Re: GPL traitor !

[Alan Mackenzie]

In gnu.misc.discuss Hyman Rosen <hyrosen@mail.com> wrote:
> Alan Mackenzie wrote:
>> Do you mean that if somebody adds functionality to GPL program, and
>> arranges for this new functionality to be called through a socket call
>> (substitute technically correct terms here) rather than a normal function
>> call, that somebody can remain within the terms of the GPL without
>> licensing his new stuff under the GPL, regardless of how intertwined the
>> new functionality is with the original program?  If so, I think you're
>> mistaken.  But please try and convince me otherwise.

> Yes, that's what I mean. Doing it by function call is OK as well.
> The only thing that forces the foreign code to come under the GPL
> is if the binary program is built as a statically linked whole
> incorporating the GPL and non-GPL code. It is that binding into a
> single work that makes the non-GPL portion fall under the "work as
> a whole" clause and requires that all of it be distributed under
> the GPL.

That's one of the things that requires GPL licensing.  The overarching
thing that mandates new code being licensed under the GPL is its being a
modification of a GPL program.

> You appear to believe that modifying the source of a GPLed program
> so that it invokes a function which is provided separately under a
> non-GPL license violates the GPL even when the modified program is
> distributed *as source*. Is that true?

No.  If the invoked function is truly separate (e.g., calling an emailing
library function from  GCC), it needn't be GPL'd.  If the invoked
function is essentially a part of the calling program, it must also be
GPL.  

This is covered by section 2 of GPL2 ("... and can be reasonably
considered independent and separate works in themselves ...") and
section 5 of GPL3 ("...  which are not by their nature extensions of the
covered work, and which are not combined with it such as to form a larger
program ....").

> If so, then you are certainly incorrect, since copyright law contains
> no concept of a computer program "working".

Sadly, neither do some software engineers.  ;-(

> Copyright law doesn't consider functionality, just text, just
> expression.

Copyright licenses for software do however deal with functionality.
Copyright law also has the notion of "derived works".

-- 
Alan Mackenzie (Nuremberg, Germany).