From: utzoo!decvax!harpo!eagle!mhtsa!alice!npoiv!npois!houxm!houxa!houxk!houxj!wapd
Newsgroups: net.arch
Title: Re: Biochips... or bull chips?
Article-I.D.: houxj.205
Posted: Wed Jan 26 10:10:39 1983
Received: Fri Jan 28 21:04:12 1983
References: brunix.1295


	I have always thought that "biological computing" was a really
great idea, and I'm sure that it will be done some day, but I am
skeptical of claims that "it will be done in 5 years" and "it will
be 100 million times faster than existing computers".

	Seems to me that the most feasible development is a large,
SLOW memory.  I don't see where the incredible speed would come from,
and since biological systems use a lot of chemical reactions wouldn't
signal propagation be fairly slow ?  A memory would be the first product
because its structure is very regular and simple.

	The benefits of biology for memories are :  they can be incredibly
large (depends on how much nutrient you want to pour in) and they can
be self-replicating (self-repairing).  A small benefit is that you
can easily configure your memory to fit your needs (you can meet peak
loads).  That is, when you don't use the computer for a week, just have
enough nutrient dribbled in to keep a handful of cells alive.  When
a big load comes through, shovel in some food and watch it grow !

	The next product after a memory would probably be a huge array
processor, with all of the configurability and self-repairing
advantages that the memory exhibited.

	Fascinating area to speculate about.  Something may come of
it eventually.

						Bill Dietrich
						houxj!wapd