Yesterday, The Globe and Mail had a piece about a new company that’s trying to change the way we search online:

SAN FRANCISCO — Powerset on Sunday unveiled tools for searching Wikipedia that use conversational phrasing instead of keywords, marking the first step of its challenge to established Web search services such as Google.

Powerset’s technology breaks down the meaning of words and sentences into related concepts, freeing users from always needing to type the exact words they want to find.

The idea is that instead of breaking your query down into keywords as you would for Google search, you just ask your question. Powerset will figure out the main concept behind your words, and then compare it to its index of other phrases that mean similar things. This sounds great, especially for the possibilities this software could give us when combined with text to speech to just let us talk to computers. Another neat feature is that Powerset uses its index of meaning to generate a straightforward ‘list of factz’ for every article.

But how well does it work?

Not badly, though my biggest challenge was to stop thinking in keywords. Here are some of my searches, some of which went well:

As did others:

Some of the answers, however, seemed to be a bit literal:

By and large, not so bad.

I also enjoyed the ‘factz’ summary of some of the pointless articles a bit too much, it was on the Star Wars lightsaber page that powerset informed me that “lightsaber emits hum”, and that “Sith Lords threw lightning” — niiice.

While those facts are pretty pointless, they are (so far as I know) correct. So the software is capable of, at a pretty rough level, figuring out what words have significance in the context of the article.

At this point I don’t see this as a service that will replace Wikipedia for me, I’m used to breaking things down into keywords, and I think I’m pretty good at it. All the same, this seems to be a pretty crucial step forwards in making computers easier to use and information more useful.