ReadWriteWeb takes a first look at Alpha, a new “computational knowledge engine” based on Stephen Wolfram‘s Mathematica, poised to go live within the next weeks. Though touted as a potential Google killer, it is in fact a different sort of beast. Rather than constantly scouring the web, it is built around curated data from various sources, which is then subjected to Natural Language Processing algorithms based on Mathematica code. The search interface is seemingly similar to that of Google et al. whereas the results stand out, providing not only proposed answers but also an array of related information. The results will also present links to the sources used for the answer.
ReadWriteWeb’s article is based on a demo which featured number-crunching searches and replies, so one could hypothesize that Alpha will be mainly geared towards such queries. It will be interesting to see how well the notion of curated data and well-crafted results plays out in real life.
For more snippets of information, have a look at The Guardian’s preview of Alpha, or get an extensive overview and analysis at CNET.
One reply on “First impressions of Wolfram’s Alpha”
Wolfram is forgetting that any new search engine needs the cooperation of webmasters and admins, in that they need to allow its search spider into their sites. Unfortunately, a great many have blocked all search engines apart from Google from indexing their sites due to the immense amounts of bandwidth used up by these leeches. And this, basically, is why all these new search engines launched with fanfare and excitement, eventually fade away and are never heard of again.