Re: [AMENDMENT]: Release Etch now, with source-less but legal and freely licensed firmware
On Tue, Oct 03, 2006 at 11:03:29PM +1000, Anthony Towns wrote:
> On Tue, Oct 03, 2006 at 02:09:50PM +0200, Frank K?ster wrote:
> > + | We allow inclusion into etch even if the way we distribute the
> > + | firmware leads to a violation of the license,
> Uh, no we won't.
> There are claims that the GPL, when applied to "sourceless" firmware,
> doesn't provide permission to redistribute because there's presumably a
> "more preferred" version of the source in existance somewhere. That's
Anthony, this is bullshit. These are no claims, but the plain GPL. All right
to distribution under the GPL are lost if you don't comply with the GPL, and
providing the source code for those firmware is mandated by the GPL. There is
ample information and hints available to prove that such source code does
> an *argument* that a violation may exist, not proof that one does. If
> that argument were accepted by Debian, we would not be distributing
> it no matter what GRs there might be, right up to the DFSG and Social
> Contract being entirely scrapped -- it would be *illegal* to distribute
> those works, both for us, for Red Hat, for kernel.org and just about
> everyone else.
This is indeed the case. The whole thing is that these sourceless GPL
firmwares are in this case, not because of a willing situation, but by error
and miscomprehension of the hardware vendors who provided them.
It is our duty to educate them, in order to solve this non-distributability,
as was successfully done in the broadcom/tg3 case, whose licence is still
non-free, but which also clearly state "based on non-pulished source code" or
some such wording, but in the meantime we can make an exception for these in
So, what about saying something like :
We recognize that there are a number of firmwares in the kernel who are
sourceless and implicitly under the GPL. We believe these are mistakes on
the copyright holder part, who didn't understand the full measure of what
they where doing in releasing those firmwares in this fashion. We will work
with those vendors to either provide the source of those GPLed firmwares,
clarify their licencing, as was done with broadcom and qlogic, or give us a
statement indicating these are just plain register dumps.
This clearly states how things are, why we are making an exception for them,
and engages us to solve this unclearness for etch+1.
> It's plausible we might decide not to risk even the possibility of
> violating the license by distributing that stuff (even in non-free) even
> though Red Hat or others are willing to, and it's also plausible that we
> might decide as a matter of policy that while we are able to distribute
> it, that it doesn't meet our standards for "main"; but we certainly won't
> decide that it's a violation of the license to distribute it, but hey,
> we'll distribute it anyway.
Well, it is a violation of the implicit GPL licence of this file, and any
amount of hiding your head in the sand will not change that plain fact.