Reuters and Calais
Big news for the Semantic web! While it is in itself nothing too fancy, and short of revolutionizing the whole information scene, Reuters’ new acquisition (ClearForest) has released the OpenCalais API. What it does is to turn “supposedly” non-structured data into that infamous RDF format.
Basically Calais extracts entities (name of a company, place in the world, …) and events/facts (merger, acquisition, departure, …) and tentatively organize them into some sort of structured graph.
I have just started to play with it a few hours ago and my preliminary impression is not a “wow” but rather an “hey, not bad”. Calais does indeed detects company names within web-pages and somehow realizes that “internet connection” has something to do with an Enterprise Term and hence aught to be in relation with Google.
I have already stated my opinion on the Semantic Web initiative. Yet short of an overwhelming applause, ClearForest’s new technology does an impressive job for an early stage system. I’ll surely keep an eye in this direction.
The semantic-web initiatives might not solve the underlying problem of semantics, but tools and technologies such as Calais are certainly pushing the limits forward.
