By SemanticWeb.com on June 5, 2009 2:17 PM
Jennifer Zaino
SemanticWeb.com Contributor
What’s next for Wolfram Alpha? Stephen Wolfram hosted a live webcast this week to respond to some of the feedback that the ambitious project “to make all systematic knowledge immediately computable by anyone” has received.
Among the plans on tap — some of them possibly to see daylight as early as next week — and issues that have cropped up are the following:
Support for more languages. Currently WolframAlpha will recognize the language something other than English is typed in, but it doesn’t yet know how to interpret it. It’s a difficult project, Wolfram says, “but the framework we have is wonderfully suited to doing it.” It already deals with international notations, as in mathematics or professional notations, some of that dealing with unit systems and date formats. As for pure words, once input is understood synthesizing the output in different languages is the tackling point.
The developers’ API is going well, Wolfram reports. He expects a preliminary version for developers to experiment with the API to be available in a few weeks. Current thinking is that a free version will be available for casual usage with one following the fee-based model to be available at some point for professional purposes. Wolfram Alpha says it is developing a ‘softer’ Mathematica-like language that doesn’t require as precise syntax as part of its API development.
Speaking of fee models, Wolfram noted that there are a variety of ways to build a business based on WolframAlpha. That is expected to include a subscription-based professional version that will roll out soon, probably to be available in a number of different versions for downloading more kinds of formats, uploading user data to the system for analysis, storing user definitions and session history, and supporting premium data services provisions paying users will be able to gateway to via WolframAlpha. But that doesn’t mean the free version is going away.
“The WolframAlpha you see today, our intention is to make it as broadly and freely available to the world as possible,” he said.
Some questions arose around providing the ability to create dynamic graphical models from data and manipulate them as users can do in Mathematica. The problem there is all the computation going on that happens locally on a PC in Mathematica, but would slow servers to a crawl if implemented in WolframAlpha. That problem may be addressed, though, with a version of WolframAlpha that lets you run manipulations created on the fly from data through a client-side notebook player.
Another goal is to increase the sophistication of technology for analyzing what doesn’t work right so far. There are tens of millions of things it doesn’t know how to do that people have asked for, Wolfram said. It needs to semi-automate or automate the processing of different domains to more quickly discover the kinds of things that aren’t working and address them.
Among the data users have been asking for have been some surprises for the WolframAlpha team — for instance, there have been a lot of requests for bibliographic data, legal data of various kinds, and history and politics. People are very interested in typing in dates to find out what happened, which WolframAlpha had not much emphasized and so it is “dramatically” scaling that up and adding lots of history information.
“It’s a linguistics challenge to make that information computable so you can ask when did such a think happen, when the description is fuzzy,” he said. Some output could emerge as early as next week but he promised more to come in this area. Just added have been median house price data for different areas in a preview version of WolframAlpha, with foreclosure rate data coming soon.
What he’s personally most excited about: Being able to do programming in WolframAlpha style. That is, type in what he is trying to get a program to do, describe it as he thinks of it and have WolframAlpha return candidate pieces of code – for example, fragments of Mathematic code that might fit the program he was describing, and then offer test cases so he can click on the one he wants and insert it into the program he might be building.

Semantic Tech & Business Conference returns to San Francisco in June! Join us from June 3-7 for complete coverage of Big Data, Linked Data, Extreme Information Management, and Semantic Web. From breakthrough approaches to solving business problems to the big data implications of fast–evolving technologies, SemTechBiz provides you with an unparalleled interactive experience and delivers tangible business value.
We're offering a special early rate when you register by February 17.
Sign up now!