We are intervening in the Apple vs.EC litigation. Become a proud supporter of the FSFE and join us in defending software freedom from monopoly control: https://my.fsfe.org/donate!

This page has not been translated yet. Please help us to translate this and other pages on fsfe.org, so people can read our message in their native language.

Six questions to national standardisation bodies

[Also available as PDF (28k)]

The following six questions relate to the application of the ECMA/MS-OOXML format to be accepted as an IEC/ISO standard. Unless a national standardisation body has conclusive answers to all of them, it should vote no in IEC/ISO and request that Microsoft incorporate its work on MS-OOXML into ISO/IEC 26300:2006 (Open Document Format).

This is a summary document. More detailed information is available online.

  1. Application independence?

    No standard should ever depend on a certain operating system, environment or application. Application and implementation independence is one of the most important properties of all standards.

    Is the MS-OOXML specification free from any references to particular products of any vendor and their specific behaviour?

  2. Supporting pre-existing Open Standards?

    Whenever applicable and possible, standards should build upon previous standardisation efforts and not depend on proprietary, vendor-specific technologies.

    MS-OOXML neglects various standards, such as MathML and SVG, which are recommendations by the W3C, and uses its own vendor-specific formats instead. This puts a substantial burden on all vendors to follow Microsoft in its proprietary infrastructure built over the past 20 years in order to fully implement MS-OOXML. It seems questionable how any third party could ever implement them equally well.

    What is the benefit of accepting usage of such vendor-specific formats at the expense of standardisation in these areas? Where will other vendors get competitive, compatible and complete implementations for all platforms to avoid prohibitively large investments?

  3. Backward compatibility for all vendors?

    One of the alledged main advantages of MS-OOXML is its ability to allow for backward compatibility, as also referenced in the ECMA International press release.

    For any standard it is essential that it is implementable by any third party without necessity of cooperation by another company, additional restricted information or legal agreements or indemnifications. It is also essential to not require the cooperation of any competitor to achieve full and comparable interoperability.

    On the grounds of the existing MS-OOXML specification, can any third party regardless of business model, without access to additional information and without the cooperation of Microsoft implement full backward compatibility and conversion of such legacy documents into MS-OOXML comparable to what Microsoft can offer?

  4. Proprietary extensions?

    Proprietary, application-specific extensions are a known technique employed in particular by Microsoft to abuse and leverage its desktop monopoly into neighboring markets. It is a technique at the heart of the abusive behaviour that was at the core of the decision against Microsoft by the European Commission in 2004 and Microsoft is until today continuing its refusal to release the necessary interoperability information.

    For this reason, it is common understanding that Open Standards should not allow such proprietary extensions, and that such market-distorting techniques should not be possible on the grounds of an Open Standard.

    Does MS-OOXML allow proprietary extensions? Is Microsoft's implementation of MS-OOXML faithful, i.e. without undocumented extensions? Are there safeguards against such abusive behaviour?

  5. Dual standards?

    The goal of all standardisation is always to come to one single standard, as multiple standards always provide an impediment to competition. Seeming competition on the standard is truly a strategic measure to gain control over certain segments of a market, as various examples in the past have demonstrated.

    There is an existing Open Standard for office documents, namely the Open Document Format (ODF) (ISO/IEC 26300:2006). Both MS-OOXML and ODF are built upon XML technology, so employ the same base technology and thus ultimately have the same theoretical capabilities. Microsoft itself is a member of OASIS, the organisation in which the ODF standard was developed and is being maintained. It was aware of the process and invited to participate.

    Why did and does Microsoft refuse to participate in the existing standardisation effort? Why does it not submit its technological proposals to OASIS for inclusion into ODF?

  6. Legally safe?

    Granting all competitors freedom from legal prosecution for implementation of a standard is essential. Such a grant needs to be clear, reliable and wide enough to cover all activities necessary to achieve full interoperability and allow a level playing field for true competition on the merits.

    MS-OOXML is accompanied by an unusually complex and narrow "covenant not to sue" instead of the typical patent grant. Because of its complexity, it does not seem clear how much protection from prosecution for compatibility it will truly provide.

    Cursory legal study implies that the covenant does not cover all optional features and proprietary formats mandatory for complete implementation of MS-OOXML. So freedom of implementation by all competitors is not guaranteed for the entire width of the proposed MS-OOXML format, and questionable even for the core components.

    Does your national standardisation body have its own, independent legal analysis about the exact nature of the grant to certify whether it truly covers the full spectrum of all possible MS-OOXML implementations?

All these questions should have answers that should be provided by the national standardisation bodies through independent counsel and experts, and in particular not by Microsoft or its business partners, which have a direct conflict of interest on this issue.

If there is no good answer to any one of them, a national body should vote no in ISO/IEC.