How To Break The GPL

Rod Dixon, J.D., LL.M. rod at
Tue Mar 7 06:30:28 UTC 2000

Interesting trick. A type of recursive code-fork. Although the examples are
not derivative works because they are not original, I am sure software
example could be.  If the code is inseparable from the freedoms (copyleft),
then Alice's modification of T' (in step 1) travels with the work, if there
is ANY distribution whatsoever. There is not an easy way to get around this
copylef requirement; that's why it is so effective.

As for step 2, document A is an instruction that is not subject to
copyright. In other words, Alice has not created anything. (Her restrictive
license lacks consideration since there is no work that runs with it. Hence,
it has no legal effect).  Alice is just offering directions, and possibly
inducing others to either infringe Trent's work or breach his license, which
is a no-no. Hence, she has not broken the chain to Trent's work. I think
this version of the hypo is too far removed from copyrightable works to help
us with the other hypo.

Rod Dixon, J.D., LL.M.
rod at

> -----Original Message-----
> From: cowan at
> [mailto:cowan at]On Behalf Of John Cowan
> Sent: Monday, March 06, 2000 2:26 PM
> To: license-discuss at
> Subject: Re: How To Break The GPL
> "Rod Dixon, J.D., LL.M." wrote:
> > Well, it is obvious why this hypo does not quite work. Assuming, for the
> > sake of argument, that Alice's derivative work is a french
> translation of
> > Trent's article
> For that case, the trick I have in mind won't work, because a translation
> into another language shares no bits with the original.  But that isn't
> parallel to the hypothetical I have in mind.
> > > Instead of sending him a marked-up copy, she tells him to get Trent's
> > > article himself and then mark it up according to her instructions.
> > This would be an absloute no-no, if you mean Alice instructs
> others how to
> > infringe Trent's copyright or breach Trent's GPL.
> Not at all.  Here's a still more concrete case.  Trent publishes
> the document T:
> "RMS supports the GNU GPL" under the GPL.  (Assume arguendo this
> is a copyrightable
> document, though it probably is not.  Also ignore defamation
> issues.)  Now watch the
> following steps.
> 1)	Alice alters her copy of the document by substituting "MPL"
> for "GNU GPL".
> 	This is a derivative work T'; the GPL allows its creation.
> 2)	Alice now creates the following document A:  "In Trent's document T,
> 	substitute 'MPL' for 'GNU GPL'".  This work depends on
> Trent's document to be
> 	intelligible, but is not technically a derivative work.  Alice may
> 	distribute this work under any restrictive license she pleases.
> 3)	Bob receives Alice's document A and applies what it says to Trent's
> 	document T.  He now has a copy of T' without Alice having
> distributed
> 	either T or T'.  The GPL permits the creation of this
> derivative work.
> 	Bob does not distribute T' either, so neither he nor Alice
> is required
> 	to license A under the GPL.
> > Hmm... for this hypo I do not think this works. Alice's
> instructions on how
> > to make a derivative work out of Trent's article does not sound like
> > copyrightable subject matter.
> Depends on how much work Alice does.  The changes may introduce
> substantial new
> content, unlike the fully-detailed hypo above.  For example, Alice's
> instructions may add a whole new chapter to an existing GPLed book.
> > Even if it is, if Alice's intstructions
> > induces Bob to infringe or breach an agreement, then Alice has
> an entirely
> > different set of problems than we have discussed so far.
> What breach?  Bob is, from Trent's viewpoint, modifying Trent's
> work for his own use,
> which the GPL permits, using Alice's proprietary additions.
> > Well, the problem with the hypo is that it ends too soon. Bob
> may have to
> > make his translation available to Trent even if he does not
> distribute it
> > according to the GNU GPL. In doing so, Alice is out of business.
> As far as I can tell, the GNU GPL does not compel distribution to anyone,
> not even the copyright owner.  Various corporate licenses do compel
> distribution, though.
> > Consequently, even if Alice is able to slide through all of the
> trap doors
> > of the GPL (and you may be exactly right that she may), there is little
> > incentive for doing so. She will not make money. As for
> software, Alice MAY
> > get away with this, but Bob may have to give up her code when
> he adds his
> > derivative work to Trent's original project.
> In this case, Bob is not a contributor, merely a user of Alice's program.
> (He thinks of it as "Alice's program" even though it includes
> Trent's GPLed
> library.)  Alice's installation routines link in Trent's library, but they
> are performed *by Bob* even though written by Alice.  Bob is the creator
> of the derivative work Alice+Trent.
> --
> Schlingt dreifach einen Kreis vom dies! || John Cowan
> <jcowan at>
> Schliesst euer Aug vor heiliger Schau,  ||
> Denn er genoss vom Honig-Tau,           ||
> Und trank die Milch vom Paradies.            -- Coleridge (tr. Politzer)

More information about the License-discuss mailing list