On Tue, Feb 22, 2000 at 12:10:25AM -0500, Raul Miller wrote:
> > > And reuse doesn't really focus on the right issue. The issue is whether
> > > all
> > > users have full rights to distribute and modify the code and all
> > > derivatives
> > > of the code.
> > >
> > > The QPL guarantees that all users do not have such rights, thus it's
> > > incompatible with the GPL which guarantees that all users do have such
> > > rights.
>
> On Tue, Feb 22, 2000 at 05:44:49AM +0200, Adi Stav wrote:
> > Is this really the issue, or is the issue the fact that modifications
> > cannot be copylefted? Consider a license very similar to BSDL, except
> > that it contains a clause saying that you can distribute binaries
> > under any license you like, but source code (modified or not) must
> > be distributed under the BSDL. Such a license would give all users
> > equal rights, but also ensure that you have no way to prevent your
> > modifications from becoming proprietary binaries.
> >
> > Would you still consider such a license free? And will it be
> > compatible with the GPL?
>
> I don't see any issue here. It's already the case that BSD software is
> distributed under the BSD license...
>
> Now maybe you're trying to suggest a clause that would prohibit
> distribution under other licenses? That, of course, would prohibit
> distribution under the GPL, since the GPL is another license.
I meant that source code could only be distributed under that new
BSDL-like license. For the purposs of discussion lets call it
Anticopyleft license. You could distribute binaries under the
Anticopyleft license, the BSDL license or any other license you like,
but not the GPL because the GPL requires that source code is available
under the GPL when distributing binaries. This would enforce
non-copyleftability.
A different BSDL-like license (lets call it Copyleft BSDL) that
requires modifications to be distributed under the Copyleft BSDL and
accompanying source code wouldn't, of course, allow its code to be
relicensed as GPL. But would it allow linking? This is not obviously
"no" for me because when you link BSDL code and GPL code you don't
have to distribute that BSDL code as GPL. The package as a whole is
GPL but not the BSDL parts. So what makes a license GPL-compatible? Is
the criterion being able to relicense it as GPL?
If so, a GPL with an added clause saying it can be linked to Qt may
not be GPL compatible... And if that Qt-compatible GPL can be
relicensed as pure GPL, then ANY modification may strip that
Qt-compatibility clause.
Also you didn't say whether you would consider the Anticopyleft a free
license. Do you equate free with GPL-compatible?
> > > > Rethinking about it, it doesn't have to be so. You wouldn't be able to
> > > > paste QPL code into a QPL program anyhow, you'd have to distribute a
> > > > patch that combines the two programs and give both authors rights over
> > > > it. So it IS possible to reuse QPL code as long as it remains within
> > > > the QPL. Am I right?
> > >
> > > Well, you're right that QPL code can be reused. But you're wrong about
> > > this being the defining issue.
> >
> > What about the patch clause? Even if the QPL didn't allow Troll to
> > demand modifications to themselves?
>
> That might be an issue too. But that's less of an issue, it's
> not about basic freedoms.
>
> Note that the requirement isn't that modifications be distributed as
> patches -- patches are just an example of how to fullfill the requirement
> that modifications should be distributed in a separate form from the
> Software -- it might very well be that as long as you can demonstrate
> the ability to distribute both modified and unmodified versions of the
> QPLed software that you're satisfying this clause.
That would make the modification a separatable form rather than
separate. Separate in potential only. Do you think a court would allow
that?
> --
> Raul
>
- Adi Stav
