|
| I have a vague memory that IBM used to filter their cadets
with
|some form of test. Can any IBM alumni out there say more about this?
|
I never worked for IBM in the 60s, but I do remember taking their
programming test (perhaps I was applying for a summer job that never came
through; I don't remember). It was a standard IQ-like test, with lots
of analogies, simple math, find the 2 figures that are the same, etc.
I remember it being quite easy, but not related to anything I
considered programming. I also took a test to get hired as a
programmer by SDC (part of Rand at the time) as a programmer. That
included a vocabulary test, which I thought was pretty irrelevant.
I did well on both those tests, and I was a good programmer (since I
don't program much now, I can't prove that to you, but I'll just point
out that as an undergraduate I was the TA for my university's graduate
level assembly language programming course). However, I don't really
see much connection between what those tests tested and what I do/did
as a programmer. I think that was about the time that some studies
showed factors of 10 differences in performance among programmers, and
this was a hoped for 'perfect' selection mechanism. I've never been
able to find any research on how effective the tests were, but they
did fall out of disfavor pretty quickly.
Dennis Egan's work on individual differences used a word processor,
not a programming language, so I don't know that you can definitively
say that spatial skills matter for programming. I only know that I am
a spatial idiot (I couldn't mentally rotate a 3-D figure to save my
life), but it hasn't made it hard to learn programming.
Another anecdote. For a study I did years ago on expert/novice
differences in debugging, my novices were CMU students taking their
first programming course. They had some amazing misconceptions (for
example, one student confused the distinction between subroutines and
function calls -- in Pascal -- with the distinction between call by
value parameters and call by name parameters). I was convinced that
none of them would ever "get it". Some of those students have gone on
to get Ph.Ds in computer science, and from their occasional names in
print, they are doing just fine. I think that research in this area
might want to ask how predictive one's grade in the initial CS course
is of one's overall success in this field.
Robin Jeffries
