Quoting Vincent Danjean (vdanjean.ml@free.fr): > I would add that I recall discussions on their ML about introducing variants for > the French language itself. The fact is that, depending on the context, some > strings may need to be translated differently. So it is possible that, in the > future, localization will be further split. This is overzealous and useless for French. It just makes localization efforts dispersed for the sake of one or two words that *might* be different. I don't want to be too rude, but I would say in first instance this is just stupid. > Ok. Many thank for all your information. > So, unless collision, I keep only the two letters in the debian package name. > The exceptions for now are: > - es-es > - es-ve I don't agree with those. From what I heard of Spanish speakers in Debian, they are perfectly able to get a consensus about written Spanish to avoid splitting translations between es_ES and the various Latin American countries. Also, what about es_AR, es_CL, etc? I would therefore suggest having one of the two labelleed as "es" and the other one kept as is. > - zh-cn > - zh-tw > For the language name, in the package description, I only talk about the > language itself with few exception where the country is also stated. > These exceptions are the same as before: > Spanish (Spain), Spanish (Venezuela), Chinese (China), and Chinese (Taiwan) zh_CN is more usually called "Simplified Chinese" and zh_TW is "Traditional Chinese". Arne can give more input on this. > and two more exceptions: Bokmaal (Norway) and Nynorsk (Norway) as it is > difficult to guess the country from the language name (but perhaps, I'm too Call them "Norwegian Bokmål" and "Norwegian Nynorsk". They are the names in ISO-639
Attachment:
signature.asc
Description: Digital signature