I agree with you on structured information, but to me the critical point with XML (and other markup languages, I go back to GML for text markup on mainframes) is transformation. XML is good – but XSLT is what makes it great in my opinion. If data I want is in XML (or a subset, such as well-structured HTML) then I can use XSLT to transform it into whatever _other_ format I need. Similarly with other forms of markup or data identification – in fact, I developed XSLT which transforms an iCalendar (RFC 2445) file into XML – but I could have used any other programming language because the elements within iCalendar, the calendar components and properties, are marked up with identifiers. The markup makes these kinds of transformations possible. In theory (I haven’t tried it) a properly-coded XSLT could transform OfficeXML into ODF and _back_. That makes the two formats more-or-less functionally equivalent: I can extract information from the document using either set of markup.

The only issue I’ve seen with Office XML (and this was in Office 2003, may have changed) is that they one-for-one translate the “control codes” in an Office Document into markup, rather than going the extra mile and building two things: one the structured document, the other a stylesheet document with the document style information.

If anyone in there is listening — the tide is shifting.

My own actions to request re-vote about OOXML approval where blocked by Microsoft partners who just abstained from voting. So, Microsoft shouldn’t be really proud about this.