Wolfram Alpha Revisited
Posted: 18 May 2009, 10:08
Now that Alpha is actually open for public usage, I'd like to share some first thoughts on the usefulness of this search engine.
First of all, describing it as a "computational knowledge engine" is fairly accurate. Imagine it as the core functionality of Mathematica coupled with an encyclopedia of facts and figures, wrapped in a natural language interpreter. As a student, it's usefulness is immense.
For example, I could input "Integrate(ln(sin(x^2)))", and Alpha will spit out a useful set of output, including a plot of the integral, and a series expansion since the indefinite integral can't be exactly defined. Although functionally, it's not as powerful as full fledged Mathematica, it is both easier to use and more importantly, free. It's also available anywhere I can access the internet with an HTML browser, meaning I could use it at an internet kiosk, or on my smartphone / MID.
I can type in natural language names of scientific constants, etc, and the engine will automatically try and match values for me, and perform any calculations I specify. For example, I tried "(mass of the earth)/(Mass of the Milky Way Galaxy)*(electron mass / elementary charge)*(human population)"
Nearly every common term I input returns a useful set of data. If I search "milk", I get a comprehensive list of physical properties and nutritional information. If I type "emmanuel", I get statistics on its use as a common name in the US, including the estimated fractional distribution of people in the US with that name, the estimated number of people alive with that name, etc etc.
And then there's the obvious comparison that is made between it and Google. Well, let me explain, that essentially, there is no comparison. Clearly, each engine serves its own purpose, with Google being an engine to search the web for web page content, and Wolfram Alpha an engine to gather data on subjects and/or do computational work.
Finally, Alpha is still somewhat limited in terms of the search terms it will recognize. For example, a search for "boston celtics" or "los angeles lakers" doesn't turn up any information / statistics at all. It's definitely still a work in progress, but it is also fairly impressive as it is.
First of all, describing it as a "computational knowledge engine" is fairly accurate. Imagine it as the core functionality of Mathematica coupled with an encyclopedia of facts and figures, wrapped in a natural language interpreter. As a student, it's usefulness is immense.
For example, I could input "Integrate(ln(sin(x^2)))", and Alpha will spit out a useful set of output, including a plot of the integral, and a series expansion since the indefinite integral can't be exactly defined. Although functionally, it's not as powerful as full fledged Mathematica, it is both easier to use and more importantly, free. It's also available anywhere I can access the internet with an HTML browser, meaning I could use it at an internet kiosk, or on my smartphone / MID.
I can type in natural language names of scientific constants, etc, and the engine will automatically try and match values for me, and perform any calculations I specify. For example, I tried "(mass of the earth)/(Mass of the Milky Way Galaxy)*(electron mass / elementary charge)*(human population)"
Nearly every common term I input returns a useful set of data. If I search "milk", I get a comprehensive list of physical properties and nutritional information. If I type "emmanuel", I get statistics on its use as a common name in the US, including the estimated fractional distribution of people in the US with that name, the estimated number of people alive with that name, etc etc.
And then there's the obvious comparison that is made between it and Google. Well, let me explain, that essentially, there is no comparison. Clearly, each engine serves its own purpose, with Google being an engine to search the web for web page content, and Wolfram Alpha an engine to gather data on subjects and/or do computational work.
Finally, Alpha is still somewhat limited in terms of the search terms it will recognize. For example, a search for "boston celtics" or "los angeles lakers" doesn't turn up any information / statistics at all. It's definitely still a work in progress, but it is also fairly impressive as it is.
