There has been some real silliness recently questioning whether the semantic Web will ever happen. Simply because an idea does not get popularly embraced within five or ten years does not mean it will never succeed. Some ideas — including, I submit, the semantic Web — respond to true underlying imperatives and drivers of change.
Henry Story in a recent post on RDF and Metcalf’s law poses the interesting thesis that "the value of your information grows exponentially with your ability to combine it with new information." I could not agree more.
And that is why the semantic Web, or something very similar to it that might come to be known by a different name, is inevitable. The Web has given global humanity the access and incentive to post all information accumulated through human history online. But your glad may be my happy and there is always that leveraged value from combining stuff ("mashups") from different sources and realms. Those issues are among the classic ones in data federation and a key driver for adoption of means to resolve semantic heterogeneities. How the semantic Web and its tools and infrastructure actually develops today remains opaque. But its need and eventual use is not.
Concerning the silliness of whether the Semantic Web will ever happen, I wrote a short comparative history of technological adoption a few years ago: A Note on the history of technology adoption, where I compare html, computers, java. People think things don’t take time, because they are usually only made aware of them at the point when they are successful.