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

Re: GPL compatibility of DFCL

On Thu, Jun 13, 2002 at 03:49:27PM -0700, Mark Rafn wrote:
> On Thu, 13 Jun 2002, Branden Robinson wrote:
> > As noted elsewhere, I'm planning on a "GPL conversion clause".  This
> > would permit the omission of the endorsements notice.
> > Actually, it would only "suspend" it.
> How could this possibly work?  If I get a work under the GPL, how can 
> some other license possibly take away my GPL permission to modify and 
> distribute the work?

Uh, it seems pretty easy to me.

Let's say I write a manual for a GPLed piece of software called Foo.
However I don't contribute this directly to the Foo source code base for
any number of reasons: maybe the work has been abandoned upstream; maybe
the author/maintainer is unresponsive; maybe the author and I are having
a spat, and while I love Foo, I find the author impossible to deal with.

So, I write a nice big manual for Foo, and license it under the DFCL.
There is an endorsement notice right under the copyright notice, and I
put my name there in my original release.

Time passes.  Someone gets the clever idea to incorporate my Foo manual
into the Foo source distribution so that all Foo users will be able to
access it easily.  Maybe they need to make changes to the content of the
manual, maybe not; if they do, they need to remove my endorsement from
it.  If they don't, they can retain or get rid of my name in the
endorsement notice at their option.  They then turn to clause X of the
DFCL and see "if and while this work is incorporated into a different
work which is licensed under the GNU GPL, version 2, as published by the
Free Software Foundation, the reproduction of the endorsement section
immediately after this work's copyright notice is optional instead of

Hmm, probably want to extend that to include the LGPL.

> > The idea being that you can't take a DFCL-licensed work, cram it into a
> > piece of software, and then suck it back out into a data file without
> > having to restore the Endorsements section.
> I'd love to hear proposed language for this.  At first blush, I don't see 
> ANY way to do this that doesn't leave it easy to turn the license into
> pure GPL.

I think you're thinking about this completely the wrong way.

1) a copyright work doesn't stop being copyrighted (or licensed) when it
is combined with another work
2) the GPL does not trump or change the license terms of any other
copyrighted work; it simply establishes conditions for modification and
distribution of the work licensed under the GPL, which may or may not be
3) you can distribute a non-GPLed work (such as an MIT or 3-clause BSD
licensed library) along with a GPLed work as long as the license on that
non-GPLed work does not run violate any of the GPL's terms.

I am as incredulous about your inability to perceive the compatibility I
am proposing as you are incredulous that it's possible.  Clearly one of
us misunderstanding something.  :)

> At the very least, I can relase a 1-line GPL program whose source I append
> to the work, keeping it GPL.

Possibly, but keep in mind that a court will not recognize a claim of
copyright on something that is too short to be an original expression.
This is why names and titles are not copyrightable.

> It seems possible that I could just claim that my copy of the work is
> a derivative of the GPL program I took it out of and therefore still
> is distributable itself under the GPL.

Wrong.  Recall:

1) a copyright work doesn't stop being copyrighted (or licensed) when it
is combined with another work

You can *claim* whatever you want, but in this case you'll be wrong.

> > Contrary to popular belief, the GPL does not change the copyright on
> > anything.  It merely requires that you not impose restrictions beyond
> > its own terms.
> Right.  Like saying that a derivative of the software (with everything but 
> the work in question removed) cannot be distributed under the GPL.

Huh?  I'm not seeing your point here.

> > The point of my tangent is this: we don't *need* statements like
> > "Neither the name of the University nor the names of its contributors
> > may be used to endorse or promote products derived from this software
> > without specific prior written permission." in our licenses.
> Aren't we then back to "we don't need a new content license."  Any free
> software license works UNLESS you need a (positive or negative)
> advertising clause.  (as long as you define software to be "a sequence of
> bytes")

This is actually pretty close to what I personally believe; however,
there are at least three complications:

A) The GNU GPL is not written in terms broadly enough to apply to all
creative works; to apply something that purports to be a software
license to something that isn't software invites argument and could land
the license in court.

B) Not everyone believes that all information can be expressed as a
stream of bytes (if sufficiently well-described), and therefore the
proposition that software == data == music == sculpture is lost on them.
I'd rather write a license that artists with intentions of contributing
to a creative commons can accept than have philosophical arguments with
all of them.  Maybe they believe in the existence of analog wavelets
below the threshold of resolution of modern physics (thanks to
Heisenberg's uncertainty principle); this belief informs their muse and
is the wellspring of their inspiration.  Who am I to argue?  It is more
important to me that they contribute their art to the world than that
they subscribe to some form of digital empiricism.  Maybe my model is
more accurate than theirs, but I don't require philosophical concordance
with my beliefs of the musicians I listen to.  I just wanna hear a good
guitar solo, man.

C) The FSF itself is not willing to endorse the usage of the GNU GPL as
a license for creative works in general.  This is the most serious
practical problem.  The FSF does not appear to share Debian's (vague)
consensus that a GPL-style copyleft as as good for data as it is for
algorithms.  Their solution, the GNU FDL, doesn't work for us in the
general case (when people employ Cover Texts and Invariant Sections).
So what do we do?  We tell people what *does* work for us.

> Unfortunately, there is a perception among people who are generating 
> "free" content that they DO need such a statement.

If we don't cause any harm, or the benefits outweigh the costs, I think
we should accommodate them.  See B) above.

> > We *could* go a route that only precludes and disavows endorsements,
> > e.g.,
> > 
> > 	Copyright (C) 200 Foo Bar
> > 
> > 	This work may be freely copied, modified and distributed under
> > 	the terms of the Debian Free Content License.  The contents of
> > 	this work are not necessarily the responsibility of the
> > 	copyright holder(s) and do not necessarily represent the views
> > 	of the copyright holder(s).
> I like it a lot.  But then I like pure GPL, X, or my favorite 
> license of all time "This work is public domain". 

I think asking people to pull all their work in the public domain is
asking too much of them.  Disney will take their idea, make a hundred
million dollars off of it with copyright law, and laugh all the way to
the bank.  I can't expect that degree of selflessness from people.

> > *However*, I think we could make the DFCL more appealing to
> > authors if we did actually permit the inclusion of endorsements. 
> Is there a way to gather some feedback from a sample of authors who we 
> hope will use this license?

There are a few forums I know of where I can shop this idea around.  I
want to have a draft first, though.  As I said earlier, Debian needs to
have something it can live with first.  Then we shop it around to
potential users.  If they don't all throw up, then we run it by an IP
lawyer and find out just how much I suck as an armchair lawyer.  Then,
assume he doesn't treat it the way Linus treats documentation[2], we
announce it to the world and wait for the MPAA, the RIAA, and/or the AAP
to decry us as a bunch of communists who will only encourage terrorist
attacks on the United States[3] by promulgating our license.  If they do
that, we'll know we're on the right track.

[1] Erik A. Young does not understand this, which is why I am so
contemptuous of the libeay license.

[2] http://old.lwn.net/2001/0118/a/lt-documentation.php3

[3] http://www.vnunet.com/News/1132563

G. Branden Robinson                |          Measure with micrometer,
Debian GNU/Linux                   |          mark with chalk,
branden@debian.org                 |          cut with axe,
http://people.debian.org/~branden/ |          hope like hell.

Attachment: pgptTDfOxE7y4.pgp
Description: PGP signature

Reply to: