Paul Jakma wrote at 10/10/06 22:11: > On Tue, 10 Oct 2006, Ienup Sung wrote: > >> Regarding the optional unit character, in particular, 'K', >> it seems people tolearate Km, km, and KM for Kilometer/kilometer. > > > You mean kilometre. ;) > > There's just one country which does not use the normal SI spelling of > 'metre' and that country does not use this unit in its road signage. ;)
Meter is also quite commonly used in many other countries and as a matter of fact more than metre in many countries. > >> Usually though km is being used and people use a ligature for such >> characters like '' in many (Asian) cultures. > > > It's pretty simple really, we have a choice between following: > > - the globally recognised symbol set defined by SI, which is familiar to > everyone with any level of basic scientific education > > or > > - Non-SI and rediscovering the lovely ambiguities that led to formal, > international, standardisation of prefixes and units in the first > place. > > In many cases, I'd bet that local deviations: > > a1) are made mostly for typographical reasons > > a2) are mistakes made by people either uneducated in, > or forgetful of science, e.g. 'km' (kilometre) versus 'Km' > (Kelvin-metre). > > b) That such deviations typically won't have any local consistency, > precisely cause they're just 'mistakes'. > >> On the proposed spec which uses 'b' character to indicate radix-2 or >> radix-10, *BSD uses a flag value and based on that, the output will be >> 'k' for radix-10 and 'K' for radix-2 IIRC. > > > For the 2**(10*n) prefixes, the correct SI-ology is: > > kibi (Ki) > mebi (Mi) > gibi (Gi) > etc. > > --paulj The point that I am and would like raise is incompatibility issue with *BSD interfaces. There is no reason for us to be different in terms of spec including function prototyping if you want to use the same C function name. If you want to deviate from the *BSD original spec, that's fine but rather should use other name in that case. Ienup
