The prefixes issue is important because it's impeding the success of
metrication. A common complaint from the public is that "metric is fine for
science but impractical for everyday life" due to the long, clumsy numbers
and names in many everyday metric quantities. Specifically, people often
object to millimeters. What to do about this?
(1) At one extreme is the laissez faire school that says people can use
whatever prefix they want--"let the marketplace decide." This is akin to the
argument from anti-metric groups like F2M. But the public has never really
had "freedom to measure." They've always been forced to deal with whatever
units or prefixes are dictated by government, industry, and business. This
attitude would encourage the further proliferation of new, awkward, and
unnecessary multiples we've seen in recent years (e.g., daPa, hPa, daN).
(2) The opposite extreme says we should banish prefixes that are not
multiples of 1000--call them the 'hdadc' prefixes for want of a better name.
This would eliminate the very practical centimeter--arguably the most common
multiple that people actually measure, as well as the first one children
learn and the one marked on countless millions of rulers and tape measures
all over the world. As others have noted, banning these prefixes would make
the expression of many areas and volumes preposterously clumsy, with long
rows of non-significant leading or ending zeroes (e.g., 8 hm3 would have to
be expressed as 8 000 000 m3 or 0.008 km3). There is a simple solution to
that problem: the CGPM could invent special short names and symbols for the
square meter and cubic meter, as they have done for many other derived
units. The names quad (q) for m2 and vol (v) for m3 have been suggested.
Then the regular prefixes would work in the usual steps-of-1000 way and we
wouldn't need the hdadc prefixes for area and volume, which confuse most
people anyway (because the prefixes are squared or cubed). However, there is
no indication that the CGPM would ever do this. Expanding the range of
prefixes currently used with liter to include kL, ML, GL, TL, etc. is not a
satisfactory solution because the liter is a non-SI, non-coherent unit that
would complicate calculations.
(3) A third opinion says we should avoid the large (>kilo) and small
(<milli) prefixes because they are not yet common, e.g., Mm, Gm, Tm, Mg, Gg,
and Tg. This strikes me as circular reasoning and self-fulfilling
prophecy--if we avoid them because they're not common, then they will not be
common and therefore avoided. This logic is akin to U.S. media editors'
refusal to allow SI units because they are not commonly used in the media.
Besides, the prefixes M and G are now very common and well known to the
public, thanks to computers, and others are becoming increasingly common as
technology improves.
A SUGGESTION
Surely we can find a happy balance between the extremes above. I suggest
this test: a prefix is acceptable with a given unit only if it is
demonstrably SIMPLER (shorter and easier to write and say and clarifies
precision) for the great majority of measurements in that range. I don't
think that tradition or force of habit is a good enough reason to use, or
not to use, a given prefix.
This test would allow (but not require) the hdadc prefixes only with volumes
(hm3, dam3, dm3 = L, cm3 = mL), areas (hm2 = ha, dm2, cm2), lengths (cm and
possibly hm, dm, dam), and perhaps the hectogram (hg). I am undecided about
the wisdom of cL and dL. Neither is coherent, and the liter is not an SI
unit anyway. The U.S. authorities recently rejected wine importers' request
to allow cL on labels.
THE STANDARDS
The BIPM 'brochure' says nothing on this issue, but it's an historical and
legalistic document, not a practical guide for everyday use of SI. The
language of the US standards (NIST, ANSI, etc.) reflects the 'simplicity'
test above, although they are aimed at a technical audience and are lacking
in examples.
