Re: On interpreting licences (was: KDE not in Debian?)
On Fri, 11 Feb 2000 12:34:29 -0500, Raul Miller <moth@debian.org> wrote:
> On Fri, Feb 11, 2000 at 05:26:47PM +0100, Marc van Leeuwen wrote:
> > Nobody in this "discussion" is claiming (as far as I can see) that
> > by some subtle shuffling of pieces you can get a (composite) program
> > from A to B without requiring permissions from all copyright owners.
>
> It seems to me that that's exactly what people are saying when they
> claim that it's legal to distribute kghostscript.
>
> > The point is that the complex conditions in the licences, in partcular
> > GPL, may lead to a situation where such permission may be obtained for
> > some particular method of distribution, but not for some other method.
> > For instance, you may distribute in source form (and under GPL) a
> > GPL-ed an application that links to Qt (because then there is no
> > requirement to distribute "complete sources"), while also distributing
> > Qt (in source and binary) under QPL; the recipient could compile and
> > link an executable program which does not "happen to spring into
> > existence", and which was the goal of distributing the GPL sources,
> > yet which could not have been legally distributed directly.
>
> Certainly: it's only if you distribute executables or object code that
> there's any requirement to distribute the complete source for the program.
>
> > I think you put forward yourself another example (a Solaris-linked
> > GPL binary and Solaris itself), where the components might be legally
> > distributed separately (assuming permission from Sun) but not
> > together.
>
> Right: that was an example of a situation which takes advantage of the
> special exception in section 3 of the GPL.
>
> > So please don't suggest any more that people are trying to evade
> > copyright law when in fact they are trying (maybe by jumping through
> > hoops) to abide by the conditions put forth in the licence(s).
>
> Are you now claiming that it's legal to distribute kghostscript?
Yes, definitely, if you are distributing sources; from your remarks I conclude
that even you would agree to this. One could include scripts to compile and
link (say statically, just to get the worst case) those sources onto a
complete executable; since all the tools necessary to do that could also be
(and are in fact) legally distributed by the same distributor, the effect is
that this executable will be identical, bit by bit, to an executable that the
distributor could assemble. So the exact same effect is achieved, although
slightly less efficiently, as if the distributor had directly distributed that
executable file. But, we agree, the latter would certainly not be permitted.
Probably your question was about another method of distribution than in source
form. But instead of diving for the umpteenth time into matters about which we
(and many others) have already amply demonstrated to have differences of
opinion, let me just restate that on which we do seem to agree. If our goal is
to give users access to an executable version of kghostscript, then there is a
reliable method of doing that (distributing sources) that satisfies all the
requirements in the relevant licences, while there is another method
(distributing the statically linked executable file directly) that does not
satisfy all requirements, in particular not those of the GPL. We arrived at
this conclusion not by trying to evade copyright law, but merely by reading
carefully the conditions of the GPL.
Now apart from distributing sources and distributing a statically linked
executable, there are other methods to give users access to an executable,
varying in the amount of work do be done at the recipients end. Offhand I can
think of distributing compiled but unlinked object files, or distributing
dynamically linked object files; maybe other possibilities exist as well.
Somewhere in this range of possibilities there is a dividing line between
legally possible and not legally possible. If you ask me where I personally
would put the line, then I would say: probably between the dynamically and the
statically linked files; I arrive at this conclusion not by general
considerations of copyright law, but by carefully reading the text of the GPL.
I will accept that differences of opinion can exist here, in particular since
it would seem that the GPL (and LGPL) was formulated without the distinction
between static and dynamic linking in mind, making it highly unlikely that it
was intended that one case be covered but the other not. But even if an
advocate of "legally possible" would give up the position of dynamically
linked executables (did anybody ever give up a position in this discussion?:-)
he could still maintain that of unlinked object files with fresh arguments
(such as: GPL 3 starts with "You may copy... the Program... in object code OR
executable form" and later "For an executable work, complete source code
means..." which apparently does not apply to unlinked object code). And then I
don't even mention the possibilities of actually modifying the kghostscript
program itself to avoid problems, for instance by ensuring that the GPL-ed and
QPL-ed code do not execute in the same process. Which is just to indicate that
these are all just battles in a desert war where no party has any chance of a
complete victory, even if in the unlikely event that the opponent concedes
defeat on a particular issue. I guess I had better get back to work.
Marc van Leeuwen
Universite de Poitiers
http://wwwmathlabo.univ-poitiers.fr/~maavl/
Reply to: