January 26, 2008

Commentary: ISO should kick OOXML off the standards bus

Author: Russell Ossendryver

ECMA, the international IT standards association, recently published its responses to comments of the ISO National Bodies in response to Microsoft's Office Open XML application for ISO standardization (the actual 2,293-page response is closed to the public). The ECMA proposals will be discussed at a Ballot Resolution Meeting (BRM) in Geneva after which the National Bodies may reconsider their original vote. Microsoft's responses make clear that within one year, it will have four different OOXML specifications to implement and interoperate with, and each of those specs will be closed. Under no circumstances should such a flawed specification become an international standard.

The main specification is OOXML 1.0 (i.e. ECMA 376 today). There are no specs available for the three additional ones, but they have been summarized by French developer Stephane Rodriguez in comments to a key issues like intellectual property rights, contradictions with other standards, and similar policies raised by National Bodies as objections to OOXML as an ISO standard. Both Microsoft representatives and ECMA are avoiding a discussion on these policy objections to OOXML. Microsoft wants this approval process to be "left to the experts" and asserts that the discussion should be only technical in nature. Confusing the process by deprecating some parts of OOXML functionality is part of the avoidance. And Microsoft still has not disclosed the migration mapping tables for OOXML, even though OOXML is a migration format! It has also raised more smokescreens in the form of vague allusions to future release information about some older, binary formats for competitors. In particular, the ambiguities in regard to the intellectual property rights involved is such that no FOSS developer, and few proprietary software developers, would care to become involved with mapping documents in binary format (.doc, .xml, .ppt) to the Office Open XML format

Microsoft is still subject to the terms of the injunction in the US federal anti-trust case, which requires the company to disclose certain middleware API specs, yet it has not released the specs for the intermediary file formats used by the Office native file support APIs -- more walled gardens.

Features such as the Microsoft Office OOXML file format with DRM, Sharepoint tags, passwords, reliance on Devmode (a method Windows uses for handling information about printer or display settings), GUID (a proprietary Microsoft Windows and .Net implementation of the UUID standard for applications to coordinate and identify resources within an operating system), migration tags, VBA macros, and other hidden system dependencies effectively prevent competing applications and even other operating systems from achieving full interoperability, while at the same time tying OOXML files to a Microsoft environment.

Here are some further examples of the dependency on Microsoft software: The "Disable Features Not Supported by Target Browser" feature is designed to optimize for various version of Internet Explorer and disregards Internet Explorer's main competitors, Mozilla Firefox and Opera. The "Disable Features Incompatible with Earlier Word Processing Formats" feature explicitly states that OOXML only considers the needs of Word 97 through 2003. This has every appearance of a ploy to keep control over all the old documents in legacy formats, newly created documents, and those that might be funneled through Microsoft Office's own custom XML format. The fact that so many features in Office 2007 are not in the ECMA spec means that competitors have no chance properly use data locked up like that. In particular, the deprecated parts with the hidden Microsoft properties will always be a permanent feature in the files Microsoft Office creates. These undocumented extensions and hidden APIs appear to be an ongoing scheme to maintain and extend their monopoly. Again, more walled gardens, not open standards.

Objections are economic, not technical

But the whole point of gaining ISO status is economic and not just technical. ISO is first and foremost about lowering unnecessary obstacles to international trade and creating even playing fields so competition can thrive. Microsoft OOXML would raise obstacles to trade by giving ISO status to only one vendor's Office productivity suite -- Microsoft Office. A 37-page report on OOXML and Open Document Format by The Burton Group highlighted this. It states, "Realistically, it's also extremely unlikely another vendor will attempt to exploit OOXML in order to produce a comprehensively competitive alternative to Microsoft Office, given the considerable resources that would be required to do so and the economies of scale that would need to be achieved in order to make the endeavor sustainably profitable."

The mere fact that a company is a de facto monopoly and offers many fragmented document formats does not entitle that company to a carte blanche right to an ISO standard simply because it dumps these formats into XML. The ISO National Bodies must not give Microsoft the ticket to dominate an industry with a file format.

Ideally, an ISO standard should represent a single international solution that can be applied by all countries. Global relevance then is the characteristic of an ISO standard through which "it can be used/implemented as broadly as possible by affected industries and other stakeholders in markets around the world." ISO standards have historically been developed using a multinational and multi-stakeholder approach, where open procedures give transparency and the principle of consensus is applied.

That has not been the case with OOXML. The ISO approval of a vendor-dependent specification that benefits only one company would run counter to the ISO goal of "one standard, one test, and one conformity assessment procedure accepted everywhere." The adoption of OOXML in this way would effectively extend Microsoft's exclusive rights to the format as a standard, and to the conversion of its old, outdated binary formats to XML. By preventing OOXML from being used by the market, Microsoft prevents interoperability with all other office suites, and OOXML becomes unimplementable and dependent on the Microsoft Windows operating system.

Further, all this smoke and noise distracts from the existence of an approved ISO standard for office documents, the industry-backed Open Document Format, which has been the recognized standard since May 2006.

In Geneva, the bus system works on the honor system and so does ISO. You buy a ticket and get on. You can also get on without a ticket. OOXML has gotten on the bus without a ticket one too many times. OOXML is an erroneous, unreviewed spec. It has never been implemented, it is not interoperable with the established ISO/IEC 26300 for office documents, and is more of an internal technical spec than something that is designed to become an international standard. It's time for an inspector to get on the bus in Geneva and throw OOXML off.


  • Commentary
  • Standards
  • Office Software