[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: The Guile license and the use of LGPL libs (like GMP).
From: |
Greg Troxel |
Subject: |
Re: The Guile license and the use of LGPL libs (like GMP). |
Date: |
04 Jun 2002 07:22:25 -0400 |
What would prohibit you from using Guile if it would be licensed under
the Lesser GPL? (I guess it might be section 6 of the LGPL that
requires you to enable all recipients of your "work that uses the
library" to replace libguile with a suitably modified version of
libguile.)
That would likely be workable. That would leave two issues:
a question of people's comfort that they really understand the terms
- with the 'guile exception' it is very easy to be sure that one
understands it correctly. The LGPL doesn't require much, but it's
not so obvious. It additionally requires that reverse engineering
be permitted. Right now, there is resistance to using guile due to
scheme. So, for the sake of widespread guile adoption (with the
goal of guile becoming mainstream), I think we should keep all
barriers as low as possible.
Static linking -- What if cisco wanted to put guile in IOS? The
current strategy of record says that this (rather than, say, tcl :-)
would be a good thing. Overall, the FSF strategy seems to be
library that can't be had elsewhere -> GPL
library that has lots of non-free competition -> LGPL
library that we really want to have used and be mainstream (eg
guile, and afaik the only one of these) -> Guile exception
All that said, a --without-gmp that basically makes integers turn to
inexact on overflow (using double) rather than into exact bignums
should be a good compromise. I don't know if Rob is up for this,
though, or if his work-in-progress already supports it.
Probably someone from the guile maintainer group should talk to RMS or
any other FSF policy folks - this issue really goes a bit beyond just
the guile project.
Sorry if I'm being redundant...