bts (11/29/82)
Greetings! This group has been created as a forum for
discussions of the Prolog programming language and logic
programming in general.
One of the first things I'd like to see is information
on which Prolog implementations people are using, what they
think of them, where they got them, etc. This seems to be
the most common Prolog request on the net, so let's share
what we know.
At UNC, we have the Edinburgh UNIX Prolog, written in
PDP-11 assembly language. It's running on one of our VAX
11/780s (4.1bsd), using the 'compat' compatibility mode
package. We found compat lurking in /usr/src/games, where
it was being used by Zork and a few of the older games.
There is almost no documentation on compat. Thanks to hints
from John Lloyd at the University of Melbourne, we were able
to fix compat to work with Prolog. (If anyone is interested,
I can mail the details.)
This UNIX Prolog is very similar to the Prolog
described in Clocksin and Mellish's book "Programming in
Prolog". That's not surprising, since they were the authors
of this interpreter, too. There are three problems:
(1) It is restricted by the PDP-11 address space, so compu-
tations frequently run out of memory. (It gives you a
nice "infinite loop" error message when this happens.)
(2) Every now and again, compat becomes confused by inter-
rupts and bombs.
(3) There is no compiler.
If you're interested in UNIX Prolog, the person to
write to is
Robert Rae, Chief Systems Programmer
Department of Artificial Intelligence
University of Edinburgh
Edinburgh, EH1 2QL
Scotland
U.K.
We'd like to have a Prolog which would give us a com-
piler and which would take advantage of the VAX's address
space. It would have to run under UNIX. One written in C
might be best, particularly if we were free to modify it.
I'm hoping that one of you will suggest a Prolog for us.
One final comment about the group. If you know Prolog
users who're not on USENET, please forward this to them.
I'm willing to do a small amount of re-mailing of articles
to people on ARPA or CSNET and to submit articles from them
to the group.
Thanks, again, for the yes votes.
Bruce Smith, UNC-Chapel Hill
duke!unc!bts
bts.unc@udel-relay