[SciPy-user] scipy code contribution question

Lev Givon lev@columbia....
Thu Jul 5 08:55:18 CDT 2007

Received from Ryan Krauss on Thu, Jul 05, 2007 at 08:36:28AM EDT:
> I certainly don't claim to be an expert on derivative work, but it
> seems like you aren't creating a derivative work if Mathworks never
> showed you the source code - which I assume they didn't.
> It seems like an overly literal interpretation of the license snippet
> that Bill posted would say that if I ever used Matlab, then I can
> never write code that does the same things as Matlab or any of its
> toolboxes.  If that is legally defended, then I can't contribute
> anything to scipy.
> I appreciate that we want to be careful, but it doesn't make sense
> that we could get in trouble by creating functions with the same names
> as Matlab functions if we've never seen their source code.  I mean,
> can they copyright plot(x,y)?  If so, than matplotlib is in trouble
> (unless John had never used Matlab when he came up with that syntax).
> What about fft(x)?

I agree with your view regarding the implementation of Matlab-like
interfaces; in my previous messages, I was specifically alluding to
functions included in Matlab whose source is visible (i.e., most of
the toolkit functions).

> The link that Bill posted with the GPL code poses as big a problem in
> my mind.  If any of us looks at that code, then any similar algorithm
> would need to be GPL'ed as a derivative work.  So, don't open that
> link :)

No need to worry; one side effect of prolonged scipy use is that
scientific C code begins to take on a marked resemblance to Greek (at
least in the eyes of barbaroi such as myself :-)


More information about the SciPy-user mailing list