🚀 go-pugleaf

In article <896.Jul2219.16.3491@kramden.acf.nyu.edu>, brnstnd@kramden.acf.nyu.edu (Dan Bernstein) writes:
>                                                      Programming is an
> art, not a science, . . . 

It can be a science sometimes, but I wouldn't want *that* code :-)

> I recently asked in comp.software-eng how you would test competence in
> software engineering. Would you test knowledge of all the latest
> terminology for decades-old concepts? A bunch of people agreed that a
> meaningful test would be very difficult.

"Story Time!" (cheers from young children) :

I worked with one person who used a 25-question multiple choice quiz
as part of his interview process.  The questions were relevent to the
positions available, and the technology level was reasonable.  The
problem was that for some of the questions there was a "normal" answer
and a "correct" answer.  For yuks, I took the quiz one day.  I
answered all the questions "correctly" but got six wrong for not being
the "normal" correct answer.  When I explained why my answers were
correct, he was quite embarassed.

[Example: You set the computer's time in:
  (a) config.sys
  (b) autoexec.bat
  (c) command.com
The "normal" answer is (b), but I knew a guy who had a device driver
in config.sys that talked to his modem - and set the time from it.]

[Example: Which file is not required to boot MS-DOS?
  (a) io.sys
  (b) msdos.sys
  (c) command.com
  (d) autoexec.bat
The "normal" answer is (d), but we all know that (c) isn't required either.]

>                                            The only positive response I
> received was a proposal that programmers be tested on actual programming
> problems.

"Story Time Again!" (more cheers from young children) :

We interviewed a person here who listed Motif knowledge on his resume.
They asked me to find out if he really did know Motif.  When he came
in, I say him in front of my workstation and said "write me a tic-tac-toe
program using Motif."  Everyone laughed when they heard what I had
done, but I *know* that guy knows Motif now.

> When most universities have a Department of Programming or at least a
> recognized programming specialization inside the engineering department,
> we might have some basis for licensing programmers. Until then, it is
> simply irresponsible to even think that we can say what a programmer
> should know. We can't.

"Story Time Again!" (snores from sleeping young children) :

One guy I went to school with scored 3.6 most semesters, but after 2
years still couldn't figure out the word processor on *his* computer.

DJ
dj@ctron.com