Quoting Lior Kesos <[EMAIL PROTECTED]>:

> The way I see it 90% of the jobs people actualy do in CS are not so
> mathimatically/alogorithmic wise complex that a simple administrator
> that can code like me can't deal with.

Well, in my rich work experience, I've had to maintain a lot of other people's
code. Let me tell you one thing - code written by someone clueful about
algorithms and efficiency and data structures is *miles* away from code written
by the "I learned by studying the Java book" or "I took a course at Sivan" kind
of programmer.

I'm not sure about administrator's coding, not having looked at a lot of such
code except shell scripts.

But the clueless really write awful code. No idea of re-using code. No idea of
uniformity of concepts (write a read loop in one way, not in ten way along the
code). Lots of hard-coded constants (I had to clean up such code only a week
ago. Two wasted hours). Heck, some of them actually compare boolean variables to
"true" or to "false" in ifs and whiles... I'm not talking about any complex
datastructures or re-implementing the hash table. I'm talking about a *clue*.

Like that famous program from my military days, which was always a nightmare.
They set it to run, and came back two days later for the results. Then one of
the more clueful programmers (who didn't actually have a BSc then) took a look
at it and discovered it was doing a bubble sort on its database. He wrote it
with quicksort or something equally efficient. It changed to running two hours
instead of two days.

As I said in the example itself, you don't necessarily have to have a BSc to
have the necessary clue. However, when someone has a BSc, you can be sure that
he *was* exposed to the required concepts and actually marked for them. So it
gives you that much certainty.

The problem with the clueless is that their programs *work* to spec. It's just
that they are not maintainable, extensible, or even readable...

However I think that if someone is a kernel guru, he must be clueful by
definition. Especially if any of his work has been peer-reviewed.

Herouth

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to