Posted by: Rob Hof on May 12, 2008
After a couple of years of anticipation, the upstart search engine-plus Powerset is launching on Monday to the public in limited form—limited not just in being in beta test but for now limited in particular to searching on the online encyclopedia Wikipedia and the open database Freebase. It’s perhaps the most high-profile of the startups looking to one-up Google with easier, more precise, or more in-depth ways to help people find what they’re looking for online. No one’s going to one-up Google as a business anytime soon. But after trying out Powerset for a little while, I do think it’s bringing some interesting new thinking to finding stuff online.
Powerset employs so-called “natural-language” search, which aims to allow people to use more natural queries instead of what Powerset CTO and cofounder Barney Pell calls the “grunting pidgin” of keyword searches. Much more than that, Powerset’s engine parses the actual meaning of words. As I explained in my story last September:
Powerset’s system will analyze the actual meaning of words and phrases that it indexes on the Web. It then will analyze the linguistic meaning of the query and find the best matches between the two—theoretically, at least, producing more meaningful results. “Our system reads every single sentence in every single document and extracts meaning from them,” says Powerset Chief Executive Barney Pell. The method is based on what’s known as natural-language processing, which Powerset developed partly in-house and partly through a licensing agreement with PARC, Xerox’s (XRX) famous Palo Alto Research Center.
Using Powerset’s method, a searcher could ask, “What companies did PeopleSoft acquire in 2002?” and potentially get more on-point search results than from typing in mere keywords such as “PeopleSoft acquisitions 2002.” Using Google, that query also brings up mentions of Oracle’s 2004 acquisition of PeopleSoft. Another advantage of the Powerset approach is it gets to the meaning of words, so queries for one word will fetch results that contain words which are synonyms. Google’s keyword searching generally won’t do that.
The much-hyped company’s launch was upstaged a bit over the weekend by rumors floating around the blogosphere that it might get bought by Microsoft, assuming Google doesn’t try to make a run at it first. CEO Barney Pell has declined to comment on that, though he can’t be unhappy about being in the middle of a potential bidding war between Microsoft and Google. That said, Ashkan Karbasfrooshan at WatchMojo.com points out that if it does sell this early—way before any revenue and even before the public has had much of a chance to try it out even in its current limited form—perhaps it is mostly hype.
I don’t actually think so, after trying it out a bit. What you notice immediately when searching on something, whether it’s in the form of a question or the usual keyword combinations, is that Powerset’s engine does in fact recognize the meaning of words. And that can make a huge difference in helping you find what you are really looking for.(I’d provide some specific examples, but the Powerlabs demo site wasn’t working late Sunday, at least not for me, but you can try it yourself now. And as usual, Danny Sullivan at Search Engine Land has a great analysis.)
Powerset also has added some serious bells and whistles to its results page, including (from its press release):
* Factz – When users enter a topic query, Powerset assembles a compact summary of interesting, and sometimes surprising Factz, extracted from pages across Wikipedia.
* Dossiers – Powerset creates a summary of information found in Freebase and Wikipedia to give you a quick overview about a topic.
* Answers – For many questions, Powerset automatically assembles an answer list from sentences in Wikipedia™ or data in Freebase.
* Semantic Highlighting – The most relevant search results are highlighted based on the meaning of users’ question.
* Minibrowser – A result can be expanded to show the snippet in the context of the full Wikipedia article.
So what is Powerset if it’s not really a search engine, as Pell takes pains to insist it’s not? “We don’t have a name for it yet,” he says. “This is actually a new category here—search engine/reader/explorer/browser.”
Hm. That can mean it’s either a really new, new thing, or the company hasn’t really figured out what its most valuable features are. In any case, I think that very uncertainty over just what Powerset is may slow its adoption, because people will need to try it out and figure out the best ways to use it. And, of course, Powerset eventually will need to prove its engine can scale to the whole Web.
So while I think Powerset has a lot of promise, it’s got a long way to go before it comes even close to challenging Google—which, of course, is already wa-a-a-ay more than a search engine itself. But it’s kind of refreshing at a time when Google’s dominance only seems to grow that some folks are still taking a stab at making search better.