Re: IBM doesn't like the GPL

[Hyman Rosen]

Alexander Terekhov wrote:
> GPLv3 is clear as mud.
>
> Outside the Definitions section it purports to define an "is called"
> term "aggregate"

It seems clear enough to me. The distinction is between putting a
bunch of unrelated code together on one disk and linking libraries
together into a program. It is only called unclear by people who
want to deliberately blur the distinction in order to try to use
the aggregation permission to allow them to create a combined work
without having to license the whole work under the GPL. They would
get laughed out of court if they actually dared put this nonsense
in front of a judge.

Where the FSF does go too far is in seeking to ignore the distinction
between static linking and dynamic linking. They are incorrect in this,
and I suspect they know it but have decided to continue insisting they
are correct in order to discourage behavior which is actually allowed.

A program which is written to use a library is not a derivative work
of that library, because it is not a transformation of that library
representing a significant work of authorship. If that program is
linked statically with the library then an altered copy of the library
becomes embedded in the resulting executable, and so making and
distributing the program requires permission from the rights holder of
the library. That is how the GPL can reach out to embrace the entire
program. But if the program is dynamically linked, it does not contain
the library within itself and it is not a derivative work of the library,
so no permissions from the library's rights holder are needed, and the
GPL cannot reach out to embrace such programs.