From: J Andrew Lipscomb (firstname.lastname@example.org)
Date: Tue Aug 23 2005 - 20:48:01 CDT
Pondering the issue of stability for a 3166-2-like set of areas: I
think you wind up having to use meaningless codes (à la FIPS 10-4).
This would allow stability at the total expense of human-
readability. While the UCS is a domain in which stability is
reasonable because there's plenty of room (relatively speaking) and
the creation of new data bits is orderly, sub-national entities
change regularly, and I'm not sure you can maintain a truly human-
readable code while forever disallowing any previously used codes.
That said, in my own project of creating codes for sub-national
entities, I've determined that you can usually get a decently human-
readable yet unambiguous code within three letters. (The only
exception I've found so far is Texas, where I had to go four letters
to code its 254 counties.)
This archive was generated by hypermail 2.1.5 : Tue Aug 23 2005 - 20:48:58 CDT