There can be no doubt that OOXML, as a standard, has severe flaws. It is incomplete, platform specific, application specific, full of contradictions, fails to adhere to existing standards, untestable, and presents a moving target for any IT worker. There is not an organization in existence, including Microsoft, that promises to actually implement the full standard. Much of this is due to the fact the final version doesn't actually exist on paper yet, but a large fraction is also do to the patchwork nature of the product.
The reason governments and companies wanted a 'office apps' standard in the first place was to release an avalanche of data from aging applications. OOXML shows every appearance of being created to prevent this escape, not enable it. The immaturity of the standard means that it remains a gamble to see if older documents will remain readable or not. The lack of testing means there is no way to determine what docs actually adhere to it or not. The ignoring of existing standards guarantees compatibility problems. All of these factors are handy for the owner of the biggest share of existing documents, as it forces users to continue to use only _their_ application or risk danger from every other quarter.