In both these cases, a coarse language code is enough.
So I support the idea of using the ISO standardised
One of the possible problem (this time, serious one) of Harald's
proposal is that future extension of ISO 639 may itself allow the
so that sublanguage tag assignment of IANA and ISO may contradict.
So, isn't it better to propose such a quite reasonable sublanguage
tag mechanism directly to ISO?
An additional advantage is that this
makes co-working with X.400 simple, since X.400 uses
the ISO codes.