On Thu, Nov 19, 2009 at 8:34 AM, David Grundberg
<address@hidden> wrote:
Judd Storrs wrote:
> 1) Octave is an interpreter. The GPLv3 license does not apply to
> interpreter input. e.g. m-files can be any license without violating
> the interpreter's license.
>
> 2) python + pytave is also an interpreter. In order to distribute a
> python + pytave interpreter, all components of the interpreter must be
> GPLv3 compatible.
>
> 3) Python scripts are also interpreter input. They are equivalent to m-files.
Yes, I think Python + Octave must be distributed under the terms of GPLv3.
I think its unnecessary to discuss Octave at all. Pytave is licensed
under GPLv3 on its own, and it doesn't matter what libraries it in turn
uses. Since Pytave consists of Python code and is GPLv3, any other
Python script that uses Pytave is a derivate work of Pytave.
You could turn the argument around and say that it's OK to take GPL'd
m-files from Octave and distribute them with proprietary m-files. I
think that is not the case.
.. am dropping in here a bit. If i remember correctly, whether a software has to be distributed as GPL or not depends on the distribution bundle.
As an example, although GCC is GPL software, applications that require GCC during compilation do not have to be GPLed.
However, as soon as you distribute a software that uses GCC internally to compile something, this software should be GPLed.
The usual work around is, user installs and provides GCC or python to the "other" application and configures the "other" application, by pointing to the gcc or python compiler. Doing so, unbundles the GPL software from the proprietary software and frees the distributor of proprietary software from the necessity to disclose his source code.
Correct ?
Ulrich