Diana Nutting Wolfram Alpha - hype and reality
Jinfo Blog

20th May 2009

By Diana Nutting

Item

There has been a great deal of comment in the last week about Wolfram Alpha, http://www.wolframalpha.com following its launch with a number of commentators, including Phil Bradley http://digbig.com/4ysqd David Gurteen http://digbig.com/4ysqe and Karen Blakeman http://digbig.com/4ysqf carrying out searches to compare the new tool with Google. FreePint's own Gary Price featured it as last week's Resource of the Week http://digbig.com/4ysqg The Atlantic Monthly for example asked if Wolfram|Alpha knew the name of Barack Obama's Supreme Court nominee (it didn't.) The reaction has been mixed. As well as a long list of questions it either could or couldn’t answer commentators have noted some long page-load times, and more than a few non-responses. And of course for those of us not in the USA, there is only American data on there. Basically Wolfram Alpha attempts to work out and provide answers to questions rather than provide links to other sites in the manner of Google. Its long term goal is to make all systematic knowledge immediately computable and accessible to everyone, but regards the present launch as just the beginning. But of course not all information is computable and it seems to fall down when asked anything that is not of a mathematical or statistical nature. Because Wolfram Alpha staff are producing the tables and graphs themselves, in many cases they claim copyright on anything downloaded and used. Its terms of use state that the service is available for ad hoc, personal, non-commercial use only, so my guess is that this might end up as a paid for service. Wolfram Alpha is accepting help with topic suggestions and other user input to its database. It is also taking applications for users to test new versions of the site and people with very specific knowledge of certain subjects can join its experts network. Wolfram Alpha has been hyped as the next stage of search engine technology -- a "Google-killer," a new way to ask the Internet a direct question. But it’s not a search engine, it is a data source. As Phil Bradley puts it in his blog if people try to use it in the way that they use Google it doesn't work - it's like expecting an encyclopaedia to give you the address of the local takeaway. For the time being Wolfram/Alpha is likely to remain a research tool for relatively sophisticated users until its knowledge base expands and its ability to understand poorly crafted queries improves. I’ll use it as part of my tool kit, and watch its development with interest, but for now I don’t think the reality matches the hype.

« Blog