Finding Answers: Compute vs. Search
Stephan Wolfram recent announced his latest endeavor: Wolfram|Alpha. Where Google will search the Internet to answer your queries, Wolfram|Alpha will calculate the answer. Where Google gives you 287,000 results for "what is the longest straight road in America," Wolfram|Alpha will give you one: the answer.
At least, that's the plan. We won't really know until May.
The difference between the two approaches was made evident to me when my daughter recently asked for help with her math homework. It's been a (very) long time since I did anything with quadratic equations and I admit to being somewhat lost by her first question. Fortunately, Google quickly pointed me at several quadratic equation solvers, graphers and just about anything else remotely related to quadratic equations. But if I typed an actual quadratic equation into Google, it was mystified.
When someone asks me a question I don't know the answer to, I often quip, "Google knows." It's indexing spiders have been crawling the web for years -- no public Web page hasn't been visited by them and placed somewhere in Google's vast Index Of Everything. But Google doesn't actually know anything. It just knows where to find things that other people know. It has meta-knowledge, just like all other search engines out there. Sure, it's better than other search engines, but that's because they've optimized the "meta" part. It's not like Google has access to more knowledge than other engines, they just index it better.
What Wolfram|Alpha is trying to do is calculate the answer to your question. If it does as advertised, it will have actual knowledge because it will answer your questions based on known first principles and it's own calculations. It will be able to extrapolate from knowledge others have published on the Web. It will add to that knowledge, thus it will know. We're not talking a HAL 9000 -- it won't be sentient -- but it will add to the collective knowledge of the Internet, not just provide a better map.
Some have claimed that it will be a Google-killer, other say it's just (yet another) over-hyped semantic-web-based search engine. We won't really know until the über-secret veil is lifted in May. If all that Wolfram is saying about his latest baby is true -- natural language query parsing, computation of answers based on facts and the entire Internet to use as the foundation of this knowledge engine, as Wolfram calls it -- it could be game changing.
Though people said the same thing about the Segway...
Post new comment