I don't agree that search is one of the keys to the semantic web. Search is needed when you have a corpus of millions of unstructured documents that might contain the information you are looking for. The semantic web, for me, is more about finding the actual answer to a question, i.e. getting it directly. Search is more about an indirect approach: you look for a document that you then have to read and understand to find your answer. Also, most of the benefits of the semantic web will be unseen at the infrastructure and technology level. That raw data will be used behind the scenes by applications to subtly improve the user experience: the application will know more about the things that are important to you and will be able to act on them.

With semantic indexing, the namespace takes over from the keyword. Page rank is a linking algorithm which extracts a mathematical meaning but with RDFa you get better metadata, less ambiguity, richer connections and a better-cleaner rank/orbit.

So a semantic arms race ensues and everyone includes (dirty) vocabulary on pages, posts, updates and feeds and the search engines index it {Fanfare} Semantic web in 12 months.

Meaning follows a little way behind - jostling for a place with advertisers, brand managers and spammers. The geolocated, the now, the personal and the semantic - four riders of the apocaweb

A service like this already exists - take a look at Freebase. As far as I know it is not indexing the web, it works more as a semantically-enhanced Wikipedia. I think the "search" experience is very interesting, and I hope to see more of this in the future. I'm not very thrilled about the "finding the actual answer" approach (= Wolfram Alpha), I'd rather see the facts in a user friendly way and then draw my own conclusions.