[SciPy-User] log pdf, cdf, etc

[Anne Archibald]

On 30 May 2010 01:20, Skipper Seabold <jsseabold at gmail.com> wrote:
> On Sat, May 29, 2010 at 3:44 PM, Nathaniel Smith <njs at pobox.com> wrote:
>> On Fri, May 28, 2010 at 9:00 PM,  <josef.pktd at gmail.com> wrote:
>>> On Fri, May 28, 2010 at 11:24 PM, Nathaniel Smith <njs at pobox.com> wrote:
>>>> On Fri, May 28, 2010 at 7:53 PM,  <josef.pktd at gmail.com> wrote:
>>>>> R's license, GPL, is incompatible with the license of scipy, BSD.
>>>>> While they are allowed to look at our code, code that goes into scipy
>>>>> cannot be based on GPL licensed code.
>>>>
>>>> You mean, they're allowed to copy our code, and we're allowed to look
>>>> at their code for reference but can't use it directly :-).
>>>
>>> We are allowed to look at their manuals but not their code.
>>> (Life ain't fair.)
>>
>> It sounds like you guys have this well in hand, but just a point here
>> -- you certainly are allowed to look at their code, just not copy the
>> "expressive aspects" of it. (Saying you can't *look* at it because of
>> the license is like saying writers can't read other people's novels!)
>> "Expressive" is a tricky term, of course -- IIUC it's basically
>> anything that could be changed while preserving functionality (because
>> the functionality, the algorithm itself, is not covered by copyright).
>> So, say, variable names certainly count as expressive, decisions about
>> which way to lay out the code, etc. If one wants to be really safe,
>> one can write down a textual description of the algorithm and then ask
>> someone else to translate back to code (the "clean room" method).
>>
>> So you do have to be a bit careful, but when you have code that
>> contains valuable information that isn't really written down anywhere
>> else then I'd say it's worth it.
>>
>
> Thanks, this is useful to know.  I've always erred on the side of
> caution and just compared the results of functions/algorithms that
> *should* be the same vs, say, R, but if I could do this and then look
> at implementation details this could relieve substantial headaches.
> It still seems like such a fine line though.

This is exactly the problem. I don't think the R community is
particularly litigious, but as a rule of thumb, doing something that
is technically legal but for which the legality is subtle opens one up
to lawsuits. The problem is that even when you are right, a lawsuit is
tremendously destructive. So things that are legal but subtle should
probably be avoided by a group as penniless as the community as scipy
developers. So it's probably better to just not read their source
code.

Anne

> Skipper
> _______________________________________________
> SciPy-User mailing list
> SciPy-User at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>