Deconstructing the Google Knowledge Graph
Mike Bergman has some interesting observations on the Google Knowledge Graph, first on its coverage and then on how it is constructing URLs for nodes in its graph.
I have to second his call for Google to release its identifiers via an API. That would be a real boon for common entities.
I say common entities because having “millions” of identifiers is fairly trivial when you consider the number of objects captured every night by optical astronomers alone. Or sequencing genomes.
Not to discount the value of a common identifier for Lady Gaga but uncommon entities need identifiers too.
Gabriel Hopmans pointed me to this post. (Morpheus)