SEARCH ENGINES
MILLIONS of web users have Google’s search engine as the homepage on their computer. It’s the first thing they see when they switch their machine on and their first port of call whenever they need to find the answer to a question.
Such is Google’s dominance of the search space that any attempt to launch a rival platform is often met with a mixture of bemusement and sympathy. The problem, as many companies quickly discover, is that it’s almost impossible to beat Google at its own game. Instead, what’s needed is a fresh approach to search which can supplement Google while also carving out its own niche.
That’s the strategy taken by Wolfram Alpha, a search engine that launches on Monday. It’s the brainchild of Stephen Wolfram, a British physicist and former child prodigy who, in 1988, created the Mathematica computer program used by scientists, researchers and engineers to number-crunch huge volumes of complex data.
Wolfram Alpha is designed to provide clear answers to specific questions. For instance, a Google search for, say, the nutritional values of a Big Mac returns 123,000 results, which web users then have to sift through to find the facts they want. A similar search on Wolfram Alpha will, in theory, return a single answer, detailing the calories, fat content and salt levels in the McDonald’s burger.
That’s because Wolfram Alpha doesn’t crawl the web for its answer, as Google does. Rather, it searches its own huge database of factual data to find the correct information.
“Fifty years ago, when computers were young, people assumed that they’d quickly be able to handle these kinds of things, and that one would be able to ask a computer any factual question and have it compute the answer,” says Dr Wolfram. “But it didn’t work out that way. Computers have been able to do many remarkable and unexpected things. But not that. I always thought, though, that eventually it should be possible. And a few years ago, I realised I was finally in a position to try to do it.”
Indeed, says Mr Conrad Wolfram, Stephen’s brother and the project’s strategic director, creating the database that underpins the search engine has been a “huge undertaking”.
“It has taken three to four years to generate this database,” he says. “But because we use Mathematica to help structure this information, you could say we’ve been working on it for even longer.”
The secret, says Conrad Wolfram, is in the value that’s been added to this data. “We believe in human expertise, which is very different to traditional search - that relies much more on computer science to pick out results,” he says. “The database is a structured environment in which we’ve added meaning and computational knowledge to the data. This allows it to come up with an answer to the question a user asks.”
Much of the hype and speculation surrounding the launch of Wolfram Alpha has framed the website as a potential “Google killer”. But Wolfram Alpha is designed to be “highly complementary” to other search engines rather than a direct competitor, says Conrad.
“If you use a search engine, it’s a bit like a librarian who gives you a whole set of papers that you need to look through to get the answer. If you use Wolfram Alpha, it’s more like having your own personal analyst who takes the question you’ve got and tries to come back with a specific set of answers unique to your question.”
Wolfram Alpha, he says, will grow in accuracy and scope as more people use it. He says that experts will analyse aggregated, anonymised data about searches performed and the success rate of results, to learn how people are using the search tool and find ways of improving the service.
Drawing relationships between complex data has long been a holy grail for web developers. For years, companies have been working on building a so-called “semantic web”, in which data contained within webpages is labelled with keywords, or digital “tags”, to associate meanings to information and enable it to be interpreted in a wider context.
But that’s a difficult idea to implement. “It’s hard to get people to tag their data in this way, and no consensus has been reached on what tags to use, says Wolfram. “Our approach is different - to suck the data in so that we can structure it, curate it, tag it with meaning and make it computable.”
It is here that Wolfram Alpha could really make a name for itself, becoming a “destination website” for students and researchers seeking an easy way to distil complex information into manageable chunks, which can then act as a springboard for further research and discovery.
Even Google, the old master of web search, has seen the potential of this, launching a tool called Google Squared that allows users to type in a specific query - such as London boroughs, or United States Presidents - and receive a list of information that relates to that search term. For example, the list of London boroughs aims to list all boroughs alphabetically, along with links to relevant websites and information about population and the leader of the council. The results are presented in a grid, and users can remove columns of unwanted information or add a column of data they would prefer to see.
Wolfram Alpha, it seems, is ruffling a few feathers. “We’re at the start of a very long project,” says Conrad Wolfram. “We’re not going to have everything right to start with, but I think people will find Wolfram Alpha a very powerful system, and it’s going to get a lot more powerful still. Many new things will come over the months ahead.” THE DAILY TELEGRAPH
From TODAY, Tech; Friday, 15-May-2009
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=70c60853-52d8-4ee7-9ef1-854147e57715)
No comments:
Post a Comment