[GRASS-dev] GRASS inefficiency and FFTW

[Glynn Clements]

stefano de paoli wrote:

> > The problem is that we couldn't then use other
> > people's GPL code, as
> > we wouldn't be distributing the end result under the
> > GPL.
> > 
> > Also, what does the "use within GRASS only"
> > restriction on the NR code
> > actually mean for a GPL'd project? If someone takes
> > the GRASS code and
> > starts modifying it, at what point does their
> > version cease to be
> > "GRASS"?
> 
> I may be mistaken, but you would also have lost the
> right to use the GPL at all if the NR code remained in
> GRASS.
> 
> Is that true?

We could release our own code (and that inherited from CERL) under the
GPL, but the licence for GRASS as a whole would be "GPL, except for
the FFT code". Consequently, we would have needed specific permission
to use any third-party code in GRASS; code which is only licensed
under the GPL can't be used in a project whose licence is "GPL, except
for ...".

That would preclude the use of e.g. the Readline library.

Before Qt was available under the GPL, it was quite common for
Qt-based code to use a licence which allowed the code to be used under
the terms of the GPL, and which also specifically permitted the
distribution of binaries which were linked against Qt (which wouldn't
have been legal had the code only been licensed under the GPL). 

However, that meant that all code had to be written anew, or be taken
from projects which also used the "GPL+Qt" license; using plain GPL
code wouldn't have been legal.

-- 
Glynn Clements <glynn at gclements.plus.com>