Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-l] season's greetings



Rick Tarara wrote:

then the relatively simple concept of significant digits


What's the evidence that there's anything "simple" about sig digs?

I'm fond of the _building block_ approach to learning. Presumably the
students already know about the Hindu numeral system we use. An
expression of the form 1.23±0.45 is simply understood as two Hindu
numerals, one of which expresses the central value, while the other
expresses the uncertainty. What could be simpler than that? I can see
that the sig-fig representation is _shorter_, but is it really
simpler? Gen-ed students ought to be delighted to use a notation that
is more explicit. They should be delighted to use a system that
involves less conceptual workload, especially when the cost is so small,
namely the time spent writing the uncertainty explicitly. This is such
a favorable tradeoff that it hardly counts as a tradeoff at all, AFAICT.

Later in the course we will do a lab where we get concerned about doing a
more sophisticated error analysis,

Then why waste time in September with a dead-end idea like sig figs?
Sig figs are not connected with anything they've used before, and not
connected with what they'll need later.


If/when people _think_, they tend to think in terms of the formalism
they've been given. How are they supposed to think about the results of
a null experiment, e.g. a result of the form 1±20? And at the other
extreme, how are they supposed to express the cm/inch conversion
factor? I express it as 2.54±0. Do you really have a "simpler" way of
doing it? If so, please explain.

I'm sympathetic to the argument that students may need to crawl before
they can walk ... but sigfigs are not like crawling; they're like
trying to waltz while wearing 27-inch clown shoes.