Via my colleague Michael Trick, Groklaw has an interesting post pointing out that the recently unveiled Wolfram|Alpha computation service makes some pretty strong claims if not to copyright, then to the right for attribution, for results the service returns.

For example, if you plug x^2*sin(x) into the search window, you will get back a graph of this function, as well as a number different series representations for the function. Wolfram|Alpha claims that these materials are protected. Individual use of them must be attributed, and any commercial use requires a specific commercial license. The problem with this, though, is that any table of mathematical formulas will provide both the graphs and series representations for this and many other functions. Furthermore, it could be reasonably argued that these are facts, which generally can't be copyrighted.

The Groklaw post contrasts this with Google's terms of service, which basically says you can't use the service to break the law.

I would also contrast Wolfram|Alpha's service with that of Economagic, which provides publicly available economic data, and, for subscribers, the ability to generate graphs, run regressions, download data to spreadsheets, and do other kinds of data analysis. None of these results are held to be protected, and Economagic requires no specific attribution. There are also no limitations or additional requirements for any commercial use of the service. The subscription fee is also easily within reach of any economics graduate student (which is the site's target audience).

So, I would have to agree with the Groklaw post that Wolfram|Alpha seems to be overreaching.

If I use freely provided software to create a graph, then if copyright applies at all, how can it possibly apply except to say that the graph is mine?

But then the counter argument perhaps becomes: If an artist or draftsman says he will draw graphs to your specifications for free, then if copyright applies at all, how can it possibly apply except to say that the graph is his?

What a conundrum! It's almost as if monopolies on reproduction aren't well suited to the realities of non-physical works...

Second thought:

Perhaps it could be argued that a graph of a certain relation of values is a "performance" of the natural facts of the statistics and mathematics involved. That way, copyright would apply, just as it does when an orchestra records music itself in the public domain.

But the orchestra's copyright is only on works that they actually record. This claim seems to be asserted on all possible displays the software is able to produce, which would be overreaching indeed. In fact, I can't imagine how their claims could reach further.