[Brian] sees semantic web technologies such as RDF and OWL as the solution to the WinFS metadata problem. That they may be, but I think they come with their own set of problems that may or may not be just as intractable as the problems facing WinFS.
I agree that the semantic web problems may or may not be as difficult as determining canonical schemas for all our metadata - past, present, and future. That’s not really saying much, though. Rather, the semantic web technologies buy us the ability to incrementally solve the problem by applying the entire world’s resources simultaneously. The best RDF schemas will float to the surface, and OWL lets vendors provide interoperability between schemas as necessary. On top of that, we get the ability to advance gracefully to new and better schemas by providing the means to express semantic equivalence between them.
As for Microsoft, they can simply release schemas for their own products. Provided they don’t completely bone them, and recent history leads me to suspect they won’t, then their large installed base ensures immediate acceptance as a de facto standard for at least half the world. The other half will begrudgingly support them anyway just to claim interoperability. And everything else can be left to the marketplace to figure out.
Which leads me back to the original point: Why is WinFS so late? I’m sure the smart people in Redmond know all of this. What’s holding them back? The world may never know.
Finally, about the Berners-Lee/Trojan Man thing: I’m just waiting for the lawsuit to roll in on that one. At least they were both knights!