Github ToS and Open Source

[Paul Smith]

On Fri, 2017-03-03 at 20:05 +0100, Xen wrote:
> Paul Smith schreef op 03-03-2017 19:32:
> 
> > We'd better hope that's not the case, because if it is then almost the
> > entire internet / computing industry is constructed from illegal
> > copyright infringement.
> 
> I mean that the restrictive terms are probably not enforceable. So
> what I mean is that although the "free use" cannot be contracted (or 
> contradicted) "non-free" use can possibly not be punished.
> 
> [...]
> 
> What I meant was the "applies to all derivative works" nature in that
> it is a _strong_ copyleft license, ie. you cannot enforce that every 
> software using any part of GPL'd code must also be GPL.


I (believe I) understood what you meant.

However, the only rights you have to the software are those granted you via the license.  So you either agree to the license in its entirety, including the "strong copyleft" aspect, or you don't get to use the software at all.  Clause 7 of the GPLv2 (there's a similar clause in v3) makes very clear that if you don't meet your obligations under the GPL for any reason (including court order) then you cannot distribute at all.

The GPL is actually in a very strong position and I disagree that it can't be enforced.

The area which is NOT so well-defined is what constitutes a "derived work" in the first place; the GPL certainly won't apply to things which are not derived works of GPL-covered software.

If you modify the code itself then distribute it, that's a derived work: there's no question about that.  And since this is mostly what GitHub does there's not much argument there.

But if you just use some code via its defined API (even if it's statically linked into your program) is that a derived work?
Is a kernel module a derived work of the Linux kernel?  Is a program a derived work of any (or all) of the shared libraries it links?  This is where the grey areas are in GPL enforcement these days.


Anyway, cheers!