Wednesday, June 10, 2009

July 19, 2006: Two Cheers for Google (or Peter Norvig)!

Candace Lombardi provided a nice account of the exchange between Tim Berners-Lee and Peter Norvig (Google Director of Search) in today's CNET News.com. This is a case where I have to side with the Google supporters' comments. Here is my elaboration on one of those comments, edited to include a hyperlink, which CNET would not let me include:

What we are really talking about here is human nature; but that is still only part of the story. What most of the techno-centric community does not understand is that every technology artifact is based in three "worlds:" the objective, the subjective, and the social. (This is a generalization of Habermas' observation that all personal interactions take place in those three worlds simultaneously.) Berners-Lee would really like the World Wide Web to be strictly an artifact of the objective world, which may be all right if you are trying to write a doctoral thesis in artificial intelligence but does not wash if the Web lies at the heart of your business model. Norvig's rebuttal addressed issues in both the subjective and social worlds, and I doubt that anyone would dispute the fact that Google's success has as much to do with their successful management (even if intuitive) of the social world as with their efforts to build the best artifacts according to criteria of the objective world.

However, beyond the subjective and social flaws that Norvig identified, the fact is that the "semantic" infrastructure of the "Semantic Web" is equally flawed. As I have tried to argue, you cannot have an ontology that is not grounded in phenomenology. In Semantic Web language, this means that the objective nature of the RDF (Resource Description Framework) will always fall short of the demands of the subjective and social worlds.

So isn't history just repeating itself? Ten years ago we had the SGML advocates who were trying to push for the formal structuring of all content. The limited amount of support for that vision was quickly blown away by the advent of HTML, and the resulting rush to search engine technology demonstrated that structure was not as important as the SGML folks wanted to believe. Guess what? NOW it is not as important as the Semantic Web advocates want to believe!

1 comment:

  1. The Semantic Web is the future of the internet, and always will be (P. Norvig 2008) Speaking as a taxonomist, the a-priori school of classification has long since been debunked using data and inference is faster, smarter and always surprising.

    ReplyDelete