Re: E-LISP licensing question

[Stephen J. Turnbull]

Glenn Morris writes:
 > Geoffrey Teale wrote:
 > 
 > > If I write some emacs lisp code does the way emacs deals with that code 
 > > at runtime mean that the code must always be under the GPL?

No.  Not if it simply uses facilities that are available in some form
in pretty much any Lisp.  If it touches an Emacs buffer or other
facilities that are available only in a GPL format linked into Emacs,
see below.

 > This GPL FAQ seems very relevant, if we consider Emacs as an
 > interpreter for the Emacs Lisp programming language:
 > 
 > http://www.gnu.org/licenses/gpl-faq.html#IfInterpreterIsGPL
 > 
 > I think it's clear that you don't need to license under the GPL
 > specifically. But you may need to use a GPL-compatible license.

The FAQ's answer is bizarre.  Under copyright law, the only time you
ever are legally bound by the GPL is when the program is a derivative
of a GPLed work.  But that requirement is *not* "GPL-compatible", it
is "the terms of the GPL."

IANAL, but that doesn't require a lot of legal knowledge to say.  I
don't trust that answer at all.

 >    If a programming language interpreter is released under the GPL,
 >    does that mean programs written to be interpreted by it must be
 >    under GPL-compatible licenses?
 > 
 > Trying to interpret the answer to that FAQ, it would seem that if you
 > just write some "pure" Emacs-lisp, you can use whatever license you
 > like.

Afraid not.  The Lisp engine is one thing, but nobody writes programs
that invoke only bytecode interpreter primitives.  Good luck writing a
useful Emacs Lisp program that invokes no functions defined in ./lisp.
So I think it's unquestionable that an Emacs Lisp program should be
considered guilty of linkage to GPLed code until proven innocent.

The only out would be if you could argue that this isn't intended to
be linked *specifically* to Emacs, which would kinda need the
possibility of running it on a non-GPL Lisp system.