RadioBanter - View Single Post

Owen Duffy · March 2nd 07, 07:47 AM posted to rec.radio.amateur.antenna

(Geoffrey S. Mendelson) wrote in
:

....
Experienced scientific and engineering programers know this and
compensate. "Teach yourself Visual whatever in a week" programmers
don't.

The biggest problem with visual whatever programmers is they tend to
be seduced by the flashy interface and ignore the substance under it.

As an experienced scientific / engineering software developer, I suggest
that much the same issues exist in a conventional 3GL language as in the
"Visual xxx" environment. Good software doesn't just happen, it takes
(aside of the relatively trivial task of writing the functional
elements), huge effort in dealing with exceptions, and testing, testing,
testing.

If you dig deep into the source of Fortran programs, you will often find
the heritage of small memory and slow processors in days gone by, where
implementers used COMMON for variables to reduce memory use, and used
COMMON for passing parameters to avoid the overhead of parameter passing
function calls. This practice was, IMHO, the greatest risk to accuracy in
such programs, and the area that demanded the greatest attention to
design and documentation. The public NEC2 code uses COMMON, and the last
bug that I demonstrated in a C port of the code was related to messed up
parameter passing in COMMON.

So, the old 3GL world wasn't that good anyway!

Owen