[Top][All Lists]

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

Re: Use of GPL'd code with proprietary programs

From: Alexander Terekhov
Subject: Re: Use of GPL'd code with proprietary programs
Date: Mon, 12 Jul 2004 16:26:45 +0200

David Kastrup wrote:
> Alexander Terekhov <> writes:
> > David Kastrup wrote:
> > [...]
> > > A program can't cause linking of itself by calling dlopen because an
> > > unlinked program is unfit for running.
> >
> > And? A video game is unfit for running without game console.
> We are talking about technical details at the moment, not legal.

Forget technical details. Focus on fundamentals. Read this (written by 
the attorney... and general counsel and secretary of the OSI):

I continue to believe that these confusing messages about "derivative
works" entirely miss the mark.  Where in the statutory or case law can
one find support for such conclusions as are reflected in these

If you don't create "a work based upon one or more preexisting works"
then you have simply not created a derivative work.  17 U.S.C. ยง101.
How in the world does an independently-written piece of software that
communicates with another independently-written piece of software
through a published API ever become a derivative work of that other
software?  Where in the GPL does it say that it can become a derivative

Nothing in the Copyright Act addresses the *use* of software in this
way.  If the GPL is enforced under the copyright law, then how could a
court ever conclude that it reaches to such API-connected pre-existing
works that merely get used together?

I think your response would be appropriate if the GPL were a contract
rather than a mere copyright license.  The GPL is intended by its
authors to be interpreted and enforced under copyright law.  There is no
basis in that law for the definition of "derivative work" that is
implied by the GPL language you quoted.  How can you assume that a
licensee accepted such a broadened definition of "derivative work"
absent his assent to a contract?

The MPL also attempts to apply to more than derivative works.  While I
don't particularly understand its reach, at least it is to be enforced
as a contract and thus the definitions in that contract are relevant.
That, for me, is the essential difference between a copyright license
and a contract.

The GPL can't do more than copyright law allows -- because that's how
its authors want it to be treated.  I don't understand how GPL licensors
can benefit from contract provisions.

> Assume that someone statically links 
> object modules compiled from G and object modules compiled 
> from H into a single executable file (call this executable file G+H).
> I believe that there is wide agreement that the GPL is 
> interpreted such that the author of G has not given 
> permission for distribution of that single executable file. 
> (I also believe there is less widespread agreement on the 
> alternative where the linking occurs at runtime.)

I don't know that there is widespread agreement to either of those
propositions.  Indeed, isn't that really what we've all been discussing?
[I must admit, I once publicly argued the point your way, but I have
since recanted because I couldn't find any statutory or case law support
for my earlier position.]

> H is not a derivative work of G. So, how does one get to this 
> widely agreed result? I believe that that interpretation 
> assumes that G+H is a "work based on the Program". So, it 
> looks to me like it is generally agreed that the GPL does 
> indeed concern itself with whether G and H are parts of 
> something larger (not necessarily every larger thing, but at 
> least some sorts of larger things). Thus, it seems that 
> stopping analysis at the point of determining that H is not a 
> derivative of G is failing to complete the analysis needed to 
> judge compliance with the GPL. 

I am still not certain what is meant by the phrase "work based on the
Program."  Under your scenario, G and H are entirely independent
creations.  If G+H requires merely the making of copies of G and H, an
act permitted without restriction by the GPL, then why is it a
derivative work?  Why is that a work based on the Program?


> You keep returning to contract obligations. But, I'm not
> relying on any contract obligations. Any distribution that 
> includes copyrightable material from B needs the permission 
> of B's copyright owner. The hypothetical that I've presented 
> includes distribution of B. Thus, B's permission is needed. 
> I'm trying to understand the conditions the copyright owner 
> has attached to the copyright owner's offer of permission to 
> distribute B (the conditions in the GPL). So, the conditions 
> specified in the GPL are relevant to what someone needs to do 
> in order to legally distribute A+B, without regard to whether 
> A+B is has some special status as a protected copyrightable
> work (B's protectable status is enough).

I keep returning to contract obligations because under copyright law
there are only a limited set of exclusive rights that may be licensed,
namely to make copies, prepare derivative works, distribute copies,
perform and display.  Where in the statute is there any reference to an
exclusive right to make a "work based on the Program" or a "Larger
Work"?  How is a court to interpret those phrases?  Why should the court
even try to do so?  Are those things more than a derivative work or
less?  Why is the licensor's interpretation of those phrases in any way
binding upon licensees?  

The GPL grants an unlimited right to make copies but a conditional right
to make derivative works (with some other words in the license about a
limited right to make a "work based on the Program").  The only way a
judge can interpret that license is to determine whether what is being
made -- your A+B, for example, where A is the GPL-licensed work --
involves making a copy of A or creating a derivative work of A.  If the
former, then the license is clear that there are no reciprocal
obligations.  If the latter, then the license is also clear that the
author of A+B must disclose his source code.  

That's the question we've all been struggling with.  Does linking
require merely the making of a copy or is it the creation of a
derivative work?  We're now back at square one.  

How does the language you quoted from section 2 help the judge perform
that analysis?  Why should the judge care at all that those other words
are in the license, given that there is no proof whatsoever that the
licensee either read or assented to that extra language?  

If the GPL is just a copyright license then none of that extra language
matters.  The only question is, has a license been granted to make a
copy or to create a derivative work?

> OK, so I thought the GPL distinguished between the two - that 
> having a GPL program (I'm not thinking of the kernel here or 
> other things reasonably determined to be part of an 
> "operating system", an allowance the GPL
> makes) on the same CD as non-GPL bits, in a situation such as 
> a Red Hat Linux CD, was OK because it was "mere aggregation", 
> which the GPL explicitly allows, and not a "collective work", 
> which the GPL states
> *would* be under the GPL.  Maybe "mere aggregation" has no 
> meaning w/r/t copyright law, but am I mistaken in thinking 
> the GPL makes the distinction?

I don't understand these subtle distinctions people are reading into the

Section 2 of the GPL grants permission to "modify your copy or copies of
the Program or any portion of it."  In that context, I have never
understood the reference within that section to "the right to control
the distribution of ... collective works based on the Program."  A
collective work is defined clearly in copyright law and is different
from a modified (or derivative) work.  One does not modify a work in the
course of creating a collective work. 

If one merely copies the original work unchanged, that falls under
section 1 of the GPL, not section 2.  

Those words in section 2 dealing with "mere aggregation" seem out of
place.  I'm even more confused about the words "work based on the
Program," but I've addressed that before and won't repeat it now.

What are the FSF's arguments, dak? Your Emperor Has No Clothes.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]