an international standard. Oddly, the open letter from Chris Capossela asserts that IBM is on the MS bandwagon:
For example, Novell, Corel, Apple, IBM, Sun, and others have already adopted (or announced adoption of) Open XML (OOXML) in their products on a variety of platforms -- including Linux, Windows, Mac OS, Palm OS, Java, and .NET.This hits at one of the core issues of the OOXML saga that I have been highlighting for months. Microsoft claims that the Office 2007 document formats == ecma Open XML. Therefore, IBM's announced support for Office 2007 document formats in a few products == support for OOXML.
IBM doesn't support ecma Open XML. But who ever expects MS to be clear communicators?
Office 2007 shipped before ecma submitted OOXML to ISO for standards approval. There have been hundreds (thousands?) of changes to the spec since it was implemented in Office 2007. In fact, Microsoft's Jason Matusow states, in a comment on his blog, that
The specification has been public (publication of Ecma 376) for fewer than 18 months. At this time, there are now independent implementations of Open XML cropping up all over the place, none yet with the fidelity of Microsoft Office, but that will come.Perhaps that is because the file formats in Office 2007 <> the OOXML specification? Or that the specification at 6000 pages is so demanding that independent implementations simply can't get the "fidelity" right?
I've asked this question many times -- what happens to all the current Office 2007 documents when ecma changes the standard to adjust those thousands of issues identified during the ISO ballot process? How can those now be called "Open XML" documents as distinct from whatever comes next?
These ambiguities and fast-and-loose play with the facts are among many reasons why IBM has not does not and will not support OOXML. IBM's Bob Sutor said it best the other day:
IBM is opposed to this specification becoming a JTC1 (ISO/IEC) standard because it was developed in a non-open manner, is ridiculously large, is technically inferior, and emerged from the Ballot Resolution Meeting with most things not explicitly resolved and more questions than ever before.And if you want to know more about those questions, Rob Weir posted yesterday regarding "How many defects remain in OOXML?" with some very concrete examples.
The six month tragedy that is the ISO process for OOXML is almost over. I can only hope, based on everything that has been documented, that it doesn't end with an endorsement of this work. A lot of clearly smart people have invested a lot of time money and effort, but that doesn't make it best, good, or even right, if the decision is driven by calendar and politics. This whole thing should be a do-over.