In our project, SPIP (www.uzine.net/spip), we're using a similar list, but in which every language is identified in its own writing. When two translations belong to the same ISO family, we just use the branches pt_BR and pt_PT, whereas when only one translation belongs to the family we use just the family code (and name), unflavoured. This can help groups join in a common effort, then split *if* needed. (However, we don't have that much experience to say if it works; but it's really an exciting process).
BTW, and slightly off-topic, the usual translation tools were really too clumsy to use (you can't expect translators to know which one to choose, and to install it and understand it -- at least it raises the bar too high --, plus you want to have a team effort, so you need people to be able to tap the same translation database). We finally decided to build a translation website with lists of "strings awaiting for translation" in every language, and a web interface to edit them. It works really well.
@ M.-A. Lemburg <mal@lemburg.com> :
Barry Warsaw wrote:
On Wed, 2003-04-23 at 08:30, M.-A. Lemburg wrote:
Why not stick to the standard English language names ?
Good idea, thanks for the link!
Here's the complete list I'm using (collected from various sources):
# # Mapping tables # # Keys must language names. The tables themselves must map # ISO codes using lowercase letters only to language names. #
-- Fil