Ernie Wright said:
I prefer to think of it as a "lie-to-students" - that is, it's a
strategy that works perfectly well until they know enough about
programming in general and floating-point in particular to realise that
there are circumstances in which it *doesn't* work
I'm not a fan of lie-to-students, especially in a
case like this where it would be very simple to
say something like:
"Always use double because it usually turns out
to be just as fast but you get extra accuracy"
and then they can file away that justification,
and more importantly, it makes them aware that
there may be cases where you use float (even if
they don't know what those cases are at the moment),
and the inquisitive students can follow it up later.
In school, I was taught electricity using the
"dump-truck" analogy: each electron is like
a little truck full of energy, and when it gets
to the bulb it dumps its load and we see the light,
then it goes back into the battery and gets a new
load.
This might explain an utterly simple circuit but
it breaks down very quickly. For example, a circuit
with two bulbs of differing resistances. I was
able to compute the right answer to the current
flow in each branch by using the formulae provided,
but I could never figure out how each dump-truck
decided which way to turn at the intersection and
why they decided that a certain ratio of them would
go one way when they didn't even know what was ahead.
In fact it was not until long after I had finished
my physics minor at university that I read some
websites with proper explanations and I was able
to finally disabuse myself of the ingrained misconception!
A similar thing applies for other people with air
flight; it is woefully common to find people who
think planes fly because the air going over the
wing has to go faster "to catch up with" the air
under the wing.
(I've even seen this printed in a kids education
section of a newspaper).
