The Linux Foundation

Architects on OOXML

From The Linux Foundation

Revision as of 21:37, 27 August 2007 by Cherry (Talk | contribs)

Desktop Architects Speak Out on OOXML


John Cherry - Linux Foundation; Global Initiative Manager

The Linux Foundation supports the adoption of open standards as a catylist for innovation and to define common frameworks for new development. With ODF (Open Document Format) as an existing ISO document standard, we do not consider it beneficial to introduce an alternative document standard. Since both ODF and OOXML are both designed as formats for editable documents, computer users would greatly benefit from multiple implementations of a single standard rather than suffer the confusion, conversions, and incompatibilities involved with multiple standards.

If you could argue that multiple standards to address a common problem is the right way to go (which I don't believe to be true), then you would have to look at the technical merits of both standards and eventually it would come down to the survival of the fittest. With the vote among the various ISO committees regarding OOXML coming on September 2, we feel that the review process for this 6546 page standard document has been insufficient and that there are significant technical issues that must be addressed. These include the continued use of binary code tied to platform specific features, propagating bugs in MS-Office into the standard, proprietary units, references to proprietary/confidential tags, unclear IP and patent rights, and more (see the reference from below).

Dan Kegel - WINE architect

One of my concerns is that Microsoft reserves the right to sue you if you implement all of OOXML. The Microsoft "Open Specification Promise" and the earlier "Covenant Not To Sue" only cover the *required* portions of the standard, not the optional sections, e.g. the ones that handle backwards compatibility. Thus Microsoft seems to be saying that any OOXML file that was produced by loading a Word 2003 file and saving it as OOXML (and that's going to be the most common case for some time) cannot be legally read by a competing implementation without licensing Microsoft's patents.

Klaus Knopper - creator of Knoppix and Gnoppix

  • The proposed standard must be implementable by everyone. Insufficient or vendor-specific documentation cannot be implemented by anyone in full.
  • The proposed standard must be architecture- and operating-system independent.
  • The proposed standard must not be obfuscated.
  • All components and interfaces must be openly disclosed and documented in full, and not given as a "container" for proprietary content.
  • The proposed standard must not contain patented/proprietary components.
  • Implementing the standard must be royalty-free.

Scott Preece - Desktop Architect

I understand that the proposed standard is really a surrogate for Microsoft's claim to legitimacy in markets where support for open standards and interoperability are critical.

However, it's also true that (a) in many problem spaces there are multiple standards and (b) the normal role of the standards process is to formalize existing practice rather than to make policy or invent "better" practices. I once chaired a POSIX working group that was trying to forge a single standard to unify practice where there were two existing practices; ultimately the effort failed and the two practices were individually standardized. And the world didn't end.

David "Lefty" Schlesinger - Chair, Linux Foundation Mobile Working Group

There are serious issues with OOXML being adopted as a standard, as the references show.

  • Interoperability is, in fact, pretty limited
  • The legal status of future versions is murky
  • The proposed standard is ginormous (that's an official word now, according to Merriam-Webster), and hasn't had a reasonable review period (and might, at over 6,000 pages, be too large to reasonably review, especially since it's pretty much duplicative of ODF)

The GrokLaw article references Google's statement on OOXML, which I endorse. If there are aspects of ODF which Microsoft would like to see improved, then let's work together to improve ODF. This new proposed standard seems unnecessary and burdensome.

John Walicki - IBM Open Client Architect

IBM is a strong supporter of open standards because they allow for fair competition among market players based on the creation and the implementation of the standards and do not allow one party to control the standard and therefore the marketplace. The ODF standard can be implemented and used by anyone on any platform, at any time, and no license or fee is required. This provides consumers with complete control and ownership of their documents, forever. ODF is truly open with contributions from dozens of companies. Changes to ODF will likewise be determined by the community. ODF provides a format that will be available to all without any specific company or companies controlling its fate or future.

We are very concerned about many aspects of the ooXML specification that is currently under ballot in ISO/IEC JTC1 including many technical issues concerning the quality of the specification.

  • It is unclear whether all vendors will be able to implement a fully compliant version of the ooXML specification. The specification references other specifications which are not publicly available and for which the Intellectual Property Rights are unknown.
  • The ooXML specification conflicts with many other International Standards (Language Codes, Paper Sizes, Colour codes, etc.) by implementing its own specific codes.
  • The ooXML specification defines capabilities (e.g. Mathematical Formulas, Vector Graphics) instead of simply referring to W3C specifications, which are internationally accepted and widely used.
  • The ooXML specifications rely on Windows specific capabilities. Document exchange with other operating systems is not guaranteed. An International Standard should be technology neutral.


[Article] [Discussion] [View source] [History]