Re: The worst that can happen to GPLed code

[David Kastrup]

Alexander Terekhov <terekhov@web.de> writes:

> David Kastrup wrote:
> [...]
> > glibc is not a derivative work of the Linux kernel, 
> 
> Sure. In spite of things like "#include <kernel-features.h>".

Which is not a part of the kernel.  Similar compatibility files are
included for _dozens_ of operating systems.  Since glibc is not even
particular to Linux, you can't claim it being an
extension/plugin/whatever that can't be used without the kernel.

> > Even if glibc happened to be linked into the kernel (which it
> > isn't)
> 
> What do you mean? That ring protection is the only boundary that the
> GPL can't cross?

Care for a somewhat less obtuse statement?

> > this would only mean that the linked kernel as a whole would be
> > covered by the GPL.
> 
> Well, you said that "the whole CD is a derivative work" and Moglen
> says that "the terms of the GPL and only the GPL" must cover the
> entire "work based on" [derivative work or whatever he might be
> thinking of]

Sigh.  Of course an aggregation is a derivative of the original works,
and the parts are covered by the individual copyrights.  But it is
only where the combined work becomes an inseparable whole that is
beyond the scope of copyright law's aggregation clauses that the terms
of the GPL demand that the combination as a whole is licenced under
the GPL.

> whole you distribute. So, again, tell me please how does it work in
> the case of kernel + stuff.

The GPL applies to those cases where copyright law says the licence
applies.

Simple as that.  It might mean different things in different
countries, but at least in the U.S., there is a pretty broad consensus
among legal practitioners what it means.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum