Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site randvax.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!sdcsvax!sdcrdcf!randvax!lseward
From: lseward@randvax.UUCP (Larry Seward)
Newsgroups: net.math.symbolic
Subject: $15K Macsyma
Message-ID: <2045@randvax.UUCP>
Date: Mon, 8-Oct-84 19:14:33 EDT
Article-I.D.: randvax.2045
Posted: Mon Oct 8 19:14:33 1984
Date-Received: Wed, 10-Oct-84 06:16:38 EDT
Organization: Rand Corp., Santa Monica
Lines: 170
Forwarded messages-------------------------------------
From: lanl-a!jlg@rand-unix
Newsgroups: net.math.symbolic
Subject: $15k for MACSYMA
Date: Thu, 27-Sep-84 10:47:42 PDT
Organization: Los Alamos National Laboratory
Apparently-To: symalg
There appear to me a few misconceptions and misstatements of fact in
James Giles' response, so I thought I should say something.
> Are there any public domain symbolic mathematics packages out there?
> Symbolics' price of $15K seems a bit steep for Macsyma. Are there any
> free packages or reasonably priced packages around?
$15K seems especially out of line when you remember that Symbolics didn't
even write the code! The VAX version was written and maintained at
Berkeley.
You seem to be under the impression that each version of MACSYMA involves
its own code in a large way. That simply is not so. There is essentially
only one version of the MACSYMA code, i.e. over 90% of it is the same no
matter what the version. And Berkeley did not provide sole maintenance of
the VAX version. MIT fixed a bunch of bugs in it that were forwarded to us
by Berkeley. Also, Symbolics did a good deal of work on the Vax version as
well.
MACSYMA was designed and developed at MIT (much of it was government funded
through DOE and others). Most of the coding for MACSYMA was done by
graduate students, DOE scientists, and faculty; few of whom benefit from
the sale of the code. In fact, the only version which Symbolics wrote was
the one which runs on their Lisp Machine.
The first version for the Lisp Machine was done at MIT. This effort is
documented in the article "The LISP Machine" which, for one, is in the book
"Artificial Intelligence: An MIT Perspective", Vol. 2, which was published
before the founding of Symbolics. You are correct that Symbolics did a good
deal of work on the version which runs on the 3600.
I don't blame Symbolics for this mess. I would probably sell MACSYMA for
$15k myself, if someone declared me to be the authorized vendor. This was
in litigation for some time and there was a strong case for declaring the
code to be public domain.
I don't recall it ever being "in litigation".
In fact, before Symbolics was licensed as
vendor, the VAX version of the code was available from Berkeley (for
government and educational institutions only) for a tape copy fee ($200 or
$500 I think). Berkeley favored the position that the code should be
public domain. Too bad they didn't prevail.
For some strange reason you are treating "Berkeley" as a monolithic thing.
A professor at Berkeley may have subscribed to the point of view you
mention, but the legal office at Berkeley is a contractee in the current
situation, and Berkeley accepts royalties from it.
Also, $15k is the price per machine - no site wide licenses are available
from Symbolics. However, substantial discounts are available for
educational institutions (still for individual machines only). Symbolics
charges an additional fee for a maintenance contract (one year).
MACSYMA and Symbolics are trademarks of Symbolics inc.
James Giles
jlg@lanl-a
...!inhp4!cmcl2!lanl-a!jlg
--------------------------------------------------
Date: Fri, 5 Oct 84 23:49:28 pdt
From: fateman%ucbdali@Berkeley (Richard Fateman)
Message-Id: <8410060649.AA23220@ucbdali.ARPA>
Subject: Re: Setting the record straight on MACSYMA
Cc: fateman%ucbdali@Berkeley
>From lseward@rand-unix Fri Oct 5 17:01:34 1984
Subject: Setting the record straight on MACSYMA
Date: Wed, 3 Oct 84 08:21 EDT
From: Richard Pavelle