Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!umcp-cs!mangoe
From: mangoe@umcp-cs.UUCP (Charley Wingate)
Newsgroups: net.cse
Subject: Re: Value of Computer Science degree
Message-ID: <3367@umcp-cs.UUCP>
Date: Wed, 26-Feb-86 09:55:24 EST
Article-I.D.: umcp-cs.3367
Posted: Wed Feb 26 09:55:24 1986
Date-Received: Sat, 1-Mar-86 01:22:22 EST
References: <1404@ames.UUCP>
Organization: U of Maryland, Computer Science Dept., College Park, MD
Lines: 56

I'd like to amplify a couple of things Eugene Miya said, and also add some
comments of my own based upon my experience in the workplace.

Here at Maryland, due to the curious division/college structure, the CS
department is totally independent of either the Math department or the
Engineering College.  It's also an important factor that the engineering
department academic programs are very programatic, whereas those of Math and
CS are rather free-form.  This has tended to divorce the CS and engineering
camps completely.

While this set-up has a lot of advantages as far as the independence of the
departments are concerned, it in practice leads to some problems.  One is
that the engineering college (and indeed, many departments across campus)
has its own programming courses, which, unless a student takes a
double-major, are likely to be the only programming courses he takes.
Things may have changed somewhat, but my experience has been that these
courses have always been years behind the times.  When I was a "first-aid"
consultant, we always saw six year old handouts and use of obselete software
as a consistent pattern in these courses.  None of these engineers were
taught any languages besides Fortran and (very bad) assembly, and little or
nothing about data structures was ever taught.

The end result of this is to produce a lot of "programmers" who really don't
know what they are doing.  This was especially true when I worked for a
defense laboratory.  At the time, it was almost impossible for Civil Service
to hire CS people: they could not pay enough, and the job classifications
caused a lot of problems.  I was in only because with my second major in
Math, I could be classified as a mathematician.  This meant that almost all
the programmers were engineers.  And these people were not just writing
stress analysis programs either; they wrote realtime applications and huge
simulations programs.  The latter were generally monuments to ignorance,
because those who wrote them had no idea about data structures or algorithms.

When I moved on to a private employer, a similar situation reigned.  At the
time, this company was hiring its first real CS people, except for one or
two who had been around for some time.  They had written two very large
personnel programs, both of which had some good ideas, but which tended to
brute force and which had been patched to death.  I came on board during the
construction of a third, which, although it contained a number of good
ideas, was hamstrung by a bad requirement on the part of the customer (who
had been told that the decision was going to cause trouble) to use a
particlar DB package, and by the presence of too many people who didn't know
anything about data structures, structured programming, or algorithms.  As a
result, the project remains under a cloud.

It seems to me that what you want is someone with the spark and flare of a
"hacker", but who has computer science education, and who can work in a
coordinated fashion with others.  There are far too many hackers whose CS
education is minimal; there are many many more who can only go off and work
in a corner by themselves.  In this respect, it seems clear to me that
courses in software engineering ought to be part of CS curricula, and they
should include group projects.

Well, that's my opinion.

C. Wingate