[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: PHP-Nuke License Conclusion?

Scripsit Branden Robinson <branden@debian.org>
> On Tue, May 13, 2003 at 02:35:05PM +0200, Henning Makholm wrote:

> > I wonder how the arguments I pointed to came into being, then, if I
> > did not construct them.

> Which arguments?


> You keep saying they exist

I keep giving them.

> > > Okay, so you and Nick Phillips will cotton to "the GNU GPL" being a
> > > variable rather than a constant,

> > Strawman.

> How so?  You and Nick are claiming that the "terms and conditions" of
> the GNU GPL as applied to a work are not always the literal text of what

I say that when one constructs at cut-and-paste licence, then the
words "this license" obviously refers to the entire cut-and-paste
license, regardless of from where those words entered the
cut-and-paste license. 

> Thus, the "terms and conditions" -- even in written form -- of the GNU
> GPL as applied to a given work can vary,

No they can't. For at given work they always refers to the license
covering that given work.

> No, that's an incidental problem.  My point is that whatever a license
> is that is the GNU GPL plus additional restrictions (consistent or not),
> it isn't the GNU GPL.

That is true. Nobody is, as far as I am aware, disagreeing with that.

> This is because the text of the GNU GPL is self-referential, and
> talks about "these terms and conditions" (clause 6).

If one creates a new text by using the GNU GPL, the new text becomes
self-referential too. It does not stay GPL-referential, because what
happens when one constructs a cut-and-paste license is *textual*
inclusion-by-reference. One does not include the *meaning* of the GPL;
only its *text*.

> What the author chooses and what he manages to express are not always in
> consistent harmony.  Copyright licensees are bound first and foremost by
> the licensor's explicit terms and conditions.  Licensees who operate
> under a "gentlemen's agreement" with the licensor in which they violate
> the literal terms of the license do so at their own peril.

Exactly. And when an author specifies that the work is goverened by a
cut-and-paste licence, then the text of said cut-and-paste license
*is* "the licensor's explicit terms and conditions".

> > On the contrary, it is obvious that if an author knowingly and
> > explicitly licenses is work under "the conditions of the GNU GPL, with
> > the following additional restrictions: [bla bla]", then the author
> > intends every occurence of "THESE TERMS AND CONDITIONS" to refer to
> > his entire combined license conditions.

> You're drawing an inference.  It is not obvious.

I think it cannot be more obvious. It is the only way to give the
author's utterings any sensible meaning.

> When an author says to me:

But that is a bad example. What the author is saying to you is

   You can copy and distribute this work under certain conditions.
   To find out what those conditions are, take the GNU GPL and add
   to it the requirement that you can't do baz.

> ..then I become confused because he is simultaneously telling me I can
> and cannot do "baz".

Yes, of course, because your'e quoting a horrible way of doing it.
As far as I'm concerned, this discussions is still not about "is there
a horrible way of doing this?" but "is there a valid way of doing

> > >   Observation Three
> > >   -----------------
> > >   The FSF does not want people "forking" the GPL, and think that license
> > >   proliferation is a bad thing.

> > This is true, but it does nothing to argue that a fork would be
> > internally inconsistent.

> And I have not made such an argument.

Then I don't see why we are arguing at all?

> People can fork the terms and conditions of GNU GPL all they want as
> far as I'm concerned, but if they create a license document that is
> self-contradictory

Since you "have not made such an argument" [that a cut-and-paste
license is self-contradictory], then what's the point?

> > >   However, in practice, most people don't do what the FSF recommends in
> > >   their FAQ.  They just say "it's under the GPL" and slap extra
> > >   restrictions somewhere else -- perhaps in a README file.

> > Which I can easily agree is internally inconsistent and may make the
> > work undistributable.

> Okay.  You can easily agree -- but you're not going to???

I *am* agreeing that the particular practise you describe is
internally inconsistent. I do not agree that this is also the case for
every *other* way of creating a (non-free) licence by adding
restrictions to the GPL.

> > But this still does not mean that adding additional restrictions is OF
> > NECESSITY internally inconsistent.

> It is if they use the GNU GPL's terms and conditions as-is.

I'm not talking about cases where people "use the GNU GPL's terms and
conditions as-is". I'm talking about cases where people *explicitly*
do *not* use them as-is, but only as an expository aid for describing
what they mean.

> > The words you used were

> > || Debian interprets "this License" and "herein" to mean the
> > || conditions of the GNU GPL expressed in its text; no more and no
> > || less.

> > which does not seem to me to acknowledge that there is a right way to
> > add extra restrictions. (Right as in "probably legally sound", not
> > right as in "free").

> Of course there is.

In that case I've misunderstood your original quote. Care to explain
what it means, then?

> It will be if they use the GNU GPL's terms and conditions as-is.

See above.

> > But I still do not think you have offered any convincing arguments
> > that "GPL with the following additional restrictions: [blah blah]"
> > is a bad name to rename it to.

> Well I think I have.  Such a license, whatever it is, is not the GPL,
> and will not be GPL-compatible[1].

Is that an argument? Of course it is not the GPL and is not
GPL-compatible. How does that make ia a bad idea to call it "GPL with
the following additional restrictions: [blah blah]"? That name very
explicitly tells the reader that the license it applies to is
*different* from the GPL.

> > On the contrary, I think it is the honest way to go about it, rather
> > than reproducing the entire load of legalese as "the Foobar Public
> > License"

> A license name one hasn't seen before is a damn good reason to read it
> closely.

Yes, but it is much easier to do so if the licensee does us the favor
of exlaining how his licence differs from something we know in advance.

> > and hiding the incriminating

> Eh?  Why are you suddenly alleging criminal activity?

I'm speaking figuratively here. s/incriminating/non-free/ if you want
to be formal.

> > restrictions in the midst of vaguely-familiar licensing text that
> > nobody is going to read closely anyway because it just looks like a
> > revamped GPL anyway.

> We here on debian-legal will read it closely.  We'd better!

Yes, but that does not mean that we always succeed. There has been
several cases recently where we found that some licence we believed
was free is really not, because we did not read it closely enough
the first time.

I don't thin we should encourage licensors to write their licenses in
a way that will make it more difficult than necessary for us to spot
non-free features of it.

> > It is clear as day that if the author starts his source files with

> > /* Copyright YYYY N.N.
> >  * This file can be copyed freely under the GNU general public license
> >  * with the added restriction that everyone who distributes it (or
> >  * derived works) must pet a cat.
> >  */

> > then OBVIOUSLY he means for "THESE TERMS AND CONDITIONS" to include
> > his additional restriction.

> Except elsewhere, when speaking with his own voice, he says between the
> "END OF TERMS AND CONDITIONS" that "Activities other than copying,
> distribution and modification are not covered by this License; they are
> outside its scope."

Yes? That is not incompatible with the added restriction of the
permission to distribute. The fact that the permission only applies to
people who happen to pet cats does not make the permission itself
cover other activities than distribution.

> > And it is what the copyright holder means that counts.

> Certainly, where that meaning can rationally be deduced from the words
> he uses.

And it can.

> But if he hands me a logical paradox and calls it a license,

There is no logical paradox here.

> > It does not matter what those who happened to write the words that the
> > author chose to refer to to partially explain what he intends meant,
> > as long as the author's intentions demonstrably differ.

> I don't question that they likely differ.  Discerning exactly what he
> *does* mean is another matter entirely.

Exactly what he *does* mean is easy to see. There is only one way of
interpreting the author's words. Why do you not think that is the one
we should use?

Henning Makholm                             "I've been staying out of family
                                   conversations. Do I get credit for that?"

Reply to: