Chris Capossela (head of the Information Worker business group) was over in the UK today and I got to join him in a couple of his briefings.  Chris is a very impressive spokesperson and I learnt a lot listening to him handle questions.  We got into some really good debates and I thought one of the discussions about XML was insightful and which had never occurred to me before in this way.

We were discussing why Microsoft didn't adopt ODF as the standard for 2007 but instead created our own standard, OpenXML.   Brian Jones has a great post on this last week where he responds to an articulate critical view on his blog - well worth a read.  I had the pleasure of meeting Brian in New York a while back - a laser sharp, very thoughtful, affable and humble guy.

In our discussion today, the arguments we debate were along these lines:

  • We could not use ODF because we have to guarantee backwards compatibility.  There are billions of documents out there from older versions of Office and we need to ensure that all of them can be opened in the new format with full fidelity.  In years to come, our customers tell us they need to be able to read Office documents without needing to dig up some archaic version of Office.  The content needs to become independant from the application that created it.  XML is the way to achieve this.  Since ODF is based on the feature set of OpenOffice which is perhaps on a par with Office 2000, it is not reasonable to expect it to fully describe documents created in officexp or 2003.  ODF does not have a way to fully describe pivot tables for example.  For this reason we could not just use ODF as our chosen XML standard.
  • ODF was not sufficiently described.  Brian Jones' blog talks more about the thousands of pages of documentation on OpenXML.  At the time when 2007 entered development, ODF had not reached a ratified standard status.  There was insufficient depth making in impossible to implement without unacceptable levels of creative interpretation.
  • Translation between applications is only ever as good as the extent to which those applications overlap.  When Microsoft decided to create strong portability between Word and WordPerfect it took thousands of person years to get that fidelity near perfect.  Even then it was never perfect.  Chris' view was that translation will never be perfect.  When you translate form French to English, you lose something.  There are words in French that do not exist in English and so something, however subtle, will be lost.  We like to hope that XML creates this Nirvana of interoperability.  It definately helps a huge amount but translation between schemas from different applications will never be absolutely perfect as long as those applications do slightly different things.

I'm no Brian Jones and will never understand the issues as deeply as he does but this helped me understand the reasoning better.  

I encourage you to take a look at his blog if you have never done so because it is an excellent example of great blogging (IMHO).  As I was saying to Chris today, great blogging is about dialogue and Brian does that superbly.  Even if you disagree with him (I find he is very presuasive), you have to applaud his true blogger spirit :-)