Berners-Lee vs. Norvig

Alright, my 2 cents on the Berners-Lee Norvig files*: Peter Norvig did not say that the Semantic Web cannot work - he just pointed out that there are serious challenges that must be overcome. I happen to totally agree: there are still big questions that must be solved, its not all smooth sailing from here on. The adoption of the Semantic Web is not a foregone conclusion. In fact I believe that Peter Norvig did a great job of posing the big questions for the Semantic Web that all to often get ignored (at least in academia): "how do we make this stuff usable", "how should this work when >90% of the availlable data is SPAM" and "whats the business model". He apparently hasn't seen convinving answers to these points - neither have I.

For me the trust question is the most important of these. I don't think that pointing to "logic" and the great big trust layer is a satisfactory answer to this challenge - it needs to be lot more concrete than that. And - dare I say it - a little bit of scepticism may be in order when people talk of building a giant system ignoring trust questions, counting on the "trust layer" that will be added on top of everything and will magically make all trust and security questions disappear. And talking about ignoring important questions: may I point out that in this year neither the ISWC, the ASWC nor the ESWC had/have any workshop that dealt with this question?

So, in the end I would add a fourth question: "how do we actually do open web-scale reasoning and what exactly will this mean" - but other than that: "Great Analysis, Mr. Norvig!"

*: A good overview of the debate is here, and here is further interesting post.