Michael Everson wrote: > So, what helps us move forward here? How about zh-Latn- for now, that can't be wrong. If 4646bis fixes this systematic variant oddity somehow we are either already where we should be, or we can add zh- Frank