[Seaside] Seriously demented multiple choice questions

Tim Rowledge seaside@lists.squeakfoundation.org
Wed, 19 Jun 2002 16:19:18 -0700


I made a fatal mistake. I demoed the current state of the
multiple-choice-question website prototype. I basked in the glory of
having something good to show at last (thanks in no small part to
the help you guys have given me) and then...
	"so how are the algorithmic questions going?"
Gulp.
I very foolishly claimed a loooong time ago that the wonders of
Smalltalk would make it possible to have questions where the numeric
quantities could be variables and an algorithm could then generate
near-random questions from the template. ie
instead of just
Q1. What is the total cost of 4 widgets at $3 each, if tax is applied at
7.75%?
	a) $10
	b) $12.93
	c) $11.07
	d) duhh...
we could have a template like (faking furiously)
what is the total cost of {x=4:7} widgets at ${y=2:17} each, if tax is
applied at {t=7:9F}%
	a) ${(x * y * (1 + t)) floor * 0.8}
	b) ${(x * y * (1 + t))}
	c) ${(x * y * (1 - t))}

etc. I think I did a dumb thing....

Now, I'm quite sure that there is a way to do something like this, even
though I don't have a clue how right now. The bit that really worries me
is finding a way to explain how to _create_ such questions to the
average community college teacher....  not to mention a sensible UI for
it. Just for starters there is the problem that most people expect
arithmetic precedence (don't they?) as opposed to Smalltalk style. Then
there is specifying the ranges of the inputs (done above as x:y is any
integer from x to y, a:bF is any floating point from a to b, etc)

Excuse me while I go and bang my head against a nice hard wall  for a
while. Any suggestions on how to tackle this most welcome. I knew I
shouldn't get involved in end user applications.....

tim

-- 
Tim Rowledge, tim@sumeru.stanford.edu, http://sumeru.stanford.edu/tim
The severity of the itch is proportional to the reach.