Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 8/23/84; site ucbcad.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!harpo!decvax!ucbvax!ucbcad!moore
From: moore@ucbcad.UUCP
Newsgroups: net.sf-lovers
Subject: Re: Qwib-qwib: A review of Bezerker Base
Message-ID: <150@ucbcad.UUCP>
Date: Wed, 20-Mar-85 03:33:49 EST
Article-I.D.: ucbcad.150
Posted: Wed Mar 20 03:33:49 1985
Date-Received: Thu, 21-Mar-85 03:44:00 EST
References: <72@milford.UUCP>
Organization: UC Berkeley CAD Group, Berkeley, CA
Lines: 21

> In the stories by Zalasny(sp?), Anderson, and Niven there is a
> different attitude toward computers (still many technical errors,
> the notion of berzerkers as being deadly because they're 'von
> Neumann machines' stands out)

    Just a quick note in defense.  There are two different concepts
under the label "von Neumann machine".  One is the computer
architecture concept of a single processor and separate memory (ouch,
an awful definition).  But it is another idea of Dr. von Neumann that
is being referred to above, an idea most recently referred to in 2010.
Basically if you wish to a job of arbitrary size, such as transform all
of Jupiter, design a self-replicating machine that can also do the
job.  This way you personally build one machine, and then sit back and
let the wonders of exponential growth increase the numbers of machines
until you have enough to complete the job at hand.  Of course you need
raw materials, but for terra-forming or planet-wrecking the raw
materials are part of the task.

    Peter Moore
    moore@Berkeley
    ucbvax!moore