Wolfram research is always doing some interesting things- and now they are aiming at providing an answer machine- they are calling it Wolfram Alpha. Its not a search engine that returns documents related to the inputed search terms- but something that computes answers to a question looking for a factual answer.

This is interesting, because often when we are searching for something via, say, Google, we are actually looking for an answer. We do it in two steps- “Country population list”- which gets us the document, and then we look up the countries we are interested in.

Unfortunately, the way Wolfram Alpha was launched and the way the media and observers in general tended to react has created a fair amount of hype, and misconception. Although Wolfram Alpha (lets call it WA) will have a natural language interface, people always get carried away with their expectations for such things. I’m certain that WA will be impressive, but I’m equally pretty sure that you won’t be able to say “Roughly how many people like to have peanut butter on their toast in Ohio” and get a reasonable answer.

In a recent interview with Rudy Rucker, Stephen Wolfram said that rather than sell WA to the search engines, “We’d rather look for things like partnerships or licensing deals or APIs. I see a new field of knowledge-based computing. Imagine a spread sheet that can pull in knowledge about the entries.”

Now this is really interesting. What if it has a way to ask questions like “what is the GDP of [country]” just like that? What if it can tell you the population of any given Zip Code? What if it knows the rate of income change by county? What if it can tell you for any geo-location/date if it was a statutory holiday or not on that day? These are things that could be very very useful in doing data analysis- and are the kinds of things that can build interesting real world dimensional tables in a data warehouse or data mart.

There are of course, various sources to get this information now- but if one, massive, super flexible, broad “answer engine” existed- this might be a real boon to business intelligence practitioners.

Imagine generating a dimensional table by accessing the web service and enriching what your business users can analyze- and knowing that the values are as up to date and as accurate as the brain trust at Wolfram Research can make them.

Although there is some question as to if the API will really be a focus its clear that there is some interest in business applications. Its just not clear if Wolfram research shares this vision.

Although the masses that are expecting to have a conversation with HAL will be disappointed, there might be a new resource in the world for dimension building for data warehousing- I will be following this with interest.