New item in ISO 639-2 - Zaza

[Frank Ellermann]

Mark Davis wrote:

>> We'd need a rule in 3066ter, that new 639-2 tags identified
>> as macrolanguage shortly afterwards will be deprecated in
>> the registry.

> That is precisely what we should avoid -- in my book any
> option that has that as a result has a big strike against it.

Where's the problem if the hypothetical new 639-2 "rkn" is
"downgraded" to a macrolanguage by 639-3, and then marked as
deprecated in the subtag registry ?

The old Ryukyu tags would continue to work as before.  They'd
also still be canonical wrt (deprecated) "rkn-exa" constructs.

Last time we argued about canonical it was about region codes,
you wanted to keep canonical regions as is, and I thought that
this can't fly for region codes.

Now I think we can keep canonical as is for language tags, and
you prefer to replace "exa" by the new and longer "rkn-exa" (?)

Is it because LTR matching works for "rkn-exa" and "rkn-exb",
but not "exa" and "exb" ?

Somebody mentioned "number of synonyms", but that's the same
in both cases, either "exa" => "rkn-exa" or "rkn-exa" => "exa".

Frank