Universal Interoperability Council

Circumstantial evidence is mounting of one or more private deals having been struck to approve DIS-29500 Office Open XML ("OOXML") as an international standard, a deal that may have played a role in several key national standardization bodies changing their voting position to approve OOXML.

Beneath the headlines of protestors marching outside the Oslo, Norway building where Subcommittee 34 ("SC 34") of the ISO/IEC Joint Technical Committee 1 was holding a meeting, SC 34 was busily adopting resolutions relevant to future work on OOXML and ISO/IEC:26300 OpenDocument. SC 34 is the JTC 1 technical committee with ISO/IEC jurisidiction over both standards.

The relevant SC 34 resolutions adopted April 9 establish two ad hoc committees to perform further work on ISO/IEC:29500-2008. Under Resolution 4:

SC 34 envisages the creation of three distinct working groups that meet the needs of:

1. ISO/IEC 29500
2. ISO/IEC 26300
3. Work on interoperability/harmonization between document format standards[.]

...

For these reasons, SC 34 hereby establishes an ad hoc group pursuant to the JTC 1 Directives, clause 2.6.2, for investigating how the first of these groups may be set up most effectively[.]

What's wrong with this picture? Several factors point to the existence of one or more deals to implement this strategy before the close of the OOXML ballot period in exchange for NBs changing their votes on OOXML. Consider the following facts and attempt to reconcile all of them in your own mind with the non-existence of such a deal:

• All NB comments on OOXML calling for its harmonization with OpenDocument were rejected at the ballot resolution meeting in February and that decision was therefore ratified by the final vote to approve OOXML as an international standard without harmonization. In other words, the final ballot of the NBs rejected the harmonization of ODF and OOXML.

• On the date of the official announcement of the DIS-29500 ballot results, an IBM spokesman responded,

IBM will continue to be an active supporter of ODF. We look forward to being part of the community that works to harmonize ODF and OOXML for the sake of consumers, companies and governments, when OOXML control and maintenance is fully transferred to JTC1.”

• Question: How did IBM know that "control and maintenance" of OOXML would involve work to harmonize OOXML with ODF prior to the April 9 SC 34 resolution?

• Question: Does it seem reasonable to believe that Microsoft would lobby to successfully defeat NB comments calling for harmonization with OpenDocument at the ballot resolution meeting and win that issue in the final ballot yet quietly allow SC 34 to impose harmonization nonetheless, unless a deal had been made to that effect?

• SC 34 decides its resolutions at the closing plenary sessions of its meetings. NBs and liaisons present at the plenary session, according to the document disclosing the resolutions, were Canada, Czech Republic, Denmark, Finland, Germany, Japan, Korea (Republic of), Norway, South Africa, Sweden, UK, USA, Ecma International and the XML Guild

  Several NBs that voted against OOXML approval were unrepresented, as were even more that voted for OOXML approval without harmonization, and as was OASIS, the maintenance organization for OpenDocument. SC 34 operates under a strict rule that agenda items must be distributed prior to its meetings.

• Question: Is it reasonable to believe that attendance would be so sparse and that the OASIS liaison would not bother to attend the plenary session if the harmonization decision were not a deal done prior to the meeting? Is harmonization of OOXML and OpenDocument so insignificant an issue?

• The government of France has adopted OpenDocument as a national technical regulation and procurement specification. France's AFNOR NB, the same body that established OpenDocument as a national technical regulation, took an understandably strong position in its comments on OOXML that OOXML must be harmonized with OpenDocument prior to OOXML's approval. See e.g., Groklaw article with translation of French NB position.

  The French NB position was consistent with the Agreement on Technical Barriers to Trade article 2 section 2.6, which contemplates that harmonization occur at the preparation stage of an international standard's development, not post-adoption. France changed its final ballot position on OOXML from disapprove to abstain.

• Question: Does it seem likely that France, deeply committed to OpenDocument by its prior decision and migration to OpenDocument, would abandon its opposition to OOXML being approved prior to harmonization, absent the existence of a deal?

• Does it seem credible that France would not bother to attend the SC 34 meeting that would decide to harmonize OOXML with OpenDocument after JTC 1 approval of OOXML if the outcome of that SC 34 meeting were not decided in advance?

• British Standards Institute submitted what are easily the most detailed specific comments calling for harmonization of OOXML features with OpenDocument prior to OOXML adoption and BSI voted "no" on OOXML approval. Post-ballot resolution meeting, BSI changed its vote from "no" to "approve," even though harmonization had been rejected at the BRM.

