Re: DRAFT for a GR proposal concerning the Sarge release
Lewis Jardine wrote:
> Thiemo Seufer wrote:
> >>As I understand it, Debian makes a point of considering the interests of
> >>'unrelated third part[ies]', especially when it comes to the chance of
> >>copyright infringement.
> >So does Debian consider the interests of SCO then? They also claim
> >copyright infringement.
> I'd hope so, in as much as Debian provides SCO (like all other users)
> with a high quality collection of Free software (No discrimination
> against fields of endeavour, remember :)
So we will stop to distribute Linux because of their claims? If not,
why do we take some hypothetically existing kernel developer more
serious than SCO, who publicly threatened to sue everyone?
> >Linking means to bind some object files together. Those firmwares
> >aren't distributed as object files. Which relies on the rather weak legal
> >theory that compiled in
> >firmware is part of a derived work, while the same firmware in
> >a ramdisk image (or even a CD image) suddenly constitutes a
> >collection of works.
> It can be argued under the new social contract amendments that many of
> these blobs are non-free, and have to go, whether or not they can be
> included in the kernel image without violating the GPL.
In its broadest interpretation it allows to exclude enough from
Debian to render the remains useless.
> If nothing else, it would make the kernel image with the firmwares and
> the kernel image without the firmwares identical; loading these blobs in
> at run-time would mean that kernel-blobs-non_free could be packaged
> separately, saving the pain of having to maintain kernel-image and
Maintaining a bunch of firmware .(u)debs and keeping them in sync with
their appropriate kernel version is surely more effort that two kernel
> >>a delayed Sarge would be annoying, but the products that are necessary
> >>for an 'anally-free' Sarge would be of great benefit to users of both
> >>Debian, and Free Software in general.
> >What exactly are these great benefits? I see diminished driver support
> >and a lack of documentation, or alternatively non-free as a rather
> >mandatory part of a Debian installation.
> Ah. I was seeing clean-roomed/relicensed firmwares, rewritten, Free
> documentation, etc. I assumed that the reason for the delay was due to
> reverse-engineering, documentation, and re-licensing. Best case, I was
> envisaging a back-down by the FSF over the GFDL, and the reintroduction
> of much of the documentation, under a Free license.
> Surely it can't take nine months just to take out the stuff that's been
> declared non-free, and not replace it at all?
Nine months for the reverse engineering of a vast and increasing array
of firmwares without hardware documentation is an excessively optimistic
estimate. It may work for some binary drivers (which aren't the matter
here), but rewriting some firmware for an undocumented and unknown
system can't be done with reasonable effort.
> > And this still doesn't count
> > the fight if a jpeg or some font descriptions can be source.
> I'm not touching that one with a 60 foot pole.
> >>Clause four of (even the unamended) social contract, in my opinion,
> >>suggests that later is better than less free, and thus the amendment was
> >>The Right Thing, even though it may delay Sarge.
> >In my opinion, invoking the Social Contract is Debian's version of
> >Godwin's Law.
> I'd say that it beats Godwin's Law, as the Social Contract is (at least
> supposed to be :) relevant to the discussion at hand.
Well, to choose a different wording: If somebody has to resort to citing
from some holy book, then you know there aren't any arguments left.