WolframAlpha - should Google be quaking in its boots?

I have spent the weekend playing around with the new search engine project WolframAlpha.
After months of preview hype, the new "computational knowledge engine" had a rolling launch on Friday, accompanied by a couple of hiccups mostly to do with dealing with the amount of traffic heading its way, and the service launches officially to the public tomorrow.

Is it any good and will it fulfill the breathless talk of it being a Google Killer? The answers are emphatically yes and no.

The site itself is pleasingly clean and simple. One search box for your query. You enter the terms and click return.

If WolframAlpha (WA) has the data to compute an answer for you, the results are quick and impressive.

Type in "London San Francisco" you get the city population and the metro population for both, the distance between both cities, the flight time between them, a map, the local time in each city and the approximate elevations of each place. For example:



It is slick and looks good.

Another example - I typed in "position of International Space Station". This is what I got:

Underneath the simplicity, lies some serious computing power. This computation, by Wolfram's Mathematica software, requires differential equations for the motion of the craft in Earth's gravitational field. When I tried to do the same on Google, results pointed me to a NASA website but the orbital tracker was unavailable. I then got lost trying to find an alternative in the maze of pages.

This is WolframAlpha at its best. It takes serious scientific data and produces answers that simply don't exist as easily or as accessibly elsewhere. It can do calculus, regression analysis and gene coding.

In short it is the best calculator on the web and it is free for everyone to use. Scientists, engineers, and math fiends are going to flock to it, I am guessing.

But it will not threaten Google's position as the search engine of choice for most of the planet. Because for all its massive databases, it falls down when you get into anything cultural.

It will not tell you the best MP3 player to buy or where to look for one. It does not tell you the latest news on MPs and their expenses claims. It has massive holes in data - most of its coverage of people gives simply a birth date, birth place and a rather useless timeline.

"Muhammad Ali" brings up:


and that was it. No details of his fights or titles. No other biographical detail.

"Don Bradman" brings up: "Wolfram|Alpha isn't sure what to do with your input."

WA acknowledges these weaknesses with links to Wikipedia and to ordinary web searches with Google, Yahoo! and Microsoft Live from every result.

Admittedly this is early days for the service. Stephen Wolfram, the British-born scientist behind WA, has said that there is much work to do and many more databases to compile. He has 150 people working for him and it will take time.

But Google is already working to blunt the impact of WA. It is about to launch a spreadsheet feature Google Squared shortly, which in effect is a very populist version of WolframAlpha.

The existence of WolframAlpha points the way to a future of supercomputing power that should excite everyone, even if they do not end up using it. It is beginning to get at the sort of role that science fiction writers have always envisioned for computers ie HAL in 2001: A Space Odyssey. But WA's "intelligence" is still severely limited.

What will I use WA for? I liked the basic financial information pages for companies which set out some nice graphs and fundamentals. For instance "Apple" returns a whole host of information that do not come with the simple Yahoo! finance pages such the daily returns versus the S&P 500:


Impressive. Economists will find WA increasingly useful, I suspect.

There are various downloads for your toolbar or widget add-ons if you want to put WA in your browser etc. Go here.

My favourite result so far is the one from the question "does God exist?"

BY Mike Harvey



Copyright 2008-2009 Daily IT News | Contact Us