Subject: RE: GPLv3 draft
From: "Lawrence Rosen" <>
Date: Tue, 11 Apr 2006 14:05:49 -0700

Forrest Cavalier asked:
> How much of a problem is that the license means what the offerer meant,
> not what the drafter meant?  

A license means what the parties to the license agree it means, usually as
expressed by the clear and unambiguous words of the license itself. The
offeror (licensor) is merely one of the parties. It is equally important to
ask what the licensee meant when she agreed to it. In the case of the GPL,
licensees around the world frequently say that they find the license
confusing. Just listen to them speak.... I hope that GPLv3 will ultimately
help eliminate that confusion.

When there is a dispute between the parties over the meaning of a license as
applied to specific facts, a court may be asked to resolve that dispute.
Courts use various techniques to divine the intent of the parties when the
language of their agreement is vague and ambiguous. Widespread and
oft-reported confusion about the meaning of a license doesn't help a court
interpret that license fairly.

No court has yet addressed squarely the interpretation of the GPL. I'm not
looking to ask a court, either, because judges are not experts in our field.

I've never known a court to care what an attorney meant when he drafted a
license (unless the attorney is also a party to the license, or unless the
attorney is party to a malpractice lawsuit over sloppy drafting).

The solution is for the GPLv3 clearly and unambiguously to say exactly what
it means so that nobody will be confused.


Lawrence Rosen
Rosenlaw & Einschlag, a technology law firm (
Stanford University, Lecturer in Law
3001 King Ranch Road, Ukiah, CA 95482
707-485-1242  *  fax: 707-485-1243
Author of "Open Source Licensing: Software Freedom and 
                Intellectual Property Law" (Prentice Hall 2004)

> -----Original Message-----
> From: Forrest J. Cavalier III []
> Sent: Tuesday, April 11, 2006 6:58 AM
> To:
> Subject: Re: GPLv3 draft
> Lawrence Rosen wrote:
> > Ben Tilly wrote:
> >
> >>It is far better to have a single license written in one language that
> >>is as clear as possible.  Particularly if said license is written in
> >>such a way that it will almost certainly be valid in any country that
> >>abides by international agreements on what contract and copyright law
> >>should look like.  And for software development, for better or for
> >>worse, the natural language to choose for this purpose is English.
> >>
> >>Which is what the GPL does.
> >
> >
> > More correctly, this is what GPL attempts to do but hasn't succeeded
> yet. As
> > a license -- even in its latest draft -- it isn't yet "as clear as
> possible"
> > or "almost certainly ... valid in any country." Wishing that result
> doesn't
> > make it so.
> >
> When the first big GPL lawsuit made it into court, (was it MySQL?)  I seem
> to remember lots of people were concerned that it get decided the
> "wrong" way and thereby open up GPL "abuses/exploits"?
> I seem to remember a collective sigh of relief when that didn't happen.
> How much of a problem is that the license means what the offerer meant,
> not what
> the drafter meant?  Sure, the FSF seems very on top of offering legal help
> in cases when it matters, but they are not copyright holders for all
> GPL'ed
> code, and therefore do not get to defend or sue.
> To really undermine the GPL, could some company (in say, Redmond,) release
> some
> GPL'ed code, sue a violator, and as plaintiff pick some unfavorably weird
> interpretation of the GPL, intending to lose the court battle on purpose
> to set
> unfavorable precedent?
> I think SCO v IBM is just a skirmish, but it goes to show that even things
> you think are slam dunk true can get disputed by flat-earthers with PIPE
> or otherwise funny money to spend on litigation.  Am I wacko to think this
> is
> a threat?