Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!agate!bionet!apple!rutgers!deimos!uxc!uxc.cso.uiuc.edu!mcdurb!aglew
From: aglew@mcdurb.Urbana.Gould.COM
Newsgroups: comp.arch
Subject: Re: Seeing the future
Message-ID: <28200243@mcdurb>
Date: 25 Nov 88 21:10:00 GMT
References: <1984@eos.UUCP>
Lines: 22
Nf-ID: #R:eos.UUCP:1984:mcdurb:28200243:000:1159
Nf-From: mcdurb.Urbana.Gould.COM!aglew    Nov 25 15:10:00 1988


>I think I have seen the future and it's not in supercomputing (or 
>"mainframes").  Oh yes there will be better machines, but I see too
>little flexibility, too little short sightedness in all people.  Difficulties
>in parallelism are grossly underestimated in some areas.
>
>Another gross generalization from
>
>--eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov

This is the best news I've heard of in a long time.
Supercomputers and super-mainframes are one of the gravest threats
to democracy in a long time. Democracy via personal-computing!

Although I'm not so sure that paralellism is as difficult as Eugene
says it is. Now that parallelism is available to the public, we'll
see how much quicker the public can be at developing new ideas,
compared to the boys in the national labs who had parallelism as
their own domain for so long.  This, plus the development of new
algorithms, like Greenard's O(n) algorithm for the n-body problem.
Greenard's algorithm in itself may spell the end of super-computing
-- or a new florescence. Have you ever considered just how many
supercomputing applications are versions of the n-body problem.