The Semantic Web was recently brought to my attention after reading an article by Clay Shirky, previously mentioned in this blog. As usual, I agree with much of what Shirky has to say. But apparently, he ignited quite a discussion, including this rebuttal and the further discussion here.
In the terms that Shirky originally states, I have to agree with him. It does sound like proponents of the Semantic Web are vastly oversimplifying the problem. Even after reading the further discussion, I’m not sure such proponents disagree with Shirky. They provide as counterexamples very simple use cases, such as indexing books, where the information structure is already well-defined. If that’s all they mean by the Semantic Web, then sure, that makes sense. But Shirky makes a good point that extending such a concept to more complicated scenarios is going to be an extremely difficult problem, where the Semantic Web folks are trying to gloss over the complexity by concentrating on the parts they already know how to do.
It sounds a lot like the debates on artificial intelligence to me. The AI researchers make grand statements about the future of AI. Critics say it’s going to be a lot harder than that. The researchers show how they can do all these cool things in environments of limited scope (like chess-playing), without realizing that such demonstrations imply absolutely nothing about their ability to handle the general case. For the Semantic Web to be nifty cool, it’ll have to be able to handle general implications on its own. And that will be tricky. In the meantime, the only metadata that is generated will be that which somebody has figured out a use for. The Semantic Web tools may make it slightly easier to deal with that metadata, but it won’t be able to extend it by creating its own metadata, for instance.
I don’t know. I thought I had something interesting to say about this topic, but maybe not. I’ll still post this, because I think the links and discussions posted above are interesting, but I’m not going to spend any more time on extending out this post.