Re: The Problem with Programming

Yes. At a management session, Borcon 2004, the speaker advised us to
"hire mediocre developers for production work" (I don't remember the
exact words).
If speed is crucial, one could use a profiler to track down the
bottlenecks afterwards, otherwise he put emphasises on readability of
the code, first and foremost. He also presented examples which I before
used to regard silly programming, like:

You see, I have a problem with that. Once you get to the point where
the product is finished and you are ironing out the rough spots there's a
good chance management will say "It's done? Well ship it out then, we
can release an update in xxx months."

Also, a mediocre developer may not know how to actually get the
performance you need out of the application the way a more skilled
developer will. Mediocre developers also lead to longer development
times, wasted time, and mediocre software.

A profiler will tell you where the bottleneck is, but not how to improve
it. A competent developer will *know* where it is before the profiler is
run, anywys, given sufficient knowledge of the system. A mediocre
developer will run the profiler and then optimize the code in a mediocre
way, resulting in an ill-fated mediocre solution. If you like marketing
mediocre solutions, by all means, hire mediocre developers.

I doubt Borland, or any company that makes its money off of softare,
is out targetting mediocre developers for hire.

Actually, I may take that back <g>

And only people who have no experience in C-based development
languages would have any problem understanding the loop at the
end of yoru post. All this unnecessary verbosity is unneeded iMo.

Ciao.

- Nate.


.