Two of the topics in the title are things we normally don’t touch in this blog. However, the tempest over Massachusetts’s OpenDocumentFormat decision is inflaming passions almost as much as religious and political issues do. In fact, I am writing about it because I woke up irritated at how ill-informed and irrelevant so much of the discussion about the state’s decision is. (Not a good way to start a blog entry!) I promised myself not to go on for more than the length of a reasonable blog-entry, so rather than dig into all the weeds, here is a short history lesson to bring out the big picture, and hopefully keep the debate focused on the real issue for Massachusetts’s and others contemplating similar decisions.
When we (in the standards community) debated open document standards 20 years ago, there was a religious and political fervor fueling the arguments of both sides. Our side (the SGML side, which included Tim Bray and Jean Paoli, now the chief XML people at Sun and Microsoft respectively), argued that nobody’s content should be held hostage by being stuck in a vendor’s proprietary format, and that the solution was a standard set of rules for describing whatever kind format was necessary that vendors were free to implement. The other side (the ODA “Office Document Architecture” side) agreed with that, however they thought the solution was for a bunch of vendors to get together and agree on a format that, instead of being proprietary to a single vendor, was proprietary to a self-defined group of vendors. This solution was even worse than the status quo for lots of reasons (lowest common denominator functionality, enhancements by slow international committee, unhealthy cabal-like motivations, …). At the time I thought of ODA as the soviet approach, and the SGML approach as the democratic approach. Fortunately, the SGML approach won, and that set in motion the developments that have given us XML today.
You can tell where I am going with this. But there is one more relevant aspect of this history to mention. One of the main arguments behind ODA was that the SGML approach was just too difficult to implement. They had a point, you have to pay for the freedom of flexibility. Their mistake was thinking there was an alternative that could anticipate all reasonable requirements. It can cost even more when you just can’t implement what you need to.
The situation today is a little different, but the need for organizations to be able to do whatever they want with their own content is exactly the same. The imposition of any single schema/format on all documents in any organization simply won’t work. Anybody who has been involved in helping organizations build IT applications knows that exceptions are the rule, and you can’t legislate them out of existence even in authoritarian corporate environments. A good decision for the state would be to simply require all documents to conform to one of a number of publicly documented and freely available XML Schemas – who cares what software did or did not create the content or did or did not design the schema? Certainly there are some complex details to work out, but there is no mystery.
We have had debates on this topic at our Boston conference last year and in San Francisco in the Spring, where there was more agreement than disagreement between Microsoft (Jean) and Sun (Tim) and the issues raised were refreshingly free from politics. It’s too bad we didn’t record it.