[Comment on “The Haskell Norm”] More on units of measure



The pseudonymous Dev L. Advocate rightly chides the S.1. people about their weird and prolific prefixes (Eos, July 30, 1996, p. 291). But propagating further standard units of measure is not really a better way of handling wide-ranging numbers.

Richard P. Feynman pointed out in a letter to the editor of Scientific American in November 1970 that scaling factors properly belong to the numbers themselves, not to their units of measurement. Rather than prescribe strange prefixes, or propagate new units like Dev's Haskell, notice should be taken of the commonly used representation of scaled numbers in so-called E-form, under which one Haskell would be given as 1E21 Pa s or 1E22 poise. The problem with this style is that any number has too many convenient representations because the scaling base is only 10.