Re: [Gnu-arch-users] re: licensing question

[Andrew Suffield]

On Wed, Mar 29, 2006 at 02:51:01PM +0900, Stephen J. Turnbull wrote:
> >>>>> "Andrew" == Andrew Suffield <address@hidden> writes:
> 
>     Andrew> As to the question of whether or not it's a GPL violation
>     Andrew> - it was our considered opinion that this is unclear.
> 
> Interesting.  The GPL says quite clearly:
> 
>       1. You may copy and distribute verbatim copies of the Program's
>     source code as you receive it, in any medium, provided that you
>     conspicuously and appropriately publish on each copy an appropriate
>     copyright notice and disclaimer of warranty; keep intact all the
>     notices that refer to this License and to the absence of any warranty;
>     and give any other recipients of the Program a copy of this License
>     along with the Program.
> 
> and
> 
>       6. Each time you redistribute the Program (or any work based on the
>     Program), the recipient automatically receives a license from the
>     original licensor to copy, distribute or modify the Program subject to
>     these terms and conditions.  You may not impose any further
>     restrictions on the recipients' exercise of the rights granted
>     herein.
> 
> I don't see how Red Hat's conditions on redistribution can be
> compatible with that: http://www.redhat.com/licenses/rhel_us_3.html,
> Appendix 1 Section 2 clearly tries to impose restrictions (at least
> diligence in finding and removing RH IP) on verbatim redistributions.

Well, this isn't strictly a restriction on redistribution, it's a
one-off hoop-jumping requirement. The distinction lies between these
two licenses:

  You must pet a cat before you redistribute this software.

(The 'pet a cat' license, classic extreme example of a license that is
not free or GPL-compatible because it contains a minor restriction on
distribution)

  You must pet a cat, and then you receive a copy of this software
  licensed under the MIT license.

The first form, which everybody agrees is not free, affects
everybody. The second form requires only that one person pet one cat,
and then redistribute the definitely-free copy they receive under the
MIT license to everybody else - effectively, they are 'purchasing' it
for the community.

[You can do the same thing for trademark-inflicted works by removing
all the marks; the result is definitely free, being GPLed]

Personally, I don't think this is right, but the argument goes that:
since a free version can be created via some trivial effort, and all
redistribution involves at least initial trivial effort (like the
effort of sending it to a mailing list), it does not form an effective
restriction on distribution - it's really just a speed bump on a path
that's already fairly bumpy.

You can certainly make good arguments against that, but it's unclear
whether a court would rule that a speed bump like this is incompatible
with the GPL. If they did not, it's also unclear whether that would
extend to something on the scale of redhat ("it is true that the
burden of removing all trademarks is not insignificant, but that
merely derives from the fact that releasing an operating system is
more complicated"). Obviously we couldn't afford to pay for proper
legal research on the subject, but there's an acute lack of useful
legal precedents that relate to the unique features of free software
licenses - and this case is borderline.

I'd really like to see a solid precedent set against speed bumps like
this one, it would clear up a number of related issues. But I don't
think it's likely to happen. Companies don't normally go to court with
things like this, it's cheaper to do the work of removing the
trademarks. Individuals can't normally afford to go to court - however
much I might want a precedent set here, it's not worth the tens of
thousands of pounds it would cost to get the job done myself.

[The other argument is that trademarked artwork does not form a
derived work, it is merely distributed alongside the GPLed work,
therefore the GPL does not apply to it. That's an old one that will
probably never go away, because it's determined on a case-by-case
basis in court.]

>     Andrew>  It's difficult to get upstream authors to fix these
>     Andrew> problems when you have other people running around
>     Andrew> claiming that debian-legal is wrong and has no authority
>     Andrew> to say anything.
> 
> [...]
>     Andrew> And we thought that was an acceptable model for people to
>     Andrew> use. A number of other projects have done similar things
>     Andrew> following debian-legal's advice.
> 
> I think you overestimate the power of unity here, and the "other
> projects" show it.  Honest redistributors will give weight to the
> (deservedly!) well-reputed opinion of debian-legal.  Others will hide
> behind any legal excuse to keep doing what they're doing, including
> (but not limited to) trash-talking debian-legal and claiming that the
> GPL's intent doesn't matter, only what's explicitly forbidden by the
> language, which is effectively nothing until a court rules, etc.

Oh, that's certainly true enough, but it's the people who are just
*ignorant* about licensing issues who are the problem case - it's
difficult to get them to fix something when you've got a bunch of
other people telling them a different story.

If there's one thing I learned from debian-legal, it's that most
people do not understand licensing, a significant number of upstream
authors do not care about it, and the vast majority of licensing
issues arise from simple ignorance.

> I certainly hope debian-legal will continue its excellent work.

AOL, but I don't have anything to do with it any more.