Path: utzoo!attcan!uunet!wuarchive!uwm.edu!ogicse!emory!hubcap!fpst
From: fpst@hubcap.clemson.edu (Steve Stevenson)
Newsgroups: comp.parallel
Subject: Re: Paper (and) systems
Message-ID: <8088@hubcap.clemson.edu>
Date: 21 Feb 90 13:33:09 GMT
Organization: Clemson University, Clemson, SC
Lines: 56
Approved: parallel@hubcap.clemson.edu


In article <8081@hubcap.clemson.edu> frazier@CS.UCLA.EDU (Greg Frazier) writes:
>
>One of the major problems in academia today (in my opinion)
>is the fad for implementing systems, be they hardware or
>software.  In many cases, the implementations are simply
>attempts to drum up money (which is often successful).  If
>a concept can be clearly demonstrated without implementation,
>then universities have no business implementing them - the students'
>time would be better spent doing original research rather than
>implementing algorithms whose validity has already been
>demonstrated.  While, indeed, one cannot *know* a parallel/distributed
>language, OS, or machine will "work" until one actually sees it
>running, if we don't even look at peoples' work until it is at
>that stage, the delay between conception and revelation of ideas
>will be HUGE!!!  I want to hear about those ideas sooner than that.
>

In many cases, a concepts useful ness cannot be demonstrated without 
implementing at least a prototype of said system. A researcher may have
missed some crucial sync-point or other unexpected external force. The point
is that while an implementation of a system (even a prototype) shouldn't be
done until the concept has been proven useful/germane/valid one can't say
the concept is *usable* without being sure that everything has been acounted
for. In most cases, the most resonable way to accomplish this is to write a
prototype and pound on it. IMHO, feasability studies are still in the domain
of the researcher. You may claim that that is really an engineer's job, but
(again) IMHO, an engineer uses *proven* technology to get the job done.

I think it's irresponsible to just sit around brainstorming about ideas
all the time. You loose track of the real world if you don't actually *do*
something tangable. Furthermore, who's going to fund you if you don't get
results. One of the advantages of funding research in the university is
that it's *cheap*.

Now, I agree with your complaint about the delay between conception and
revelation of ideas is too large. However, that is due to the inadequacy
of proto-typing tools, NOT "The System". Fortunately, lot's of work is
being done in that area as of late and I don't think it will be too long
before we have tools powerful enough to quickly implement prototypes of
new ideas with much less overhead (in fact, the tools will pop up much
sooner in universities).

	Charlie Fineman

------------------------------------------------------------------------------
"Don't worry, be a happy fatalist"

InterNET: fineman@pluto.arc.nasa.gov


-- 
Steve Stevenson                            fpst@hubcap.clemson.edu
(aka D. E. Stevenson),                     steve@hubcap.clemson.edu
Department of Computer Science,            comp.parallel
Clemson University, Clemson, SC 29634-1906 (803)656-5880.mabell