axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] Re: [Maxima] Re: GCL used commercially?


From: Richard Fateman
Subject: [Axiom-developer] Re: [Maxima] Re: GCL used commercially?
Date: Wed, 23 Jul 2003 09:22:42 -0700
User-agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02



Richard Stallman wrote:
    The commercial version, which was enhanced by Symbolics and then
    "Macsyma Inc" for about 20 years, is not public.  The Macsyma INc
    people supported and enhanced AKCL.  I do not know if their
    changes to AKCL were made public, but I suspect not.

Can anyone find out?

What was the license of AKCL at the time?  Did it permit them
to distribute an improved version and not release the source?
The LGPL does not permit such a thing.

_______________________________________________
RMS and others:

1. My recollection is that the Kyoto people wanted
their code separated, so that to run Austin-Kyoto Common Lisp
you had to set up the Kyoto code and then run Bill Schelter's
 "change script" to make it into Austin-Kyoto.
Officially you also needed to write to the authors
to tell them you were using it. And various other things.
I do not believe the Kyoto people made a fuss about re-use,
though I don't know if IBM or Ibuki or Austin Code Works
(see below) made specific deals.

2. The KCL licensing can be viewed in
http://www-2.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/lisp/impl/kcl/kcl/license.txt

among other things it says
.........
(c) Copyright Taiichi Yuasa and Masami Hagiya, 1984.  All rights reserved.
Copying of this file is authorized to users who have executed the true and proper "License Agreement for Kyoto Common LISP" with SIGLISP.

.........
I see no evidence that these authors ever surrendered their
copyrights, though in GCL documentation, also at CMU it says


   Versions 1.0 and above of GCL (aka versions 1-625 and above of
   AKCL) no longer require the kcl.tar file, and are covered by the
   GNU General Public Library License.

It could be that that declaration happened without the true original
authors' permission
and that someone (misunderstanding the nature of GPL?) thought that
(say, because he was willing to give up rights to HIS contributions to
AKCL) that ALL of AKCL, reborn as GCL, would be covered by GPL.
This could have been Bill Schelter, who also asked DOE to release
their 1982 source of Macsyma under GPL.  (Rather than, say, a BSD-like
license). In fact the DOE license is not GPL, if you look at it,
since one cannot send copies to Cuba or such countries. I don't think Bill was much interested in legal issues,
but he was generous with his code, intending it to be given
to anyone interested in using it. He made not have understood that
a GPL would inhibit some people from using it, RMS notwithstanding.
It is not possible to ask Bill Schelter, but the authors are, so far
as I know, still alive.

3. At least two companies (Austin Code Works, Ibuki) sold improved versions of KCL, or tried to. These are mentioned in the
various license and info files at cmu,... see the text file
http://www-2.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/lisp/impl/gcl/0.doc
for example.

As for stuff built on top of AKCL....

  a. I think that some version of the Axiom computer algebra
system required KCL, and that Bill Schelter consulted for
IBM to make it work better for that purpose.

  b. As previously mentioned, commercial Macsyma used a version
of KCL.

RMS' question seems directed to find out
if these various people (macsyma inc, ...) might be
in violation of some version of GPL.  My point is quite
different.

To reiterate:
I think that these 4 companies relied on GCL or AKCL or some predecessor
NOT being under GPL. They had no wish to distribute to their
competitors any improvements they developed.


  For good reasons or not, if AKCL were under GPL those companies
probably would not have used it, enhanced it, productized
it, or (especially) sold it.

 They wanted a piece of code that they could use without
paying, and that they could improve at their own expense, and
then use for their (proprietary) purposes and sell.

Whether they could have been convinced to forego their
capitalist impulses and celebrate free software is,
I think, not possible to answer at this time.

If I understand the objections of some people of the
maxima group to GPL, rather than LGLP, it is that it
would inhibit some future company that might pick up and
enhance maxima at its expense, and for its own profit, and
that it would start with the
initial work of this group.

 For a small enough
specialized and highly technical market, such a
company MIGHT make sense.  This is a different situation
from a very common, widely distributed, technically
"shallower" piece of software where 100,000 people
might plausibly contribute useful additions and corrections.

One solution for such a company would be to not use GCL;
using a commercial lisp might seem to increase its costs,
but note that keeping a full-time GCL expert alive and
well in a company might cost $100,000 or more
 per year (salary, support,benefits, overhead). Using
a "non-free" lisp might be cheaper than a free one. So
I am not personally so worried about using GCL for
maxima, so long as it also runs on other lisps.

Regards,

RJF





reply via email to

[Prev in Thread] Current Thread [Next in Thread]