Alain LaBont� a �crit: > So the discussion started with stability of "country" codes... If a code > represents a political territory, it can NOT be stable by nature, whether > it is alphabetic or numeric. The only stability you can expect has to do > with the non-reassignment of a code within a reasonable period, to a > different body.
To meta-paraphrase Albert Einstein (*), country codes should be as stable as possible, and no more so. It is true that country codes inherit a certain instability from the nature of countries. There's not much we can do about that. However, it is possible to limit the instability. As I said earlier, to my knowledge Romania has not experienced any inherent instability that would require the code to be changed. Neither its boundaries, nor its political status, nor its form of government have changed in any way. Instead, what happened is that somebody looked at the alpha-3 code "ROM" (which was assigned DECADES ago and which is derived from the first three letters of the country's name in its own language), discovered that it spells out the name of an oppressed ethnic group within that country, and panicked that they might be identified, by the use of the country code "ROM," as belonging to the ethnic group "Rom." This would be tantamount to assuming that all people from Benin, Guyana, or Kenya are named Ben, Guy, or Ken, respectively. The relevance of all of this to Unicode (besides the link to internationalization) is that, unlike ISO 3166, Unicode has a policy that forbids changing the name or position of a character once it has been assigned. This policy frustrates some people, such as the North Koreans and Cambodians, but in general it serves to assure users that following today's standard will not result in invalid data next week. (*) Albert Einstein is popularly (but probably incorrectly) quoted as having said, "Everything should be as simple as possible, but not one bit simpler." -Doug Ewell Fullerton, California
