Fw: Query about GNU-GPL

Laurence Finston lfinsto1 at gwdg.de
Wed Mar 23 12:52:38 UTC 2005


On Wed, 23 Mar 2005, Alfred M. Szmidt wrote:

> Not sure if I understand you correctly, but if "the Program" uses the
> library, then the library is part of "the Program", and as such the
> whole thing must fall under the GNU GPL[0], since "You must cause any
> work that you distribute or publish, [...], to be licensed as a whole
> at no charge to all third aprties under the terms of this License".

I don't think so.  For example, my package uses GNU Bison and
I include a Bison input file in my sources.  If I choose, I can
distribute Bison with my package, since its license permits me to
do so.  It doesn't matter that both my package and Bison are licensed
under the GPL;  I could do the same with a package with a different
license, if the license permitted me to redistribute the package.
This doesn't make Bison part of my package or my package part of Bison.

Another example would be if I were to use the libplot library supplied
as part of the GNU plotutils package.  Here, I would include its header
files in my sources and link to the library when compiling.  In this
case, the packages are mixed more intimately, but the comments above
still apply.

>
> [0]: I'm ignoring the case when the non-free library has a similar
> clause that states that the programs linking to the library must be
> under the terms of the non-free license.

I am pretty certain that merely linking to a library does not give the
owners of that library any rights over my code.  I am also pretty
certain that linking two packages, in the literal sense of using
a linker to generate a file of object code, does not unite the
packages in a legal sense.  Even so, I will certainly not make
any free software I write dependent on a library with a license that
makes similar claims, whether they are ultimately enforceable or not.

If anyone wants an official answer to this or similar
difficult questions, I suggest he or she contact
`licensing at fsf.org'.

Laurence Finston



More information about the Discussion mailing list