Wolfram|Alpha – A number-crunching search engine

Every now and then you hear news of the launch of yet another would-be “Google Killer” – a search engine that can produce even better results than everybody’s favourite search engine.

Whenever I see such news, I’d usually roll my eyes and move on to something else. But this month will see the launch of a really different kind of search engine that can do some amazing things that Google can’t.

It’s called Wolfram|Alpha, and it’s designed to answer natural-language search queries. Now, there are already a few other natural language search engines out there but none like this. It’s not surprising that Wolfram|Alpha is one of the most hotly anticipated web offerings of the year.

Founded by mathematician cum entrepreneur Stephen Wolfram, this search engine is an attempt to be an authoritative source of answers for questions relating to some key areas like science and finance– basically any information that can be numerically quantified.

Call it a search engine for the technically-inclined, if you will. Commentators are saying it will become the preferred search engine of scientists, researchers, analysts and so on but it’s not going to replace Google for general-purpose search. And it can’t. It’s not geared to help you find, for example, restaurant listings or movie times.

So what can it do that’s so different? One example that its founder, Wolfram, offered in his blog is this: A computational search focused on finding the number of calories in a recipe. The basic data involves the calories per gram of each of the ingredients. However, turning that generic information into the actual total calories for a specific recipe requires computation.

Google or Yahoo! won’t be able to do that. That’s because search engines usually deliver results based on keywords. They aren’t designed to answer questions. Wolfram|Alpha is. It computes the answers to queries tapping onto the huge collection of data that the company has amassed.

Like Google, its secret sauce is in the algorithm. Its algorithm considers not only the words in a search query, but also what they mean when taken together. It can also perform computations involving numbers or formulas users enter. Computation is what turns generic information into specific answers, Wolfram wrote in his blog.

The math formula used in Wolfram|Alpha is something its founder developed called “Mathematica”, which has been described as a symbolic language that lets searchers express complex computational processes in a fluid, intuitive way.

That sounds like vaporware to me but Wolfram|Alpha has gotten good reviews from its beta testers.

It’s worth mentioning that Wolfram|Alpha is different in another way. Unlike Google or Yahoo!, it does not even gather data from the Web. Instead, it has its own databank of information that its researchers have compiled over the years.

Bottom line: Is it a Google Killer? The simple answer is no. The more nuanced answer is that it’s better in some ways but inferior in others.

Because it’s not a traditional search engine that taps onto the web, there are many things that it can’t answer well or perhaps can’t answer at all because it doesn’t have the necessary data in its knowledge bank. It’s not a general-purpose search engine. But for certain things, it will provide remarkably precise and detailed answers.

Wolfram|Alpha is expected to be launched publicly on May 18 at www.wolframalpha.com.

Oon Yeoh invites you to read his blogs, join him on Facebook, follow him on Twitter, view his Flickr photos and watch his YouTube videos through www.oonyeoh.com.

Comments

[Sivin Kit]

I'm pretty excited about it.

[JQ]

Hi Oon, enjoyed your column. It's a disservice to call Mathematica vapourware however, considering that it was initially released in '88 (albeit the algos are / will be continually revised), and widely used in universities. Thks.