Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!news.sei.cmu.edu!cis.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!swrinde!sgiblab!darwin.sura.net!ms!physics.physics.wm.edu!lasota
From: lasota@physics.physics.wm.edu (Chris LaSota)
Subject: Replicating systems from scratch....
Message-ID: <1994Oct3.172855.22766@cs.wm.edu>
Sender: news@cs.wm.edu (News System)
Nntp-Posting-Host: physics.wm.edu
Organization: /usr/local/lib/rn/orginisation
Date: Mon, 3 Oct 1994 17:28:55 GMT
Lines: 34


I've written a rather large C program whose purpose is to produce replicating
systems that have characteristics somewhere between L-systems and chemistry.
I'm writing to the Net to see if anyone else has heard of this kind of work
or has done it themselves.  It could save me a lot of time (not that you may
care).  Here's the description:

I take alphabetical characters to be the fundamental components of each
system.  There is a population of each type of letter, and two letters are
chosen at random proportional to their population.  They may be the same
letter.  One letter is called the TEXT, the other is called the EDITOR.
The TEXT is written down vertically over the EDITOR, and both are left-
justified.  Then a rule table is consulted to see what action must occur
between these LETs (as I call them).  There are rules such as swapping
position, inserting LETs which are available from the population, and
deleting LETs back into the population.  The rules apply only to a vertical
LET pair.

By using these types of rules, one can get STRINGs formed from sequences
of LETs.  These STRINGs go back into the population as separate entities
which are available for random selection during another "reaction".  During
the "reaction", the LET pairs react from left to right, and therefore some
rules which insert or delete LETs will change the future LET pairs which
are still to the right, but have not had a chance to react yet.  This
introduces a large amount of complexity into the system, yet it comes from
only a handful of simple rules.

I'm hoping to be able to generate large replicating STRINGs from a large
initial population of LETs.

Lemme know what you think.  Thanx.

Chris LaSota

