Stephen Wolfram is a British physicist perhaps best known for his work in creating the computational software Mathematica. That could all soon be forgotten if his Wolfram Alpha computational data engine proves to be the Google killer that some are suggesting. After all, it claims to be able to answer any factual question whether that answer can already be found in currently indexed documents or not. But then again, Cuil claimed that it was the biggest search engine on the planet right from the get go and then mixed these claims with more hype about unique content-based relevance methods for the engine. The media ended up in a feeding frenzy with reports suggesting that Cuil was a Google killer, that by going beyond link analysis and traffic ranking it was the ideal search engine.
A former Google search index architect, Anna Patterson, along with a Stanford University search research pioneer, Tom Costello, were the brains behind Cuil and threw promises of "leveraging our expertise in search architecture and relevance methods" to the hungry press. Of course, a hungry Internet population were quick to take a look when Cuil launched on July 28th 2008, and it seemed to live up to the promise by grabbing 0.26 percent of the global search market in just 48 hours. Maybe not quite a Google killing effort, but a pretty impressive one for a search start-up nonetheless. Unfortunately for Cuil the reality did not live up to the hyperbole, and testing here at DaniWeb quickly concluded that "overall there is little doubt that Google delivered upon my search requirements in a much more effective manner. Indeed, Cuil only managed to find one out of the three pieces of information I was after, at least as far as first page hits are concerned." By the end of August 2008 Net Applications figures showed the Cuil market share had fallen to just 0.01 percent.
So can Wolfram Alpha fare any better?
Good question. Considering the amount of hype surrounding it, some might say it is right up there with the whole 'Ginger' publicity madness back in 2001 that promised a product that would change the world and which ended up being the Segway.
You see Wolfram Alpha can promise all it likes, but ultimately it has to deliver. Which is where things get a little sticky, because at the moment Wolfram Alpha remains firmly behind closed doors and not open to public scrutiny. There is no demo, there is no public beta of the alpha. There is a statement from Stephen Wolfram himself which tells us that he is "excited to say that in just two months it’s going to be going live." But two months is a long time in the marketing of Internet technology, and to be honest I want to see something to back up the claims of being able to "ask a computer any factual question, and have it compute the answer."
At the moment all we have to go on is the explanation of how Wolfram Alpha works from the man himself. Wolfram reckons it just comes down to being able to "explicitly implement methods and models, as algorithms, and explicitly curate all data so that it is immediately computable." Easy huh? Well, no, not really, as Wolfram himself admits. "Every different kind of method and model—and data—has its own special features and character."
Wolfram has used a combination of his Mathematica tool to provide a symbolic language that can represent anything, plus the algorithmic power needed for any kind of computation, and the concepts of NKS to produce this equivalent of the Star Trek computer. NKS? That's a New Kind of Science, an empirical study of very simple computational systems that was the focus of Wolfram's work for many years. Wolfram reckons that NKS provides "a paradigm for understanding how all sorts of complexity could arise from simple rules." He admits that getting computers to deal with natural language is incredibly difficulty, but actually not needed if knowledge has already been made computable. "All one needs to be able to do is to take questions people ask in natural language, and represent them in a precise form that fits into the computations one can do" he explains.
Yet Wolfram insists that a mixture of "many clever algorithms and heuristics, lots of linguistic discovery and linguistic curation, and what probably amount to some serious theoretical breakthroughs" his team has made it work.
In a couple of months we will be able to find out, I guess, by popping over to www.wolframalpha.com and entering a question in the Google-alike simple text input box and seeing how the "huge system, with trillions of pieces of curated data and million of lines of algorithms" really can provide a new paradigm for using computers and the web as Wolfram suggests.