The gnomes of ISO (err, ISO/IEC JTC1 SC34 WG5) have released the second draft of their Technical Report comparing ODF and OOXML (PDF). It is up to 126 pages now, and much more fleshed out than the first draft.
One of the issues that came out during the intense lobbying of 2007/2008 was that a lot of work needed to be done comparing the formats before anyone could really justify dogmatic statements that ODF could already do everything needed by OOXML on the one hand, or that ODF's rationale and approach was so different and incomplete that it was and always would be completely unsuitable on the other. What we had was working theories and a mix of prejudice and expertise.
So now after a couple of years effort, we are starting to get a better shape of things.
I have made a little scorecard, based on a quick count of the comparison of features and functions in section 6 of the report. Combining the word processing and metadata counts:
- Features or functions whose translatability (between ODF and OOXML) is rated High: 33
- Features or functions whose translatability is rated Medium: 50
- Features or functions with low translatability, from lack of support in OOXML: 16
- Features or functions with low translatability, from lack of support in ODF: 11
So that gives us about 25 percent of the broad features counted as having low translatability (I don't know that 11 and 16 are significantly different, btw.) And less than a third of features with high translatability.
These differences reflect not just file format capabilities, but reflect the different capabilities of the software which implements each format natively: Word and OpenOffice. Getting reliable interoperability with these current formats is not a matter of picking the format, but avoiding rare or fancy features (or only using them when they are ephemeral decoration.)
One opportunity that this report may give us is to be able to write a kind of validator that tells users when their document has features that may not survive round-tripping well.