Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!husc6!ut-sally!utah-cs!shebs
From: shebs@utah-cs.UUCP
Newsgroups: comp.edu
Subject: Re: Software Engineering vs. Computer Science
Message-ID: <4422@utah-cs.UUCP>
Date: Fri, 27-Mar-87 21:55:06 EST
Article-I.D.: utah-cs.4422
Posted: Fri Mar 27 21:55:06 1987
Date-Received: Sat, 28-Mar-87 17:04:51 EST
References: <193@darth.UUCP>
Reply-To: shebs@utah-cs.UUCP (Stanley Shebs)
Organization: PASS Research Group
Lines: 67

In article <193@darth.UUCP> gary@darth.UUCP (Gary Wisniewski) writes:

>There seems, in general, to be a lack of understanding of what Software
>Engineering actually is.  If you've written a game for an Atari, are you
>a Software Engineer?  If you've written an rule-based simulation for a
>Symbolics, are you a Software Engineer?

Well, what's the answer?  *I* sure don't know!  The typical definitions
for SE seem to focus more on the process than the result, so you can get
well-engineered games and hacked-out simulators...

>Are educators working to produce curricula which focus on CS and SE as
>distinct areas of study?  Very few of my co-workers have an understanding 
>of Software Engineering which was not gained from experience.  In fact,
>to become good engineers, many people must unlearn principles taught to
>them in CS courses.  Producing large, complex software systems is a different
>kind of problem than writing a 300 line program.  Data abstraction is an
>interesting twist in a small program, but essential in a large system.

This is exactly the kind of attitude I was decrying in an earlier article.
Students have to be able to build little things before they can build big
things.  This is true of any engineering field.  EE's build circuits with
one resistor and one capacitor, then one transistor, then one IC.  They
don't even think about VLSI or continental power networks until fairly late
in their student career, and will not generally *build* sizeable things
until grad school.  The same is true of other engineering disciplines.
Let's not start flagellating ourselves because students don't write
10K+ line programs as undergraduates!

The remark about "not understanding SE without experience" touches on an old
bone of contention between academia and industry.  Just last week I noticed a
letter in an IEEE rag, stating that academia did not prepare students
adequately for work, and proposing (among other things) that tenure be
abolished, and that university faculty should consist of one-half industry
people.  No mention was made of how those people were going to be coerced
into teaching...  I believe that it's *never* going to be the case that
universities will prepare students for industry to industry's specifications.
That's just not the way the world works.

>My point of all of this is focus.  Students who wish to unravel the secrets
>of computing (and probably the mind as well) are rarely satisfied with the
>engineering of software.  Those interested in practical solutions to real
>problems are rarely interested in the science of computing.  Many software
>people lack an understanding of this simple duality.

CS hasn't been around long enough to develop standards on how much theory
the engineering types should know, or how competent a theoretician ought
to be in the practice of programming.  Presumably you don't want ignorant
software engineers to spend a lot of time and effort trying to write a
"program to detect all infinite loops in other programs"!  If you divided up
theory and practice now, the theory would spin off into the ozone (a la
Soviet computer science), and the practice would drift into the reinvention
of exponential algorithms for sorting.  I personally prefer to be able to
do both theory and practice.  I suppose many people in CS departments like
that too, which is maybe why there hasn't been much impetus for splitting
departments (can you imagine physicists and EEs and MEs bundled into the
same department?  It would last one week, or until a seminar, whichever
came first...)

>Gary Wisniewski
>uucp: {allegra,bellcore,cadre}!pitt!darth!gary

							stan shebs
							shebs@cs.utah.edu

(I've been in both industry and academia and am now considered too theoretical
for industry and too practical for academia.  What a bind! :-) )