Re: comments on Wolfram Alpha
- To: mathgroup at smc.vnet.net
- Subject: [mg100352] Re: comments on Wolfram Alpha
- From: dbreiss at gmail.com
- Date: Mon, 1 Jun 2009 07:12:09 -0400 (EDT)
- References: <gvtmjt$gjb$1@smc.vnet.net>
It is important to remember that Wolfram|Alpha is, at the present, a framework for answering calculationally-oriented questions. And that framework currently has access to only some databases and knowledge areas. In the course of time many more will be added. But users that first approach Wolfram|Alpha--and who see a simple natural language input field--often expect uniform response to questions that query arbitrary fields of knowledge. So, it is currently a question of the Wolfram|Alpha team getting the message out about what are sensible exceptions that users should come to it with. Rolling out a project of this sort is, I suspect, very complex--not the least because most users have their key experiences with a very different conceptual model of how to ask questions (a la Google, largely). As for Wofram Research's resources and how they might direct their energies... I hope that much of what has gone into creating more and more curated data sets for Wolfram|Alpha will end up being accessible directly from Mathematica in time. I imagine the W|A will accelerate aspects of Mathematica's development--as well as its acceptance as a far more general tool than most people realize it is. Just some random thoughts.... On May 31, 6:36 am, Beliavsky <beliav... at aol.com> wrote: > input: arima > output: Arima,Trinidad and Tobago > Population: > city population | 34997 people > my comment: Nothing about ARIMA time series models, which comes up as > the 3rd link on Google. > > input: garch > output: Wolfram|Alpha isn't sure what to do with your input. > * Did you mean:march > my comment: nothing about GARCH time series models, widely used in > finance > > input: united states budget deficit > output: -$347 billion per year (US dollars per year) (2005 estimate) > my comment: Users would want more recent data, such as the 2008 > deficit and the estimated 2009 deficit. > > Wolfram Alpha appears much less useful than the combination of Google > and Mathematica, so I wonder if resources are better spent adding > functionality to Mathematica.