Path: utzoo!mnetor!uunet!husc6!bloom-beacon!mit-eddie!uw-beaver!ssc-vax!dickey
From: dickey@ssc-vax.UUCP (Frederick J Dickey)
Newsgroups: comp.edu
Subject: Re: conceptual basis of CS
Message-ID: <1899@ssc-vax.UUCP>
Date: 4 May 88 13:44:32 GMT
References: <4823@june.cs.washington.edu>
Organization: Boeing Aerospace Corp., Seattle WA
Lines: 35
Keywords: fundamental concepts
Summary: programs = algorithms + data structures

In article <4823@june.cs.washington.edu>, jsnyder@june.cs.washington.edu (John Snyder) writes:
> I recently set myself the task of making a list of 5-10 fundamental
> concepts of computer science.  The sort of idea I was after had to
> (1) be such that current CS theory/practice would be inconceivable
> without it and (2) it crops up (maybe in various disguises) in all
> or many areas of computer science.  
> 
> I would really like to see your list, perhaps with a brief comment

It strikes me that a very, very fundamental concept of CS is the notion
that computers perform manipulations on data. This seems to manifest
itself in various guises. For example, Wirth's 
              Programs = Algorithms + Data structures,
(hope I got it right).

Essentially the same view is taken in AI
          Expert system = Inference engine + Knowledge representation,
except that different terminology is used for basically the same thing.

Digital computer architecture has the notion
          Instruction = Op code + Addresses of data.

One of the reasons I claim that this is a fundamental concept of CS is that
it is not clear to me that it is the only possible view of a computer,
and thus it is a distinguishing characteristic of CS.
For example, a computer might be viewed as a device that transforms
inputs into outputs. From this viewpoint, it is not clear that you
need algorithms and data structures. To elaborate further, consider
a "learning" neural net. To program this device one does not design
algorithms and data structures, but presents it with a series of
input/output relationships. The neural net may have something 
corresponding to algorithms and data structures internally, but you're 
not concerned with these. It seems to be the case that a neural engineer
is not nearly as concerned with algorithms and data structures as a
software engineer.