Brian Jones has announced the availability of the final drafts of Ecma Office Open XML.
Once Ecma endorses it, as seems 100% likely, this is the standard that will be hot-footed to ISO for fast tracking. I think it is 75% likely that ISO member countries will accept it, for two main reasons:
After all the years of complaining about lock-in, binary formats and lack of documentation, it is pleasant to be able to report that paradoxically it is the scale of OOX that is probably its main technical obstacle to acceptance at ISO: the amount of scrutiny that such a large spec can have is fairly limited. When a 10-page spec has a serious problem, it can be a show-stopper for standardization: when a specification for thousands of elements has some problem but thousands of things right, what should a reviewer (for a national voting body) do?
I think the answer to a lot of issues with standardization (what should be standardized, what quality should it be, what implementations should there be, what buy-in should be required, what should be done if there are competitive standards, and so on) can be readily solved by going back to first principles. The first principle of standardization at ISO is that a standard is an agreement. The nature of the things to be agreed on when the technology is pre-existing (i.e. Office) is quite different to when new standards are made, either from nothing or by combining different developer’s best practises.
Should your goverment use ISO/OASIS ODF or ISO/Ecma OOX or XHTML? Err…, yes. All of them, where they are appropriate. For converting archives of Word .DOC and .RTF documents into an XML format that retains fidelity, clearly OOX is the winner. For maximal interchange between applications and preventing application-specific formatting conventions leaking through, ODF is probabably the winner. For text which needs to have the maximum reach in different media, XHTML is the winner. For documents which have embedded data that non-COTS applications use and which will be re-targeted to different kinds of publications (not just the same publication in different media/styles) then making your own home-made schema (subsetting public vocabularies whereever possible) is the appropriate. Outside of that, it is a dogfight, and the top-dog will emerge over time I suppose.