Determining supported locales is difficult and slow.
Not all API layers publish their locale support (E.g. FormatterEngineFactoryService, FormatterFactory)
Compiling a list of supported locales may require reading a large set of configuration files
The list of specifications becomes large when every input/output format is multiplied by up towards 200 locales.
Many TaskGroups don't differentiate between locales, and perhaps a default locale (i.e. some texts may be in english, is better than no result at all, at least to evaluate the usefulness of the software as a whole)
Determining supported locales is difficult and slow.