Wolfram|Alpha API to be released later today

We’ve just been told that the public API for Wolfram Alpha will be made available later today. The API documentation will be available at http://products.wolframalpha.com/api . As of noon, PDT, that page only redirects to the Alpha home page, but they’ve promised it will be available sometime this afternoon.

It’s hard to write much about an API that hasn’t appeared yet, but let me tell you why I think it’s important. Lots of writers have compared Alpha to Google, but I think that’s a mistake. it’s a data source, not a search engine, and that’s a significant difference. What matters with a data source is the ability to ask a question, get an answer back, and use it as easily as possible. An API minimizes the impedance mismatch: you can do computing directly with Alpha’s curated data.

But there’s another comparison that’s even more relevant: Twitter. What has made Twitter success isn’t so much the web application that lives at twitter.com. What has made Twitter valuable is the huge ecosystem that has grown up around that application: alternate clients for all sorts of platforms, web sites for searching, slicing, dicing, and remixing. Those have all been enabled by a simple and well-thought-out API for dealing with Twitter programmatically. The web isn’t about web pages; it’s about interactions between data sources.

Wolfram|Alpha gets this. The web site is just the down-payment; the real value is in the APis that will let developers build an application ecosystem around Alpha. As they put it: “The end-game [is] to drive new and innovative applications of our technology that we literally never could have imagined, driving computational knowledge into new domains.” A huge cache of data is interesting, but not enough: an API that allows you to explore that data, to remix that data with other data sources is what’s needed to enable creativity and discovery. What will the results look like? I don’t know, but I can’t wait to find out. I’m excited.

tags: , , ,
  • jtrav

    How is this going to drive innovation if it costs so much money to use?

  • http://blog.kiwitobes.com Toby Segaran

    Echoing jtrav: you wrote this whole article without mentioning the fact that the API is not free.

  • http://paulmwatson.com Paul M. Watson

    A shame about the cost. “Cheapest” plan is $60 for 1000 requests and $0.08 for every request after that. That is priced out of this developers range.

  • http://www.alexandertolley.com Alex Tolley

    The high cost of the API will marginalise the system for widespread use. At best this will be usable in an enterprise setting. If the idea was to make software for domain experts to build knowledge domains, then I could see limited use in corporations, but I do not see how the innovation will take off without the breadth of small developer involvement.

    Wolfram may do well with this product, but I doubt we’ll see it in mainstream use.