Divide supported locales into two categories
There should be two types of locales supported by I18n package. One collection of Locale object would represent a group of locales fully supported by the system (e.g. locales for which translation files and other localized resources are available) - as it is now. The other category would consist of all locales supported by CLDR itself. There should be possibility to use functionality based on CLDR data (like number formatting, parsing etc) for locales which does not belong to the first category (the "fully supported" locales).