Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!scramble.lm.com!news.math.psu.edu!news.cse.psu.edu!uwm.edu!newsfeed.internetmci.com!in1.uu.net!news.thepoint.net!not-for-mail
From: myrddin@iosys.net (Myrddin Emrys)
Subject: Re: Reproduction
Message-ID: <319c035f.80793249@tisXnews.thepoint.net>
Date: Fri, 17 May 1996 05:03:30 GMT
References: <4lit47$ekt@news.magi.com> <4lln6h$t6a@soleil.uvsq.fr> <Dr1FL3.16A.0.staffin.dcs.ed.ac.uk@dcs.ed.ac.uk> <phvii6sd5y.fsf@destiny.mcs.anl.gov> <4n2pbo$2l4@news.euro.net> <ph91eyzzpz.fsf@destiny.mcs.anl.gov>
Organization: SimBusiness
Reply-To: myrddin@iosys.net
X-Newsreader: Forte Agent .99e/32.227
Lines: 36

We intercepted this transmission from pusch@mcs.anl.gov (Gordon D. Pusch):

:In article <4n2pbo$2l4@news.euro.net> Pierre Baume <pbaume@euronet.nl>
:writes:
:
:> I guess you are thinking about making an ancestor that will generate
:> random children... Is that possible within the rules of Tierra ?
:
:Well, actually it *wasn't* what I was thinking of --- but it certainly
:=is= an interesting idea...
:
:It would be very awkward in Ray's original instruction set; while it
:can be shown to be computationally universal (i.e., it can be used to
:simulate a Turing machine), it lacks instructions to move data from a
:register to the soup, or vice versa --- it can only copy data from one
:location in the soup to another. Hence, while one could easily write a
:random-number generator in tierran, there is no simple way of writing
:a random =cell=. However, a dodge that would accomplish =almost= the
:same thing, would be to move data from a random address to the next
:address in the ``daughter cell''...

 <Rest of posting snipped>

While Tierra cannot 'create' an arbitrary cell, it can (as you said, and I take
on faith :) be a random number generator. Therefore, an 'array' of 32 opcodes,
representing the 32 possible opcodes, could be used as the source for the random
cells. I am unfamiliar with Tierra's opcodes, so I would like confirmation or
refutation of this method's viability.

I get the impression that arbitrary numbers to reference specific cells are
non-existant. Therefore, it is quite likely harder than I envision (ie, generate
random number, then copy opcode from position x to next code in new random cell)
but I doubt it is impossible, given that Tierra is capable of being a universal
computer.
--
Myrddin Emrys                                 mailto:myrddin@iosys.net