• Question: Is it reasonable to believe that BSI, which has one of the best reputations among NBs globally for quality input on international standards development and as a stabilizing influence, would turn its back on its substantial political and technical investment in the harmonization issue and reverse its vote absent a deal to achieve harmonization?

• The OASIS OpenDocument Technical Committee ("TC") is dominated by IBM and Sun Microsystems. ODF editor Patrick Durusau had stayed out of the public fray of the document file format war. On February 6, Durusau broke his silence and began publishing a series of articles steadily growing more supportive of OOXML approval by JTC 1, implicitly endorsing the notion that competing standards are a Very Good Thing. (One might agree that they are a Very Good Thing for professional standard editors like Durusau whilst still believing they are not so good for the software users of the world.)

  Just prior to the OOXML ballot resolution meeting, Durusau endorsed the concept of co-evolution for OpenDocument and OOXML, a concept later appearing to be adopted by the SC 34 resolution. Immediately prior to the ballot resolution meeting, Durusau endorsed OOXML adoption. Shortly after the ballot resolution meeting, Durusau attacked IBM for its opposition to OOXML.

• Question: Smokescreen or candor? Does it seem reasonable to believe that Durusau would jeopardize his paid position as OASIS OpenDocument Editor by endorsing OOXML and publicly attacking IBM, or is it more reasonable to suspect that Durusau was laying down a smokescreen for a deal already struck that had IBM and Sun's approval? The fact that he has a contract with Sun to edit OpenDocument v. 1.2 would not seem to tie IBM's hands and IBM has more votes on the OpenDocument TC than Sun has.

• The OpenDocument TC has been all but moribund for months, with a strong appearance that folks were waiting to see what JTC 1 decided about OOXML approval. This is understandable given that many NBs were pushing for harmonization of OpenDocument and OOXML prior to the final OOXML vote.

  However, between the close of the ballot resolution meeting and the close of the period for NBs to change their votes, on March 17, a South African government official announced that OpenDocument v. 1.2 would be available for public review as a TC draft during May and June of 2008. Such availability is prelude to a ballot among OASIS members whether to adopt ODF v. 1.2 as an OASIS standard and to send it off to JTC 1.

• Why would IBM and Sun suddenly decide to ship OpenDocument v. 1.2 to JTC 1 as rapidly as possible some two weeks before the final count of votes on OOXML at JTC 1 if there were no deal?

• SC 34 claims that its authority to embark on this work is conferred by JTC 1 Directives (PDF) section 2.6.2. But that is an argument that eats its own tail. Section 2.6.2 simply provides procedures to create a special working group. It provides no authority at all for countermanding the vote of the NBs on an international standard.

  OOXML, like it or not (and I do not), was approved with specific rejection of harmonization. New SC 34 work items are required to be submitted as a New Work Item Proposal ("NP") pursuant to JTC 1 Directives, section 6. One would think that an SC 34 decision to contravene a vote on an international standard would at least require approval by ITTF or the JTC 1 Secretariat. But if Google search is any guide, no NP was submitted for this work item.

• Question: JTC 1 has spoken. What is SC 34's authority for harmonization work post-OOXML approval and where is the NP if this work has any authority other than a private deal?

• Simon and Garfunkel once made the record charts with a song that included the lyrics,

  I turned my collar to the cold and damp

  When my eyes were stabbed by the flash of

  A neon light

  That split the night

  And touched the sound of silence.

  Occasionally silence speaks louder than words.

• Question: Why have all involved been so quiet about all this prior to an SC 34 decision being made?

• Question: Is there any commitment by anyone --- most especially Microsoft --- to support a harmonized version of OOXML/OpenDocument? I note that there are precisely zero full-featured implementations of either standard that are capable of writing to the official standard without vendor-specific extensions.

• Why should anyone bother to contribute to this effort if there are no such commitments?

Alex Brown, convenor of the OOXML ballot resolution meeting and newly-appointed convenor of the new SC 34 ad hoc group, has bloggged about what happened at the Oslo meeting. Rick Jelliffe has too.

A cynic might suggest that telling the world what you have done only after you have done it falls a fair bit short of the ideal in decision-making process transparency. It would be nice to be told that such issues are under discussion. Especially when all signs point to a back room deal before the final vote positions on OOXML were decided.