It's Like Déjà Vu All Over Again
"You could probably waste an entire day on the preceding links alone. But why take chances? We also give you Paul Snively..." — John Wiseman, lemonodor
Tim Berners-Lee speaks at Stanford tonight about the Semantic Web. BTW, I think I finally got the vision. It goes like this. There's data in this blog post. It relates at least three things. 1. Tim Berners-Lee. 2. Stanford. 3. Semantic Web. 1 will be at 2. Link to Semantic Web from the arc between them. Link to www.stanford.edu from 2. With this information properly stored in a relation, you could query it. Now, I gotta figure out why this is different from Google, which probably already does a pretty good job of extracting all that meaning from this post. Postscript: TBL is not speaking at Stanford tonight. Not sure what that says about the Semantic Web. [Scripting News]
Yup, you got it in a nutshell: the Semantic Web is about making multiple types of relations explicit and providing them a simple, foundational semantics. And you're right; the human brain is quite capable of extracting this meaning from your post, which a human being can find via entering "tim berners-lee stanford semantic web". But the point of the Semantic Web is to make the Web machine-understandable. The tools to make it human-understandable are already pretty good, an effort to which you've made huge contributions. TBL is just trying to bring the machines to the party now.
Just to offer a tiny, localized example, thanks to an RDF-based tool I've been experimenting with at work, I can enter some new observations about a model of car, and once I've specified its make, that model will also show up under queries for "European" or "Asian" or "German" or "American" or "British" vehicles, because the system knows the nationality and continentality of all of the makes.
That's a poor example, because it'd be pretty easy to do with straight relational database technology. RDF and the Semantic Web only become interesting when you get into relationship types besides set containment and subclass/subproperty relationships.
8:37:10 AM