Re: please release sarge instead of removing binary firmware
* Anthony Towns (firstname.lastname@example.org) [040416 14:10]:
> [-devel dropped]
> On Fri, Apr 16, 2004 at 11:00:08AM +0200, Andreas Barth wrote:
> > * Anthony Towns (email@example.com) [040416 10:25]:
> > > Because when you have a compile-time dependency you create a derived
> > > work -- vmlinuz -- of both the GPLed work and the firmware. Creating
> > > derived works is an act covered by copyright, and can only be done with
> > > the permission of the rights holder; and if the only permission you have
> > > comes in the form of the GPL, then you're required to make available
> > > the complete source to the entire derived work.
> > It is not necessarily a derived work, but could also be just a
> > collection of works. (And I consider this to be the case here.)
> Collections of works are derived works; the GPL has a specific exemption
> for collections made up of a derived work that's "merely aggregated"
> with other works.
I meant "merely aggregated". However, I was a bit too tired this
morning for realizing that I should've re-read the GPL before writing
> I don't think it's reasonable to claim that you're
> "merely aggregating" the works, when you encode the firmware in hex,
> and stick it into an array in the source code -- that is, I don't just
> think that's a questionable enough claim that it shouldn't be relied on,
> but I think the opposite claim is strong enough that it could be.
> As far as I can see, the argument that says firmware inserted into a
> GPLed program isn't affected by the GPL because it's "mere aggregation"
> would also have to apply to someone inserting GFDL'ed documentation
> into a program too, or a non-modifiable image for an icon, or a host of
> similar things.
Well, this can IMHO only be considered on a case-by-case basis. The
safe assumpation is (and can only be in my opinion) that we treat the
whole as one derived work, and I really appreciate that you resist all
trys to get you away from that.
However, in the case like the NVidia driver (where Linus himself said
that they are not derived), or also for some of the binary blobs, they
are really only aggregated and not derived. Having said this, I want
to emphasise that even if some of them are aggregated, the right way
to handle this is of course to treat all as derived, unless it has
been proven on a case-by-case-basis by someone with good legal
knowledge that they are only aggregated (or at least following the
guideline set up by such a person), and my considering is definitly
not the right measurement for the decision.
PGP 1024/89FB5CE5 DC F1 85 6D A6 45 9C 0F 3B BE F1 D0 C5 D1 D9 0C