WolframAlpha explained: The world's most academic search engine
One of the most ambitious search projects has just launched. WolframAlpha is the product of British physicist Stephen Wolfram. We've been investigating what it does, how it does it, and if it's any good.
92,518 days ago on 27th January 1756, with just 92.90 per cent of the leap year 1756 left to tick, Wolfgang Amadeus Mozart was born. It was a Tuesday. The moon was a waning crescent, and the sun shone for eight hours and 53 minutes until setting at 4:42pm.
Quite the historical paragraph, huh? Thank WolframAlpha -- the Web-based "computational knowledge engine" from British physicist Stephen Wolfram. Its simple interface belies the power behind its doors, however, and Wolfram's motivation differs notably from the organisational drive of Google.
In reasonably simple terms, Wolfram Research takes over 10 trillion pieces of data from vetted sources -- what time the sun rose on a particular day, or the boiling point of a chemical for example -- and feeds it into WolframAlpha's supercomputer-powered database.
Thousands of different algorithms and search technologies then grind away to pull together all the different pieces of information into a single page. These often comprise tables, graphs, charts and diagrams to explain whatever query you asked the engine to compute and spit out.
'The year Mozart was born' was our question, and the document presented gave us the data for our opening paragraph. What's good is that it didn't link us to a Web page, but rather created one. And it could be taken away as a simple downloadable PDF document.
What about 'the life expectancy of the UK population in 2009'? That'll be 79 years, putting us in 26th place for longest-livers worldwide, apparently. How about calculating Pi to 39,000 digits? Pfft, a walk in the park. What about, 'Will Dawn Porter go for a coffee with Nate Lanxon?'. Ah, no, that stumped it.
So who's it for, and what can't it do?
Researchers, scholars and students seem to be prime candidates for reaping the benefits of Wolfram's work. So we asked Andrew Hoyle -- a psychology student at the University of Lincoln -- to use Wolfram for some research.
He had mixed experiences, finding useful results for questions relating to chemistry and biology, but fewer for his subject, psychology. "I looked up details about the neurotransmitter noradrenaline for example," he told CNET UK, "but Wolfram explained only its chemical composition, rather than its effect on the human body, which is what I'd be looking for."
"I think for subjects often of a more theoretical nature, including psychology, searching peer-reviewed journals and medical documents, will not be replaced by Wolfram right now. Though I may find it useful for extra 'nuggets' of information that may otherwise have been missed."
The future
Wolfram seems to leave theory and conjecture to Google's archive of journals and documents. Indeed, a search for 'Is there a God?' produced an error page reading, "Human Discourse -- Additional functionality for this topic is under development," suggesting the lack of an answer for this question, rather than a plain error page as usually seen for failed queries, is only temporary.
As beautiful and computationally impressive as Wolfram's engine is then, it currently has its limits. But unlike Cuil, it's not aiming to replace Google, so its shortcomings shouldn't hinder its acceptance in the search space too much. And maybe, just maybe, it'll soon lend its computational horsepower to expand into realms beyond that of scientific fact, and into that of glorious unproven metaphysical debate.
Maybe it will even make some money.