Re: Mozilla/GNOME interoperability

[Michael Robinson <robinson netrinsics com>]

Mark Galassi <rosalia@cygnus.com> writes:
>    Michael>  1. It does not violate the GPL to link GNOME apps
>    Michael> against the non-free C libraries on commercial Unix
>    Michael> distributions.
>
>These are basic OS functions; there's a clause about that,

No there isn't a clause about that.  There is a clause that releases "major
components (compiler, kernel, and so on) of the operating system on which
the executable runs" from the source code availability requirements for
binary distributions.  Nothing about linking.

>makes good sense (both historically and now) to be able to use free
>s/w on a proprietary OS.

We're not talking about "makes good sense"; we're talking about the
specific wording of the GPL.

>The X libraries are not GPL, but they use a compatible
>university-style free license.
>Python libraries are compatible with the GPL.
>NPL is incompatible with the GPL, and these libraries are not a basic
>OS function.

There is nothing in the GPL about "compatible" and "incompatible" licenses.  
What the GPL says is this:

"b) You must cause any work that you distribute or publish, that in whole or
    in part contains or is derived from the Program or any part thereof, to
    be licensed as a whole at no charge to all third parties under the terms
    of this License."

And:

"But when you distribute the same sections as part of a whole which is a
 work based on the Program, the distribution of the whole must be on the
 terms of this License, whose permissions for other licensees extend to the
 entire whole, and thus to each and every part regardless of who wrote it."

Three points:

 1.  These sections of the GPL apply exclusively to the distribution of
     derived works.  In fact, the GPL specifically states: "Activities other
     than copying, distribution and modification are not covered by this
     License; they are outside its scope."  Nothing about linking.

 2.  A strict interpretation of the GPL prohibits the distribution of a 
     derived work that combines GPL and BSD-style (so-called "compatible")
     code, as the BSD-style license grants "use, copy, modify and distribute"
     rights.  Nothing about relicensing, sublicensing, or colicensing.
     Those are rights retained by the copyright holder, unless explicitly
     granted.  Under the terms of the BSD license, all modifications
     explicitly fall under the terms of the BSD license.

 3.  The basis of Richard Stallman's "incompatibility" claim seems to be
     that modifications to NPL-covered code explicitly cannot be copylefted.
     However, modifications to BSD-style code implicitly cannot be
     copylefted.  In order for the "compatible/incompatible" license
     distinction to have any validity, one must assume non-existent
     relicensing rights for code covered by BSD-style licenses.

All that notwithstanding, there still is nothing in the GPL which prohibits 
linking of GPL code with any other kind of code.  The only restrictions are
on how such a linked executable may be distributed.

>    Michael> If my understanding is now correct, I have another
>    Michael> question.  Are there any reasons for this besides:
>    Michael> "because Richard Stallman says so"?
>
>I hope I've explained the details, but since RMS crafted the GPL, I
>guess ultimately it *is* because he says so :-)

My point is basically this:  Is the GPL an actual legal document, is it
some sort of "oral tradition", or is it just whatever happens to suit
Richard Stallman's personal agenda at any given point in time?

	-Michael Robinson