In principle this is a great topic. I got some interest in the semantic web technologies in early 2000'.
My conclusions were that:
- The approach was extremely rigid, basically all info (data and metadata) had to be classified mostly in a hierarchical way (first order logic) with only a few exceptions when most real life cases were unable to fit in this scheme.
- Companies in Semantic Web just sold search. Classification was often done by hand or by "connectors" that had to be chosen or even tailored by hand.
A few years later search technologies were much more powerful and AI enabled impressive results in classification.
