[R] licensing of R packages
Berwin A Turlach
berwin at maths.uwa.edu.au
Fri Nov 14 17:01:30 CET 2008
On Fri, 14 Nov 2008 11:07:11 +0000
"Barry Rowlingson" <b.rowlingson at lancaster.ac.uk> wrote:
> A "strict" interpretation of the GPL does not stop numerit from doing
> what they do. They do not distribute the GSL in any form. They tell
> you to go get the GSL dll from somewhere.
> This misconception of the license terms comes about because of the
> use of the word 'use'. If I distribute a short C program that has a
> call in it to a function that has the same name as something in the
> GSL, does my C program use the GSL? No. Maybe it _mentions_ the GSL,
> but the GPL has no problems with that. I'm distributing my C program,
> and not the GPL-covered code, so I can license it how I like.
I do not believe the situation is as easy as you put it. But, IANAL
and the last time I entered a discussion about the GPL I was reminded
that the GPL is about distribution rights and was corrected on some
But I remember that a situation as you describe was hotly debated on
gnu.misc.discuss in the mid-90s; thus, I am talking obviously GPL 2.
Unfortunately I do not remember which software/company was involved and
how the dispute was solved. But they did something like you
described: distributed a binary and asked the user to download
additionally some GPL software and then run both together. If this
were allowed, the GPL would have a hole in it that you could drive a
truck through. :)
As far as I remember the discussion, it was deemed that if the software
that you distribute relies on libXXX and there are several
implementation of libXXX (including one that is GPL'd), then there is
no problem of saying to users "here you are, you can download my
software [binary only] but you will need additionally libXXX to run
it. libXXX you can either get commercially from YYY or a GPL version
But if only one implementation of libXXX exists and it is under GPL,
then your software is pretty clearly a derivative of libXXX; in
particular if it cannot be run without libXXX. Hence, you are bound by
the GPL. The argument is more or less that when you compiled the
binary on your machine, requiring the GPL'd libXXX, you created a
derivative work and to distribute this derivative work would require
you to adhere to the GPL. The fact that you distribute all but libXXX,
ask your users to download it separately from somewhere else and then
let them link the two pieces together, does not absolve you from your
responsibilities under the GPL. Essentially, by distributing your
binary not under the GPL, you violated the licence under which libXXX
was licensed to you. So you have no right to use it; hence you have no
legal way of producing the thing that you are distributing. I guess
the charge would be that you are distributing something that violates
the copyright of the copyright holder of libXXX; and that holder could
go after you under copyright law (on which the GPL is based, IIRC).
Given the naming scheme that GSL is using, I doubt that there is an
alternative implementation of libgsl (or gsl.dll); thus there might be
a violation here. But, as I said, IANAL.
> I can't find any rage from the FSF about what Numerit are doing, so I
> assume they're not considering it a violation.
First, the FSF might not be aware of this. Someone who cares should
perhaps notify them.
Secondly, Numerit's website seem to contain only e-mail contacts but no
snail-mail address. It is not clear to me where they are located and I
guess this information would be needed in case that it comes to a
suit; where would it be filed? But, perhaps, a lawyer would find out
easily where they are located.
Thirdly, even if the FSF knows about this and are aware where Numerit
is located, they may not be able to do anything. As far as I can tell,
the copyright of the GSL is held by the GSL team, not the FSF. My
understanding is that the GPL is based on copyright law and violations
of the GPL are handled under those laws. Presumably that means that
only the copyright holder can take action.
Perhaps the easiest solution would be to transfer the copyright of GSL
to SCO. They could then write to all the users of Numerit (as per
their testimonials page) that they are using a software that violates
SCO's copyright and ask them to cease and desist under threat of
compensation law suits. That should fix the problem, except that
we wouldn't like it if SCP were given a life-line, would we? ;-)
Finally, I am aware that the GPL allows you to sell the software for a
price that allows you to recuperate your costs; and of course you can
sell services. From that point of view, I find the the second option on
Numerit's web-site on how to obtain the GSL dll curious:
Another option is to order a CD from Network Theory Ltd (for $495).
The CD includes a precompiled version of GSL (DLL), complete source
code, and online documentation. Cheers,
How can Network Theory justify $495 as their cost for putting all these
together? But, apparently, if you go to Network Theory's side, they do
not distribute precompiled CDROMs anymore. I wonder whether somebody
has had a word with them about the price they were asking?
=========================== Full address =============================
Berwin A Turlach Tel.: +65 6515 4416 (secr)
Dept of Statistics and Applied Probability +65 6515 6650 (self)
Faculty of Science FAX : +65 6872 3919
National University of Singapore
6 Science Drive 2, Blk S16, Level 7 e-mail: statba at nus.edu.sg
Singapore 117546 http://www.stat.nus.edu.sg/~statba
More information about the R-help