🚀 go-pugleaf

I am the non-programmer who decided to write a chess program for the fun
of it.  I've reached an impasse and decided to post this question here,
rather than in rec.games.chess, since it's about programming and not
chess.

Anyway, my program is at the point where it recursively searches
depth-first to a specified depth and prunes the search tree of
unnecessary searches.  Before it starts a given search, the moves are
sorted in order of most significant capture.

So far, so good.  I was able to accomplish this much without outside
help and, in fact, when I finally got a copy of Levy and Newborn's _How
Comoputers Play Chess_ I was pleased to discover that my program
followed very closely the structure that they describe.  I am now at the
point where I want to implement (a) arbitrary deepening of the search
for non-quiescent positions and (b) "iterative deepening", so that one
level at a time is completed.

I actually have a primitive version of (a) running; the search will not
terminate on a position where a king is in check.  But the scoring of
these deepened searches doesn't seem to be right.  I can probably fix
this on my own.  My real problem is with (b).  Levy and Newborn say
something about searching at a given level and storing lots of positions
in a hash table before moving to the next level.  Is this the only way
to do it?  I don't know a thing about hash tables and, from what I have
read it looks like it's over my head.

So my question is, can iterative deepening be usefully (if not
optimally) accomplished without using hash tables?

Incidentally, when I get this miserable program running decently, I
will release it and the source code (sorry, Turbo Pascal) into the
public domain for anyone else who might find this sort of thing
amusing.

-- 
Todd Moody * tmoody@sju.edu  (That's St. Joseph's Univ., Philosophy Dept.)
            "In what furnace was thy brain?"  -- William Blake