Re: GPLv3 Position Statement

From: Linus Torvalds
Date: Thu Sep 28 2006 - 10:41:01 EST




On Thu, 28 Sep 2006, Jörn Engel wrote:
>
> My graph traversion code I did for my thesis should have been merged
> into gcc, but I didn't even bother sending a patch. Copyright assign
> my ***, thank you very much.

Yeah, I don't see why the FSF does that. Or I kind of see why, but it has
always struck me as
(a) against the whole point of the license
and
(b) who the F*CK do they think they are?

I refuse to just sign over any copyrights I have. I gave you a license to
use them, if you can't live with that, then go fish in somebody elses
pond.

I had some code I tried to give to glibc back when I was doing Linux/axp,
since glibc was really in pretty sad shape in some areas. I think I had a
integer divide routine that was something like five times faster than the
one in glibc, and about a tenth of the size. Things like that. So I wanted
to give things back, but ended up just throwing in the towel and said "ok,
if they don't want the code, whatever..".

If somebody pays me real bucks, I'll work for them, and I'm perfectly ok
with letting them own copyright in the end result (ie I've done commercial
software too, and I think it's only reasonable that since I did it for
pay, I don't own that software myself). But copyright assignments "just
because we want to control the end result"? No thank you.

> And that is in fact the primary reason, hacking gcc has been fun and I
> would like to do more, from a purely technical point of view. But
> having to sign a large amount of legalese is the kind of thing I may
> have to do for a job, and they pay me for it. It is not the kind of
> thing I do for fun. No fun, no money - hell, why should I do
> something like that?!?
>
> Thank you for not requiring copyright assignments, Linus.

I _hate_ copyright assignments. Maybe it's a European thing.

Of course, I also hate paperwork, so maybe it's just a "lazy" thing.

Linus