Before the official launch of much hyped Wolfram Alpha “Computational Knowledge Engine”, there was talk that it was the next Google or Yahoo, or that it may even be a Google killer. It could change the way we find information on the web. But that has proven to be far from the case. In fact, for many, using the site first hand seems to have produce more frustration and head-scratching than actual results.
In case you haven’t seen or heard of Wolfram Alpha, you should check it out. It isn’t a search engine – it is a self-described “computational knowledge engine”. And that is where the problems begin. Much like Google or other search engines, Wolfram Alpha’s interface is centered around a simple text field where you enter your query. But instead of searching the web for pages related to the search terms, it returns factual results from it’s own internal knowledge base. Or doesn’t, as is far more often the case. From my own experience, and what I’ve heard from others, Wolfram’s failure rate far exceeds the instances in which is actually returns any results at all. Because it’s a closed system, Wolfram Alpha can only find information in its own database. If the Wolfram team doesn’t know it and hasn’t entered it, the system won’t find it. Compare this with a search engine, which taps into the vast wealth of information that has been published by millions of people on billions of web pages. The probability of finding the answer becomes so much greater with a resource like that.
And that leads to the second major problem with Wolfram Alpha – it is the opposite of the social web. By internalizing all its content and seeking answers only from what is contained within its own internal data stores, Wolfram Alpha is moving against the crowd sourcing, social aspect in which most of the rest of the Internet is moving. The problem solving and information gather power of crowds is much more powerful than that of a few, and unfortunately, Wolfram seems to be relying on a few, while search engines rely on the masses. It bucks the wiki model of cataloging knowledge and bringing back the old model of the encyclopedia. It’s creators seem to be looking to fictional knowledge systems such as the Hitchhiker’s Guide to the Galaxy or EMERAC for inspiration rather than the modern state of collaborative knowledge generation and sharing.
And that leads me to a third problem with Wolfram Alpha – is it actually something that people want? It claims that it is not a search engine, and after using it for a few minutes that becomes painfully obvious. It does do a good job at quickly delivering factual data, if that data is contained within its knowledge based, that I’ll give it. But search engines, though they are not perfect by any means, deliver content that is far more diverse, deep, more interesting and more useful than the limited body of facts, figures and data that is available in Wolfram. Imagine if you combined the computational power of this new tool with a search of the world wide web to answer a query. That would be a powerful way to search the web, and would truly be a huge step toward the realization of the semantic web. Until then Wolfram Alpha seems best suited for limited fact checking, mathematics and a bit of geek humor (and more geek humor).