Re: Hungarian notation

[engebret@sg2.cr.usgs.gov (Chris Engebretson)]

In article <4pk29i$nhj@tpd.dsccc.com>, kcline@sun132.spd.dsccc.com (Kevin Cline) writes:

|> Actually, I think Mr. Silverman knows exactly what student code looks like:
|> It looks a lot like professional code.  Or should I say that a lot of
|> professionally written code looks like 2nd-year student code?
|> 
|> Maybe we should have a contest to find the worst examples of professionally
|> produced C++ code?  

Apologies in advance if I'm going way off-base here. :-)

In my place of work, at least, the worst examples of professionally produced
code (independent of the language that it's written in) usually don't come 
from coders with degrees in Computer Science.  Rather, they come from people
like remote sensing scientists and electrical engineers.  These individuals
have the technical know-how to develop algorithms and justify their correct-
ness, but they tend to have difficulty expressing them eloquently and
efficiently when it comes down to actually writing the code.  Comments are
sparse if not totally nonexistant, variable names are not descriptive, and
"goto"s are a dime-a-dozen.

As I often have occasion to maintain code like this, I gain a deeper
appreciation for some of my college programming-oriented classes, because
they serve a dual purpose.  Not only do they introduce you to and allow you
to get your hands dirty with various topics and theories (i.e., compiler
fundamentals and language translation), but by requiring you to write several
thousand lines of code to successfully complete the class, they also make it
crystal-clear how important such things as efficiency, structure, and
documentation really are.  Most of the scientists I know who play a dual role
as a programmer really haven't had this kind of exposure; in fact, with
regards to C/FORTRAN, a good number of them are self-taught.  As a result,
the code that they produce works, but not as well as it probably could (and
should!), and it's certainly not as clear as it should be.

It's been my experience and observation that our academic community is
beginning to require more computer science "techie" courses from its 
scientific and engineering students.  And this is a Good Thing, because a
good deal of the world's existing base of code was not written by individuals
with strong Computer Science backgrounds, and that's not going to change
anytime in the near future.  Since these scientists and engineers will
inevitably be asked to express themselves in the form of a high-level
computer language, this experience will prove invaluable, both to themselves
and to the individual(s) whose job it is to maintain that code.

PS: Of course, this is not to say that a scientist with little background in
Computer Science can't write truly excellent code, and I'm certainly not
saying that when a student graduates with a degree in Computer Science, he/she
automatically will write wonderful programs, because we all know _that_ isn't
true.  But the old saying holds true: Practice makes perfect.

--
/*-------------------------------------------------------------
Chris Engebretson // CSB - SED - Scientific Systems Development
United States Geological Survey   -   National Mapping Division
Mundt Federal Building, USGS EROS Data Center
Sioux Falls, SD 57198  http://edcwww.cr.usgs.gov/eros-home.html
email: engebret@edcserver1.cr.usgs.gov      phone: 605-594-6829
-------------------------------------------------------------*/