New Web tool WolframAlpha launches test run

While not a traditional Web search engine, a challenger to Google emerged on Friday -- WolframAlpha, named after the man behind the venture, British-born computer scientist and inventor Stephen Wolfram.

Wolfram, who earned a PhD in theoretical physics from Caltech at the age of 20, is careful not to call his latest invention a search engine, describing it instead as a "computational knowledge engine."

Unlike Google, which takes a query and uses algorithms to scour the Web and return a series of links to relevant websites, WolframAlpha.com takes a query and crunches through its databases to return answers.

"The basic idea of WolframAlpha is very simple," the 49-year-old Wolfram said in an online presentation of his venture, which went live for a test run on Friday. "You type your input, your question and WolframAlpha produces a result."

WolframAlpha had been scheduled to kick off at 8:00 pm Friday (0000 GMT Saturday) but was late out of the gate because of what was described on the company blog as "some kinks."

"We got off to a late start, but so far, so good," it said.

Danny Sullivan, editor-in-chief of SearchEngineLand.com, said in a telephone interview with AFP that WolframAlpha is a "really interesting tool."

"I try to describe it as a 'fact search engine' to help people understand the degree that it's different from Google," he said. "Google tends to point to stuff while (WolframAlpha) actually have some answers."

He said WolframAlpha was not presenting itself as a rival to Google "although they want to capture the general search audience too.

"They're saying they're not trying to wipe out Google but they feel they do the kinds of searches that Google doesn't handle," Sullivan said. "If you're trying to get facts this might be a handy kind of encyclopedia for you."

However, Sullivan said WolframAlpha "can be kind of finicky."

"It doesn't have answers to everything that you might try," he said. "So you tend to get sort of a dissatisfied feeling if you've done your search and it comes up with nothing for you."

Another problem with WolframAlpha may revolve around sourcing, he said.

"Anytime you do a search they'll tell you where the data has come from and where they're pulling it from," he said. "But, you know, sources from all sorts of places can be wrong.

"So there will still be that issue where some people may feel like 'Hmm ... I don't know if I want to trust this,' in the same way that they don't want to trust Wikipedia sometimes," he said.

Wolfram said WolframAlpha is a "long-term project."

"We're trying to take as much of the world's knowledge as possible and make it computable," he said. "So that anyone, anywhere can just go to the Web and use all that knowledge to compute answers to their specific questions."

Sullivan said WolframAlpha was the most amibitious search project in recent years outside of Microsoft's unsuccessful efforts to steal market share from Google.

"It's a fairly large project they've put together," he said. "They've got like 150 people."

Google dominates online search in the United States with a more than 64 percent share of the market followed by Yahoo! with 20 percent and Microsoft with less than 10 percent, according to industry tracking firms.

Sullivan said that as impressive as WolframAlpha may be, Google just might still be one step ahead.

He noted that the California company earlier this week unveiled a laboratory project called Google Squared which it plans to release later this month.

"What it does is actually trying to go beyond what Wolfram is doing," he said. "Wolfram is gathering information that's been all nicely and neatly put into databases and spreadsheets and everything.

"Google Squared is trying to find information from across the Web and automatically build those kind of spreadsheets for you."

Read PhysOrg.com story: Wolfram Alpha 'Knowledge Engine' is Like a Modern Farmer's Almanac

(c) 2009 AFP