skef@cmu-cs-spice.ARPA (Skef Wholey) (03/15/85)
Macrus Leech asks:
Does anyone know of a PUBLIC DOMAIN version of COMMON LISP written in C?
I'm interested in getting source for such a beast.
A Common Lisp written in C was done in Japan by some people at Kyoto University
for the DG Eclipse MV/8000. They have since ported it to the Vax and Sun
running 4.2 Unix. I do not know if their implementation is public domain, but
they seem interested in distributing it (the implementors are coming to CMU
early in April to show it off to us). It is supposedly slower than other
Common Lisps for the same machines (DEC's Common Lisp on the Vax and DG's
Common Lisp on the MV). I don't know how much slower.
We (the Spice Lisp Group) at CMU have done a public domain Common Lisp written
mostly in Common Lisp. Our code has served as the basis for most existing
Common Lisp implementations, including DEC's and DG's (Symbolics Common Lisp
and Kyoto Common Lisp are the notable exceptions).
Our system consists of a compiler, an extensible text editor (with a default
command set very much like ITS/TWENEX EMACS), and the Lisp system proper, which
has debugging facilities and so on. It is probably much harder to port to a
new machine than Kyoto's Common Lisp, although the better performance and
richer environment would probably make such a porting effort worth the extra
trouble.
Although Spice Lisp is in the public domain, CMU will charge something like
$100 for a tar tape of our sources and hardcopy of our internals documents to
cover handling. ARPANET access to the sources can be obtained for free. The
system weighs in at about 300K lines of Lisp code. Machine-specific support
for the runtime system will probably add something like 5K to 20K lines of
code, depending on the machine and level of tuning you're interested in doing.
Common Lisp is BIG. Doing a high quality implementation (working from our
sources) will probably take at least 1 man year of effort by someone who can
code fast and knows what he's doing. That's not too bad compared to other big
systems, but it's not something you can expect to do over a weekend, either.
If you're interested in getting access to our sources, the best person to
contact is Scott Fahlman, head of the Spice Lisp Group. He can be reached at
this address:
Dr. Scott E. Fahlman
Computer Science Department
Carnegie-Mellon University
Pittsburgh, PA 15213
When I find out more about Kyoto Common Lisp, I'll post.
--Skef
--
uucp: ...!seismo!cmu-cs-spice!skef
arpa: skef@CMU-CS-Spice