> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On > Behalf Of Antoine Leca
> What is the problem? > Do not tell me you worry about the size of the resulting package... All I am worried about is cleaning up ISO 639, including making sure that it's clear what the denotation of every item is, identifying any synonyms, and resolving any other such issues. I'm not at all concerned about the cost of having 101 locales vs. 100. I'm concerned about things like the costs to implementers of mismanaged resources due to duplication (e.g. UI strings or Web content that gets out of sync), or the cost to users who can't find the content they want because it's tagged with a different but synonymous identifier. Having synonyms in a coding system is not a great idea (or, at least there are added implementation costs if a coding system has synonyms); having undocumented synonyms is simply bad. > Furthermore, you (and your company) do not want to play politics, do you? I'm not trying to resolve a specific issue that my company has; I'm trying to evaluate what the committeee charged with maintaining a particular ISO standard should do. In this particular case, deprecation is one option that the committee can consider; I'm trying to examine whether that would create problems with implementations (not problems resulting from geo-political sensitivities) that might make that option too costsly. BTW, just to be clear, my question wasn't about the deprecation of "Moldavian"; it was about the deprecation of "mo". There is no reason why the denotation of "ro" couldn't be documented as "Romanian; Moldavian". There may still be political issues as to whether the two can be equated, and perhaps some may complain that the identifier shows a bias; but it is not about whether the "Moldavian" identity has validity. > By the way, the users may not be a good source for informations on such a > question... The ones which do have an eye about the ID are computer people Just to be clear, the particular question I was asking was not meant to be directed at users; it was meant for people involved in IT implementations. That's why I asked it on this list. > As you have already noticed in this thread, the issue is debatable A bigger issue than the question I asked is certainly debatable; the question I had in mind, I think, is not. Let me put it this way: what kind of implementations might break if it were declared that mo is a synonym of and, therefore, it is recommended that systems interpret mo like ro and stop any further generation of mo? Peter Constable
