Re: Significant figures - a Modest Proposal

[brian whatcott <inet@INTELLISYS.NET>]

At 12:11 8/27/99 -0500, Joe Bellina wrote:
> Try just asking everyone in the classroom to look at their watch and
> write done what time it is.  Very common place, but it makes the point.
>
> cheers

I have never quite lost my sense of wander at the timepieces we
now casually carry round on our wrists.
    This Timex five year analog has no more visual complexity
than a second hand. It is the kind you can buy for around $30.

   And yet it provides performance comparable to expensive
 chronometers of earlier years.  While I type this note on a
screen editor, an inset digital time of day window is updated
 at one minute intervals from a time server (in my case, from
'Muss' at McMaster) which computes the average internet packet
 delay time from there to rural SW Oklahoma and indicates
accordingly, a fair representation of the International time standard
shifted to my preferred time zone.

 With this sort of standard easily available in the classroom or lab,
it would be quite interesting to ask for the time to be noted at
a given moment (ready, steady, now!)

 Given this signal could be expected to provide an absolute
indication of zulu within a very few seconds, the statistical
manipulation of the samples, showing a raw mean, the effect of
eliminating 1,2, or 3 outliers at each tail of the distribution
and reworking a mean to see how well your sample can bracket
the absolute value might well be instructive.

This note was held for transmission at 10:35 pm local
(allowing 5 secs advance to handshake the local mail server)





brian whatcott <inet@intellisys.net>
Altus OK