Microsoft, Powerset and Chasing the Tail of Semantic Search

We wrote about Evri earlier this week and how it falls short of the mark on being a valid semantic search tool. Now Microsoft is jumping into the quest for true semantic search by purchasing Powerset and lumping it under the Microsoft Live Search umbrella.

Microsoft has made no secret of its great desire to be more of a presence on the web in recent months. It has continued a torrid on-again, off-again affair with the idea of buying Yahoo (recently off-again, but today's rumors say perhaps it isn't as dead an idea as we hoped, though). Behaving like the inappropriate uncle who wants to buy his way into the country club certainly hasn't done Microsoft's "street cred" any favors online.

Now that their first significant online purchase out of many made this year to bolster their web presence has come to pass, you have to wonder why they have chosen to pursue semantic search. Microsoft hasn't specifically come out and said as much, but my guess would be to better compete with search giant Google, since he who gets to the heart of true semantic search capability "wins" to some extent.

I think semantic search is a pipe dream. So far it is an idea that sounds fantastic on paper, but in nearly every iteration of the idea it falls on its face. Too often the main issue is that it doesn't search enough places, or that the search is too limited in scope and direction. Many people don't want to find vague "human sorted" results, as you can't really guarantee that humans will produce the quality of content sorting you need - people are fickle.

The Live Search team at Microsoft is certainly excited about the new addition to their roster, waxing poetic about their natural language technology: "Powerset will join our core Search Relevance team, remaining intact in San Francisco. Powerset brings with it natural language technology that nicely complements other natural language processing technologies we have in Microsoft Research."

Natural language technology. Semantic search. The people driven web. These all sound like what we already have, just realized for a future web interface (this is where other bloggers might introduce a token Web 3.0 analogy. Meh.) . I'm following closely to see if Microsoft can pull this off. If they do, it will be quite the coup.