Does anyone actually want semantic search?

Lots of new semantic search engines now. Never mind that it's not entirely clear what semantic search is, what is clear is that liberal use of the word 'semantic' is enough to gain some interest and, in the case of Powerset, to get bought up by Microsoft for integration with Live search. But there's no evidence that users want semantic search, as the recent puff about Wikipedia issued by Powerset confirmed.

The latest US search engine usage data provides a good starting point for seeing if Powerset can put a dent in Google's search engine market share. Michael Arrington passes judgement on Powerset-Live integration to date. Plenty of lively comment follows that post.

Here is an example of the great leap forward arising from semantic search. What caused the great depression? Amongst the causes identified by Powerset it:

was partly caused by the after-effects of the Stock Market Crash of 1929 in the United States, and partly by internal factors in the Netherlands...

Even as an early days demo this isn't good enough to win over users. The problems for Powerset in trying to differentiate around question answering are twofold. Firstly, for simple questions, non-semantic search engines will get a reasonable answer - how tall is everest? for example. Secondly, most questions don't have simple answers. What caused the great depression? Well, lots of things. A non semantic search engine will get you close enough to begin exploring the causes of the great depression; semantic search doesn't seem to add any value and may even make things worse.

Related Post:

Widget by [ Iptek-4u ]