Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!hou3c!hocda!houxm!houxz!vax135!cornell!uw-beaver!tektronix!hplabs!sri-unix!decvax!mulga!munnari!Lee@Ucb-Vax
From: Lee%Ucb-Vax@munnari.UUCP
Newsgroups: net.lang.prolog
Subject: BF search of SLD tree: Solution
Message-ID: <742@sri-arpa.UUCP>
Date: Wed, 6-Jun-84 22:12:38 EDT
Article-I.D.: sri-arpa.742
Posted: Wed Jun  6 22:12:38 1984
Date-Received: Wed, 13-Jun-84 01:46:33 EDT
Lines: 20

I didn't pay much attention to the recent discussion about
breadth first traversals and that news has been removed
from our system.  However, I believe the original question
concerned the BF traversal of the SLD (proof) tree for a
given goal.  The advantage of BF is that it will find all
solutions eventually (depth first can get lost down an
infinite branch).  I have just been working on an algorithm
which needs a fair search strategy for SLD trees so, as a
first step, I wrote the following program.  It has been
briefly tested on MU-Prolog and C-Prolog.


% Breadth first search of the SLD tree of a goal.
% Returns all solutions.
% The main data structure is a list of 'buds' (starts of
% branches which have not been explored).  Each one has

***Sender closed connection***

=== Network Mail from host su-score.arpa on Mon Jun 11 02:40:55  ===