Python license (2.3)

[Robert Kern]

Antoon Pardon wrote:
> Op 2005-04-14, Robert Kern schreef <rkern at ucsd.edu>:
> 
>>Antoon Pardon wrote:
>>
>>>Op 2005-04-14, Robert Kern schreef <rkern at ucsd.edu>:
>>>
>>>
>>>>Antoon Pardon wrote:
>>>>
>>>>
>>>>>Op 2005-04-13, Robert Kern schreef <rkern at ucsd.edu>:
>>>>>
>>>>>I would do that if I were just writing code I thought others could
>>>>>find usefull. I then would feel no problem "burdening" those users
>>>>>with the same kind of license I found in the product I took some
>>>>>code from. But I also think that readers of documentation should
>>>>>be free to use any code included in any way they see fit.
>>>>
>>>>If they have issues with distributing code derived from Python, why are 
>>>>they reading a Python tutorial?
>>>
>>>
>>>Try and look it from a students viewpoint. He is learing languages,
>>>algorithms and so on. Now he is ready to write his own program.
>>>Chances are high that he will rely on examples from the
>>>courses/documentation he read. It is just not practical for someone
>>>like that to figure out all the possible different licenses under
>>>which he can use the examples from the various documenation sources.
>>
>>The PSF License is about as light as they come.
>>
>>
>>>Now if this documentation refers to code from yet another source
>>>with its own license, using it becomes an utter nightmare for
>>>the student, because now he has to figure out which piece of
>>>the code is original from the author of the documentation and
>>>which was copied from the other source.
>>
>>Then write your own code and don't use anyone else's. You can't offer 
>>extra permissions for code that's not yours.
> 
> 
> Well then I'll just have to do that.
> 
> 
>>>Consideration like this, let me come to the conclusion that
>>>code included with documentation should come with no strings
>>>attached for the students to reuse.
>>
>>No such thing, really. Copyright law requires almost as much as the PSF 
>>license. The MIT license is shorter, possibly more easily 
>>understandable, but practically amounts to more-or-less the same thing.
> 
> 
> If I read a tutorial or a course on algorithms both with examples.
> Does copyright law require that I attribute if I reuse code
> from these examples?

If the amount copied is large enough.

> Even if it was pseudo code that I had
> to translate in an actual language.

Probably not. Copyright controls copying (and a few other things, but 
they have less relevance in a software context).

> Suppose some time has passed and I have to write similar code.
> I cant find the documentation but this time I'm experienced
> enough so that I can recreate the code. Do I still need to
> attribute the code?

Again, probably not.

> What if the code is so short that basically everyone that
> solves the problem writes the same kind of code?

No, copyright requires creativity.

Rosen's book should answer these questions for you.

>>In short, don't worry about it. Don't sue people, keep the attributions 
>>intact, and probably no one will care.

Sorry, that list should also have had "follow courteous practices with 
other people's code" which includes listing changes and a reference to 
the license of that code.

> If they don't care, why did they attach such a license in the first
> place.

Lawyers. The original license (the "CNRI License") was much briefer and 
vaguer, although it amounts to the same requirements, practically. The 
PSF license made those requirements, disclaimers, etc. explicit.

-- 
Robert Kern
rkern at ucsd.edu

"In the fields of hell where the grass grows high
  Are the graves of dreams allowed to die."
   -- Richard Harter