What’s wrong with OOXML anyway?
ISO has now approved OOXML. In FSFE, as our press release title says, that makes us concerned about quality of standardisation process; but if you’re like most of the world and haven’t followed every OOXML blog entry from Andy Updegrove, noooxml.org and Groklaw, you may be wondering: what’s wrong with OOXML anyway?
I’ll try to summarise points from those blogs and FSFE’s previous writings on the issue. Because this news is unfolding fast, I’ll put this blog entry online section by section as I write it.
Patents
When people pointed out the danger of Microsoft’s patents, Microsoft published an "Open Specification Promise" saying that they won’t use their products against your project "to the extent it conforms to" the OOXML specification. However, Microsoft’s own word processor does not fully comply with the OOXML specification, so if you try to make software compatible with their word processor, they can still use their patents against you. (This was pointed out FSF Latin America)
Further problems with their promise have been detailed in SFLC’s document No Assurance for GPL.
Technical
Archive your data here?
It is known that the OOXML specification contains hundreds of problems. The Grokdoc website lists many, and the national standards bodies submitted more than 1,100 in September 2007. When these were discussed in Geneva at the end of February, it was concluded that there was nowhere near enough time to discuss all the problems. A bulk vote was proposed on the Thursday and on Friday a 6:4 vote approved, without discussion, Microsoft’s suggestons for 900 problems.
ISO not up to the task
It seems ISO was not up to the task of making a technical decision under pressure, and failed to fix their process when the cracks became obvious. ODF remains the only open standard. There are claims of voting irregularities in United Kingdom, Germany+Croatia, Norway, Poland and others.