mark.davis at icu-project.org
Fri Oct 21 21:54:15 CEST 2005
I don't know that I would put it quite so strongly; many nations have
stable internal divisions. However, I agree with the main point, which
is that in its current state, it is not anywhere near complete enough to
use, does not distinguish between stable and unstable codes, and is not
accessible. Thus the ISO 3166-2 codes are essentially useless for
application within any other standard, including language tagging.
>Harald Tveit Alvestrand scripsit:
>>>In the USA there are a number of states, and there are well-known
>>>abbreviations for the states. The same in Canada, for its provinces.
>>>India and China and Brasil and Germany and Switzerland and probably many
>>>more have such subdivisions of their nation. Great Britain too.
>>>I even think there is an ISO 3166 part for such subdivisions, but I am
>>>not sure of the part number, nor whether the standard has been finalized.
>>ISO 3166-2. The code for Scotland is "SCT".
>Unfortunately, the ISO 3166-2 codes are stable neither in principle nor in
>practice, neither in code nor in referent. The U.S. is a special case;
>its states are not mere subdivisions but are mutually sovereign entities
>whose existence and territorial bounds are constitutionally guaranteed.
>Other countries can and do revamp their subdivisions at will, and
>ISO 3166-2 serves as a mere registry of the code elements and subdivisions
>in effect at present. Furthermore, its contents are not freely available.
>So ISO 3166-2 is completely useless for language tagging.
More information about the Ietf-languages