Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ames.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!ihnp4!zehntel!dual!ames!eugene
From: eugene@ames.UUCP (Eugene Miya)
Newsgroups: net.arch,net.ai
Subject: Re: Alternate Computer Architectures (Naive Questions)
Message-ID: <770@ames.UUCP>
Date: Tue, 29-Jan-85 15:14:53 EST
Article-I.D.: ames.770
Posted: Tue Jan 29 15:14:53 1985
Date-Received: Sat, 2-Feb-85 00:36:01 EST
References: <596@marlin.UUCP>
Distribution: net
Organization: NASA-Ames Research Center, Mtn. View, CA
Lines: 63
Xref: watmath net.arch:774 net.ai:2496

> 
>   I have been reading some articles about alternate computer architectures
> and am curious about the following:
> 
>   *** Optical Computers ***
> 
>   1. What work has been done in actually building an optical computer??
>      (That is a computer whose logic states are determined by 
>       properties of light)

There is a company trying to build a hybrid electronic/optical [Guiltech].
See mention of other places to look in the next paragraph.

> 
>   2. What are the advantages of such a computer?? (My understanding is 
>      that processing speed is MUCH quicker than conventional architectures.
>      How much faster?? Why??) Also, the amount of memory in a "crystal" drive
>      is much greater than a normal disk drive (say, of the same volume).

The way it was explained to me by vax135!alan, is infinite parallelism.
The suggested reading material should be Proc. of the IEEE, July 1984,
survey on optical computing.  Photons have the advantage of not interacting
with one another (also a disadvantage since this is how electrons have an
advantage with transistors).  Alan Huang at Bell made me a convert on this.
You can have many beams (slide projectors) thru the same space.

>      
>   *** Molecular and Chemical Computers ***
> 
>   1. I read some eons ago about efforts to use organic molecules in a
>      supercooled liquid as a memory device. Whatever happened to that
>      line of research??
> 
>   2. Since human brain activity is electrical and chemical in nature,
>      what efforts has been made to simulate the METHODOLOGY of human
>      thought on a computer? For example, has any research been done
>      to find the electrical and chemical sequence that occurs when
>      someone is typing a line of text (like what I just did)??
> 
>                                                 Mike
> -- 
> 
> Real Life: Michael Bloomberg
>    MILNET: bloomber@nosc
>      UUCP: [ihnp4,akgua,decvax,dcdwest,ucbvax]!sdcsvax!noscvax!bloomber

The human brain is not necessarily a good model for computation.  Witness
robots that do not necessarily look like humans like Unimates.  Research is
being done in this area.  When I was an undergrad, I held a work-study
job in a psychology department and a summer position with a firm (Infomation
Magnetics) designing magnetic media.  In the latter, we discussed using
media such as DNA: problems: e.g., UV light, advantages: reproducibility,
low cost.  The biggest problem is speed.  This is all chemical at the
molecular level (no flames please).  It could be argued that "parallelism"
might make up for the speed, but I remain unconconvenced. As for "software"
talk to the AI people (in particular minksy, since he reads the ai list).
There are people working in this area, but it doesn't excite me like optical
computing.

--eugene miya
  NASA Ames Research Center
  {hplabs,ihnp4,dual,hao,vortex}!ames!aurora!eugene
  emiya@ames-vmsb.ARPA