Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!cs.utexas.edu!rice!fontenot
From: fontenot@rice.edu (Dwayne Jacques Fontenot)
Newsgroups: comp.theory.cell-automata
Subject: source for Gosper's Hashlife
Message-ID: <1991Jan28.203508.27886@rice.edu>
Date: 28 Jan 91 20:35:08 GMT
Sender: news@rice.edu (News)
Reply-To: fontenot@comet.rice.edu (Dwayne Jacques Fontenot)
Organization: Rice University, Houston, Texas
Lines: 29

Hello,

Does anyone know where I can get some source or a pseudo-code algorithm
for Bill Gosper's "Hashlife"?

The information I have says that the hashlife algorithm was designed in 1982
and that it uses self-referential hash tables to store lookup patterns for
John Conway's game of life. The algorithm is somewhat adaptive in that it
"learns" which patterns to store and works up from relatively small patterns
(4 x 4) to arbitrarily large ones.

As you might deduce from the above, the algorithm starts out slowly and then
accelerates logarithmically. I deduce that it uses memory in the same way ;-)

The algorithm has the intriguing property of storing, in the hash table, all
the information to recreate the life universe at any time in its history.
Gosper was described as running backwards and forwards in time to see how
interesting forms were produced.

The description I saw (in the book, Mind Children) was not complete enough
(in my opinion, anyway) for me to implement the alogorithm.
I would like to implement this algorithm, but I believe I need a better
description than was given in the book. Source code of any kind would be
even better...

Thank you for your time,

Dwayne Fontenot
fontenot@uncle-bens.rice.edu