🚀 go-pugleaf

Received: by munnari (4.44)
	id AA28083; Mon, 2 Dec 85 09:49:14 EST
Date: Mon, 2 Dec 85 09:49:14 EST
From: gln@munnari.oz (Gary Nelson)
To: INFO-RIDGE@hopkins

Here is a description of MU-Prolog that was supplied to me by the authors
at the University of Melbourne.  Although I have not used it yet myself,
they seem to be making very productive use of it.  I am forwarding this
information merely as a service to the Ridge community.  Any questions
should be directed to:

		jas%munnari.oz@CSNET-RELAY

_________________________________________________________________

                             MU-Prolog Release 3.2



       Release 3.2 of the MU-Prolog system will soon be  available,  for
       educational  and  research purposes only.  It is written in C and
       is especially suitable for Unix environments, although it is also
       available  for  VMS  systems.   It  is  currently running under a
       number of different versions of Unix (System III, System  V,  BSD
       4.1,  BSD  4.2,  OSx,  UTX)  on  a  variety of architectures (Vax
       11/780, Vax 11/750, Perkin-Elmer  3240,  Gould  PN-3000,  Pyramid
       90x,  Sun,  other  MC68000  systems).   Note  that because of its
       memory requirements (around 0.5Mb) it will  not  run  on  systems
       without  virtual  memory  or  large  real  memories.   There is a
       handling charge of $A200.00 associated with the distribution.

                             Features of MU-Prolog

       (1)  All  UNIX  Prolog  and  most  C-Prolog  and  DEC-10   Prolog
            facilities  are  provided.   The  syntax  is compatible with
            DEC-10 Prolog (the de facto standard).

       (2)  There  are  extra  control   facilities   which   enable   a
            coroutining  style of execution. User-defined procedures may
            have wait  declarations  which  can  delay  calls  to  these
            procedures.  Calls  to  many  system  predicates may also be
            delayed.

       (3)  Three sound forms of negation are  provided:  not  (~),  not
            equals (~=) and if-then-else.  Calls to these predicates are
            delayed if they are insufficiently instantiated.

       (4)  A logic pre-processor (lpp) allows programmers to write in a
            style  closer to first-order logic, which is then translated
            into   standard   Prolog,   and   control   information   is
            automatically added.

                            New Features in MU-Prolog 3.2

       (5)  A dynamic loading facility for linking compiled object  code
            into a running MU-Prolog interpreter (Unix only).

       (6)  The database system has been extended to allow the  user  to
            incorporate  (via  dynamic  loading)  his/her  own  indexing
            schemes to access external  databases  from  MU-Prolog.  For
            systems  on which dynamic loading is not available, database
            access schemes must be compiled into MU-Prolog.

       (7)  Two external database indexing schemes are supplied with the
            system:  a  recursive  linear  hashing  scheme  to store and
            retrieve ground Prolog facts; a superimposed codeword  index
            scheme to store and retrieve general Prolog terms (including
            variables).  Both indexing schemes provide efficient partial
            match  retrieval, even on very large databases (greater than
            1 million facts).

       (8)  A simple clause indexing scheme can be  invoked  to  provide
            fast access to facts and rules in the internal database.


       To obtain a copy of the  MU-Prolog  system,  two  signed  licence
       agreements  and  a cheque for $A200.00, payable to The University
       of Melbourne, must be sent to:

           MU-Prolog Distribution
           Department of Computer Science
           University of Melbourne
           Parkville, 3052
           Australia

       Please  include  details   of   your   installation   (that   is,
       make/model/location  of  host  machines) and what mediums you can
       read (1600 bpi tape in Unix tar format is easiest for us, but VMS
       BACKUP format can also be arranged).

       The MU-Prolog system has been  developed  at  the  University  of
       Melbourne  primarily  by  Lee  Naish. The recursive linear-hashed
       database system was developed in conjunction with James Thom. The
       superimposed   codeword  database  system  was  designed  by  Dr.
       Kotagiri Ramamohanarao and implemented by him and John  Shepherd.
       The clause indexing scheme was developed by Jeff Schultz.

       MU-Prolog 3.2 will be  fully  available  around  October-November
       1985.

       Work is underway at the University  of  Melbourne  to  develop  a
       compiler   for   the   MU-Prolog  language  which  will  offer  a
       substantial performance improvement over  the  interpreter.   MU-
       Prolog  3.2 will be the last release of the MU-Prolog interpreter
       before the release of the compiler.