There's been a lot of buzz about WolframAlpha, a new search engine that's not really a search engine. It's been touted in some circles as the next big thing in search.
The brainchild of British computer scientist Stephen Wolfram, the WolframAlpha has the daunting title of "computational knowledge engine."
WolframAlpha takes a specific question, searches its knowledge base for the information required and then uses computation to arrive at an answer. It then displays the answer in an organized way, including grids of facts and graphs.
To do this, WolframAlpha must understand your question, have a vast reserve of facts, which its developers have been entering into a database and know how to display the answer in meaningful form. All that is no small feat.
WolframAlpha is sort of a cross between Google and Wikipedia with differences from both. Like Google, it searches for an answer to a query, and like Wikipedia, it searches through a database of information that has been entered.
Unlike Google, it does not search the Web. Unlike Wikipedia, the facts have been compiled by experts, not volunteers, and the answer does not return in the form of an article. And unlike either, it applies computation to the facts it finds.
Here's an example Mr. Wolfram gives on the blog (blog.wolframalpha.com). Suppose you asked how far the moon is from the earth. Google and Wikipedia would return the traditional answer as part of a Web page or an article.
WolframAlpha, however, would know that the moon's path around the earth is elliptical, so the distance between the two is constantly changing. It might give you the textbook distance, but then do a computation and tell you how far away the moon is right then. Those two answers could be 10,000 miles different.
If you want to get good examples of queries on WolframAlpha, there is a screencast on the blog.
One example -- entering "2 cups OJ" will return nutritional information like you would see on the side of the carton, a breakdown of the vitamins and a nutritional comparison to other foods, among other data.
But there are uncertainties about WolframAlpha. We don't know the size of its knowledge store. The team behind it says it is a work in progress so its database will grow. But some early indications are that there are obvious holes in its store of data.
We do know that WolframAlpha was scheduled to launch late last week with an official date to go public tomorrow at www.wolframalpha.com. We also know that it is free.
The team planned to Webcast the process of taking the site live on the blog. Webcasting began Friday night and was to continue all weekend.
Mr. Wolfram goes into this endeavor with a reputation. He is well known for developing Mathematica in the 1980s, probably still the most widely used scientific computation program. Mathematica is the basis of the WolframAlpha computation engine. And he is the author of the controversial 2002 book "A New Kind of Science."
WolframAlpha does best with fact-based questions that have specific answers. No "What is the meaning of life?" kind of questions.
Mr. Wolfram says there are five computing centers housing two supercomputers and an impressive array of other equipment. That will provide the ability to handle 175 million queries per day. But you still may experience an overwhelmed site the first few days.
So you might want to try out WolframAlpha. It may already be available today at www.wolframalpha.com, for sure tomorrow. You be the judge of whether it is the next big thing.



Comments
Welcome to Post-Gazette subscriber-only commenting community. Please review our community guidelines.