(I don't know how to post to the CLDR list, so sorry if this question goes to the wrong place, this message may be forwarded there)
ISO 3166 has made some privisions for codes reserved as they are already in use by ITU and/or WIPU, even if they are not strictly "assigned". These code do exist as there are legitimate needs for use within the area of what we call "locales", even though they are not strictly countries or semi-autonomous regions and territories, justified by the fact they they are separated from their mainland and thus need some separate technical solutions to cover them. Here I think about the following codes, used outside of strict ISO 3166 assignments: * [EA] Ceuta and Mellila, two dependencies of Spain [ES] on the Mediterranean coast of Morroco [MA]. Their special status do not make them autonomous regions like other regions of Spain, and they can't still be considered as part of Morocco, as the local population considers itself as Spanish and not Moroccan, and mostly speak Spanish. The two cities have a status comparable to Gibraltar [GI], a dependency of Great Britain (GB]. * [IC] Canary Islands, a plain "Autonomous Region" of Spain [ES] (with a status similar to other regions of Spain since the adoption of the new post-Franco constitution which gives some local governance with the right to legalize another language than Castillan Spanish.) The Canary Islands speaks both Castillan and one of the four main regional variants of Catalan spoken in Spain. This code is used in ITU, but would find also applications to make a distinction between the continental forms of Catalan [ca] spoken in Spain [ES] or France [FR]. There's probably another code in ISO 3166-2 for this region of Spain. But all references I found about it was with [IC] rather than [ES-IC] or something else in [ES-*]. * [AC] Ascension Island, a dependancy of Saint Helena [SH], itself an island dependancy in Atlantic Ocean of Britain (GB] both located West of Africa. However the Ascension Island is far enough from Saint Helena that the ITU needs a separate code for it. May be there's a [SH-AC] subcode in ISO3166-2. * [DG] Diego Garcia, an island which is part of the British Territories in the Indian Ocean [IO], with a joint military presence of [GB] and [US] with an important air-naval force. The special administrative status of this island make it inappropriate within [IO], and some databases will need a separate code for it. * [CE] the "Council of Europe", or CoE (no confusion here with the "European Council" which is one governing institution of the European Union [EU], grouping the ministries of the full EU member countries, today 15, 25 countries in a few days on May 1st). In addition to the 15 existing EU members and the new 10 EU members, the CoE includeds these 19 + 1 countries, by order of membership: - Norway [NO], (in EFTA too) - Turkey* [TR], - Iceland [IS], - Switzerland [CH], (in EFTA too) - Liechtenstein [LI], (in EFTA too) - San Marino [SM], (in EFTA too) - Bulgaria* [BG], - Romania* [RO], - Andorra [AD], - Albania [AL], - Moldova [MD], - Macedonia (the Former Yugoslav Rep. of) [MK], - Ukrainia [UA], - Russia [RU], - Croatia [HR], - Georgia [GE], - Armenia [AM], - Azerbaijan [AZ], - Serbia-Montenegro [YU, CS?], - plus Monaco [MC] which is in discussion since several years for a later membership to the CoE. (* those counties marked with a * are also candidate for the next round of E.U. membership). The only few remaining European countries that are not members of the Council of Europe are: - Belarus [BY] - the Holy See (State of the City of Vatican) [VC] - ... (sorry, I may have forgotten a few of them) The code [CE] is used in many juridic databases related to the European legislation. I have been told that membership to the European Union required first to become a member of the CoE (I'm not sure of that, even though every of 15+10 EU member countries comply to this rule, which gives competence to the European Court of Justice in La Hayes, where the EU and EFTA are both represented too). Localization is often driven by language identification, but not only. Official status of countries and territories also have its importance (currencies, import/export controls on products, services and technologies, tax, and even the time format or definition of timezones and daylight rules...). It's hard to put all these distinctions into a language code, and a locale is effectively a set of conventions (which goes even further than what the POSIX locales allow to set with a single environment variable like LANG or LC_ALL). My question is then: if Unicode will concentrate to provide a "stable" reference for localized data, based on historic RFC 3066 and ISO 3166 and ISO 639, what will become these reserved codes? And what about other international organizations (notably the United Nations [UN] when it administrates or regulates regions like Kosovo in Serbia [YU -> CS?], or the Antarctic territories, or historically the Neutral Zone in Cyprus.
