Re: defining "distribution" (Re: A few more LPPL concerns)
On Sun, 2002-07-21 at 16:49, Boris Veytsman wrote:
> > Date: Sun, 21 Jul 2002 14:32:39 -0700 (PDT)
> > From: Mark Rafn <firstname.lastname@example.org>
> > > Suppose I take a GPL'ed program, change it and put the closed version
> > > (sans sources) on my own machine. I did not violate GPL yet. Now
> > > suppose that I make the drive NFS-exportable and encourage my paying
> > > customers to mount it and access the program from there. Would I
> > > violate GPL? I think yes.
> > Absolutely. This is distribution in the classic sense. No additional
> > definition of "distribution" is necessary.
> Note that in this case the customers might not copy the program but
> just execute it from this location.
This doesn't matter. The fact that you have allowed remote access, and
the third party has taken advantage of it, makes it distribution.
> > It's not distribution for me to install a package on a system I administer
> > (or just have an account) and allow others to execute it. They can
> > "access" it in terms of execution, but if they copy it, they do so without
> > my permission (and without yours).
> This is the root of our disagreement. I think that a sysadmin that put
> a changed copy of latex.fmt in the $TEXFORMATS directory to be used by
> his users, *distributes* a changed LaTeX. You think he does not; the
> problem with your theory is that it undermines both the intentions of
> LPPL AND GPL. You see, there is no reasonable difference between a
> sysadmin who put a closed copy of a GPL'ed program in /usr/bin, and a
> cunning manager who made this program NFS-accessible "for execution
> only" by the people paying a fee.
Sure there is. In one case, a single computer can use the resulting
binary; in the other, multiple computers can, after paying an access
> When I execute a program, this is not a distribution. When I allow
> others to execute it, I distribute it -- even if there is no actual
> copying of bits between magnetic media.
Actually, it's not clear that this is true. For example, technically a
CD player "makes a copy" of a CD into its internal memory in order to
play the music encoded upon it. This is not considered to be "copying"
because this copy is required in order to gain access to the copyrighted
work you supposedly paid for when you bought the CD. In other words,
playing a CD is a "fair use" copying right, and I can execute that right
whether I'm the person who plunked down $20 for the particular CD or
(Of course, the whole mess that is music distribution rights is way
beyond this discussion.)
This is different from the so-called "clickwrap" licenses because you
generally have to agree to those at *installation* time, not every time
you run the program. Of course, clickwrap licenses have their own
To UNSUBSCRIBE, email to email@example.com
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org