PROLOG-REQUEST@SCORE.STANFORD.EDU.UUCP (02/08/87)
PROLOG Digest Monday, 9 Feb 1987 Volume 5 : Issue 7
Today's Topics:
Query - Debugging & Conference Proceedings,
Announcement - Dissertation available,
Implementation - Debugging
----------------------------------------------------------------------
Date: 28 Jan 87 05:01:56 GMT
From: cybvax0!frog!primerd!bob@EDDIE.MIT.EDU
Subject: Debugging
I'm doing some research on debugging techniques, and I'm interested in
the special kinds of bugs a Prolog programmer encounters. (There are
obviously some things about Prolog which make it unique in this
respect.)
I'd like to hear about problems people have had, and how they uncovered
them. Bug stories are always interesting. I'd also like to hear your
opinion about what sort of tools would make Prolog debugging easier.
-- Bob Pellegrino
Prime Computer
------------------------------
Date: 27 Jan 87 11:34:27 GMT
From: mcvax!kddlab!titcca!kossvax!koeavax!yang@seismo.css.gov
Subject: I will be pleased to send my dissertation (about P-Prolog)
on request (In English)
Hello, I'm a Ph.D student, majoring in logic programming, at Keio
University. I have just finished writing my doctoral dissertation
titled 'A Parallel Logic Programming Language and Its Implementation'.
You can find an abstract of my dissertation in the rest this message.
Does anybody have an interest in it? I'd like to send you a copy upon
your request. I'm sorry that I will basically send it by sea mail
because of a financial reason. If you'd like to read it as soon as
possible, please mention this in your reply and send me about $15.
Then, I will be able to send you in air mail.
I look forward to hearing from you !
Thanks you.
-- Rong Yang
------------------------------------------------------------------------------
My mail address: | My net mail address:
|
Rong Yang | yang%keio.junet@utokyo-relay
Aiso Lab |
Dept. of Electrical Engineering| or
Keio University |
3-14-1 Hiyoshi, Yokohama 223 | ..!mcvax!kddlab!titcca!kossvax!koeavax!yang
JAPAN |
------------------------------------------------------------------------------
Abstract
This dissertation presents a parallel logic programming language named
P-Prolog, and discusses its implementation.
To date, two main research areas for parallel logic programming have
been studied:
(1). To develop methods, or control strategies, for executing
pure Horn clause programs in parallel.
(2). To develop new parallel languages based on guarded Horn
clauses, and implementation methods for them.
For (1), it is likely that to execute pure Horn clause programs with
both and-parallelism and or-parallelism is quite difficult, so
restrictions are usually imposed. The drawback with (2) is that
guarded Horn clause languages sacrifice completeness, because they do
not incorporate don't-know non-determinism.
This dissertation presents an alternative proposal, P-Prolog, which
provides the advantages of guarded Horn clauses while retaining
don't-know non-determinism where required. P-Prolog programs are
composed of a kind of extended guarded Horn clauses called classified
Horn clauses} which are introduced in this dissertation. A novel
concept introduced in classified Horn clauses is the exclusive
relation of guarded Horn clauses. Two advantages resulting from the
introduction of classified Horn clauses are:
(1). The language combines and-parallelism and or-parallelism. Therefore
the execution of P-Prolog can incorporate both don't-care non-
determinism and don't-know non-determinism.
(2). The input/output pattern of predicates need not be fixed. Its
synchronization mechanism allows the direction of data flow to be
determined dynamically.
In terms of implementation, combining and- and or- parallelism is
regarded as a difficult subject. This dissertation presents an or-tree
model and an implementation scheme for it, to combine these kinds of
parallelism with reasonable efficiency. The model and implementation
scheme discussed in this dissertation can be applied not only to
P-Prolog, but also to other parallel logic languages.
This dissertation is divided into two parts: language and
implementation. In Part I, a brief review of the theory of logic
programming and a survey of parallel logic languages are first given;
then the concepts, syntax, semantics and characteristics of P-Prolog
are described. In Part II, the three main problems of implementation
are discussed: representation of data, management of multiple
environments and communication between and-processes and or-processes.
------------------------------
Date: 3 Feb 87 11:05:21 GMT
From: Bruce T. Smith <ihnp4!inuxc!iuvax!bsmith@ucbvax.Berkeley.EDU>
Subject: 2nd International Logic Programming Conference?
Does anyone know how to get the proceedings from the Second International
Conference (Uppsala, SWEDEN, in July, 1984)? I've seen the 1st (Marseille,
FRANCE, in September, 1982), and the 3rd is a Springer LNCS. The Second,
however...
Thank you.
-- Bruce T. Smith
------------------------------
Date: 2 Feb 87 16:15:18 GMT
From: Peter Y.F. Wu <rpics!wup@seismo.css.gov>
Subject: Debugging
Typo is a common mistake. When that happens with the functor name of
a predicate, it results in undefined predicates in the Prolog program.
It may be nice for a Prolog system to *optionally* flag undefined
predicates so that one does not have to trace it all the way to
realize it is a typo in the functor name. It has to be optional since
undefined predicate may be a deliberate feature to generate failure.
-- Peter Y.F. Wu
------------------------------
End of PROLOG Digest
********************