Date: 8 Feb 90 18:11:02 GMT
From: mnetor!utzoo!henry@uunet.uu.net  (Henry Spencer)
Subject: Re: Imperial measurements

In article <9002072000.AA23577@cmr.ncsl.nist.gov> roberts@CMR.NCSL.NIST.GOV (John Roberts) writes:
>... As I
>understand it, there are (or recently have been) two acceptable formats
>for SI expressions, chosen because they produce "convenient" values for
>many everyday measures: centimeter-gram-second (arg! hate that one - and its
>in all my old books!) and meter-kilogram-second (the most common one now, and
>probably the "true" standard). ...

You're slightly confused here.  The problem is that the original metric
system did not tie down a lot of the details, particularly as scientific
and technical use spread into new areas like electromagnetism.  So there
was a wild profusion of different metric systems with different units in
the more esoteric areas.  For example, c-g-s and m-k-s defined different
units of energy, the erg and the joule respectively.  SI is the definitive
attempt to clean up the mess and settle on a single unified system.  There
is *one* acceptable format for SI expressions, and only one, basically a
cleaned-up version of m-k-s.  C-g-s is pre-SI and has nothing to do with it.

> Both of these formats suffer from an
>idiosyncrasy left over from the beginnings of the system - measures are in 
>terms of either a hundredth of one basic unit (length) or a thousand of
>another basic unit (mass). Also, common usage includes units that are not
>part of SI, such as the liter (the vulgar term for the cubic decimeter).

Again, this is confused.  SI stresses 10^3 as the basic multiplier very
heavily, with tenths and hundredths strongly discouraged.  One or two
exceptions are made for everyday use, notably the centimeter and the liter,
but these are officially considered concessions to convenience and their
use for technical purposes is theoretically discouraged.

>It is very seldom that an astronaut desperately needs to know the diameter of
>a bolt in terms of an eighteenth-century survey of the distance from the
>North Pole to the Equator through Paris - it's more important to quickly
>pick the right size wrench. Thus, bolt sizes should be "number 1", "number 2",
>and so on...

In fact, this is what is normally done.  Sometimes there is a unit attached
to the arbitrary numbering, but it is often meaningless -- a "one-inch"
pipe has no dimension that measures one inch, and a "two-by-four" is never
two inches by four inches (not even when rough-sawed, nowadays).  As you
point out, in applications like this the actual measurement is not very
important, so long as X matches Y and the user has a feel for what the
numbers mean.  That's actually true of many of the non-metric uses in
aviation as well:  if the unit of altitude, the foot, magically changed
in size by 5% overnight, and all the instruments and charts changed to
match, almost nobody would notice.  The issue is seldom how many meters
up you are, but whether you are where others expect you to be, whether
you have adequate clearance above that mountain ahead, or whether your
rate of descent is too high for your landing gear to absorb.
-- 
SVR4:  every feature you ever |     Henry Spencer at U of Toronto Zoology
wanted, and plenty you didn't.| uunet!attcan!utzoo!henry henry@zoo.toronto.edu