Re: KDE not in Debian?
Chris Lawrence wrote:
> If you have something to say, say it to the lists.
Sorry, I was trying to get you to respond to the particular issues I had made
rather than continue to make the generalized statements "It just isn't so" or
"The GPL requires this" w/out bothering to indicate where in the GPL this is
required. But, alas, I have failed :-(. While I could respect your different
reading of the GPL, I cannot see how your reading of the GPL allows linking with
XFree code but not Qt code. To date, nobody has explained this to me, except by
claiming that the XFree code can be licensed under the GPL. When I went through
a thorough exercise of showing why this in fact can't be done (my post bearing
Message-ID <38955EF0.52C053CC@mieterra.com>), nobody has responded, perhaps b/c
you agree that I am right.
So you really have a dilemna here: how can you read the GPL to permit linking
GPL code with X code while at the same time preventing linking GPL code with QPL
code?
> (This will be my last post on this topic, barring egregious factual
> errors that need correction---mine or those of others.)
>
> > You haven't required it, AFAIK, from Gnome or other programs that
> > link with X. And under your reading of the GPL (at least if you
> > agree with the others I have been debating this issue with), if Qt
> > is incompatible with the GPL, so is XFree. Oh right, it would
> > really suck if you couldn't distribute XFree, so you can just ignore
> > that transgression. Or am I missing something? (Please respond to
> > my post with Message-ID <38955EF0.52C053CC@mieterra.com> so I don't
> > have to drown this list in repeating it).
>
> XFree is not distributed under a license more restrictive than the
> GPL. Qt is. That's it. End of story.
Hmm, the problem I have is that I do not see language in the GPL that says
anything about code being more restrictive being a criteria; you wishing it to
say that does not make it so. If you refer to Section 6's requirement that you
impose no additional restrictions, here again is the language:
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. You are not responsible for enforcing
compliance by third parties to this License.
Now, let's look at this closely. The provisions kick in when you distribute the
"Program" or "any work based on the Program". The former part would be
distributions under Section 1 and the second part distributions under Section 2
(and of course both could apply to Section 3, depending on whether changes were
made).
Now what happens when you do this? The recipient gets a license from the
Program author (i.e., for the GPL'd code) to copy, distribute or modify "the
Program", subject to the terms and conditions. This would not include the added
part, as is obvious since the original licensor cannot grant a license w/r/t the
added part, not being the author.
Next, it says you cannot impose further restrictions on the rights granted
herein. My question is, where does Qt impose further restrictions on the rights
granted "herein"? Well, that boils down to this question: what rights does the
GPL grant with respect to the "added" code?
This is where the heart of this debate is. Some people see Section 2 as
requiring that the modified part be licensed under the GPL. However, as I have
pointed out in my other e-mails, and which nobody has credibly disputed, XFree
code cannot be licensed under the GPL. Thus, if you read Section 2 thus, you
cannot distribute X code with GPL code. However, Debian does this.
The other way to read Section 2 is how I do, that it only requires that the
source to the modified code be made freely available to all third parties and
that no charges be made for redistributing the modified code. Both Qt and X
code comply with these requirements.
So I am interested to know, even if I agreed with the statement, why it is
relevant that "XFree is not distributed under a license more restrictive than
the GPL. Qt is." Even if that is true, you have not given any reason why that
would make a difference under the GPL.
I realize this license exercise is not simple. If you don't understand how all
this fits together, that's OK. But in that case you should not criticize others
for their reading of it, or claim that they are doing something wrong.
> >From section 2 of the GPL:
>
> These requirements apply to the modified work as a whole. If
> identifiable sections of that work are not derived from the Program,
> and can be reasonably considered independent and separate works in
> themselves, then this License, and its terms, do not apply to those
> sections when you distribute them as separate works. But when you
> distribute the same sections as part of a whole which is a work based
> on the Program, the distribution of the whole must be on the terms of
> this License, whose permissions for other licensees extend to the
> entire whole, and thus to each and every part regardless of who wrote
> it.
>
> There are no provisions of the XFree license that stop XFree code from
> being aggregated with GPLed code (the result being GPLed).
Yes, I have shown that there are. The XFree license says that if you
redistribute XFree code you have to include the XFree license. It does not
permit you to relicense the XFree code under a different license. Of course, it
does not speak to other code that you add, you can license that as you want, but
nonetheless you cannot change the license of the XFree code.
I will also note that there is nothing in the QPL which prevents QPL code from
being aggregated with GPL code.
> By
> contrast, the "QT FREE EDITION LICENSE" (which is what Qt 1 is
> distributed under) specifically forbids the modification of Qt:
>
> Your software does not require modifications to Qt Free Edition.
>
> Nothing in the X license forbids modification of X by third parties.
Right, you have shown that the X license allows modification by third parties
but the Qt 1.x license does not. But you have not shown why this particular
difference is relevant under the GPL.
> You are also forbidden from distributing modified versions of Qt1;
> nothing in the XFree license prohibits distribution of modified
> versions (see, for example, xfs-xtt, which is based on XFree's xfs
> implementation).
>
> As far as the QPL (Qt2) license goes, I leave the discussion to Joseph
> Carter, who actually helped Troll write it (with the specific intent
> of allowing Qt2 applications to be "pure GPL") before the lawyers got
> involved. My belief is that the provisions that might require you to
> give your source code to Troll, even if the binary code is not
> distributed to others, were the most egregious
Where does it say that? I think you are referring to Section 6(c), but Section
6 explicitly says that clause (c) applies only when you distribute your work.
> (the GPL only obligates
> you to give source code to people you give binaries to; if you don't
> give someone binaries, you don't have to give them source either).
> For example, if I modify Emacs, RMS can't demand that I give him my
> changes unless I give him a binary of my modified Emacs; however, if I
> make a modified Qt2 (to create my own whizz-bang desktop that only I
> use), or even a program based on Qt2, Troll can demand a copy of it.
> That seems more restrictive than the GPL to me; I'm amazed it even
> meets the DFSG.
Of course QPL is different. Again, you have not shown how that difference
matters under the GPL.
> Anyway, if you've followed this thread, you know that Debian has also
> thrown out (or not let in) other software with ambiguous license
> terms. XForms and Motif-based programs have gotten the same
> treatment (as have Qt-based programs from people other than KDE).
Well, a recent post shows this has not been the case, that in fact a number of
QT-linked programs, including to qt-1, were included, although apparently now a
bug report has been filed.
Ciao,
Andreas
Reply to: