Re: GPL 2(b) HUH?

[Hyman Rosen]

Ciaran O'Riordan wrote:
> > No, it would not. By statute, in the U.S., a derivative work is a
> That's not about software, and that's only one jurisdiction.

Unless you can find something in the statute that defines derivative
works of programs differently than derivative works in general, the
general definition will apply. And granted that the U.S. is only one
jurisdiction, it's a pretty big one, and it's the home country of the
GPL.

> The example you're replying to was a scheduler which is an intimate
> part of a kernel, definitely not a separate program that interacts
> with the kernel.

So what? The scheduler code does not incorporate the kernel into
itself, it is just designed to work together with the kernel. The
scheduler code itself is therefore not tied to the copyrights of
the kernel.

Let's do the analogy thing again. Suppose I write a mathematical
paper in which I prove a theorem. As it happens, the proof of this
theorem involves a series of steps each of which is proved using
existing theorems, and for each step, I say that some statement is
true and include a citation to the paper proving the sub-theorem.
I've used obscure results, so that even to a specialist, my paper
cannot be understood without going to the references and reading
them. Nevertheless, the copyright on the paper is mine alone. It
is not a derivative work or compilation or collection of the papers
I cite.