23 Jun
2004
23 Jun
'04
10:06 a.m.
On Wed, 23 Jun 2004, Jules Richardson wrote:
I'm not sure if I agree with the current trend
toward open book exams -
yes, in the real world, chances are you will have reference material
handy. But the current attitude of most people I've known who have taken
such exams is that they don't need to actually know anything, because
they can just bluff their way through it in an exam by reading the book
and get sufficient marks to pass.
If a closed-book exam is offered open-book, then that is the case.
A test that is all about memorization is useless in an open book
format.
The exam needs to be different according to the format.
An open book test must be written to test application of
knowledge instead of memorization.
for example:
How many bits are used for a single precision IEEE floating point number?
is only appropriate for closed-book
int N = 10;
while (N--) printf("&d\n",N);
What are the first and last numbers displayed?
is appropriate for open-book.
Then these people go out into the real world and they
can't think for
shit - as soon as an oddball problem hits them they're just incapable of
working it through to a solution as they're too used to just being able
to read the answer in a book right when they need it.
Then these people go out into the real world and they can't think for
shit - as soon as an oddball problem hits them they're just incapable of
working it through to a solution as they've forgotten everything that
they memorized.
I know I'm perhaps a little younger than the
average age of people on
this list, but I feel I was one of the last generation who was lucky
enough to do an old-style degree course. We had access to real (and
diverse) systems rather than things being emulated, and we were given a
lot of grounding theory in the way things actually worked, and more
importantly we weren't given an easy ride - no such thing as an open
book exam then, no whizzy graphical tools to do half the work for us
etc.
In my day, we didn't have the option of using a calculator.
Did that help or hurt? Are the aspects that it helped or hurt
relevant to what is being tested?
In my day, we were to supply our own scratch paper for standardized tests,
including graph paper. We were not allowed to bring in sliderules,
but there was no rule against making one during the test!
OTOH, in the UC Berkeley School of Information Management and Systems,
I was the first student ever to use a word processor for the PhD
written exams. I managed to convince them that grading
penmanship was no longer valid.
I'm amazed at how often the fundamentals that we
were taught have helped
me work some problem out - and I've lost track of how many of the later
generations of graduates I've had to deal with who just can't think
properly because all they've been taught is how to push a mouse around a
screen.
How many current students can find a square root without a sqrt or x^y key
on a calculator?
How many even know the square root of 2 and 3?
I even get some who have been TAUGHT that pi is EXACTLY 22/7.
(I got in major trouble when I was 10 years old for telling the teacher
that pi was not 22/7)
industry on the surface of things think they want. I
do wonder quite
where things will be in ten years when there's almost nobody left who
can actually think for themselves though...
Anyway, rant over :-)
mine is just beginning.