proposed ISO 639 change for "arn"

[Phillips, Addison]

> >
> > However, sounds like ISO 639 is also concerned about stability and doesn't
>> want to withdraw the code altogether. Moving it to ISO 639-5 doesn't help,
>> since those codes aren't deprecated in BCP 47. This might be a very special
>> case? (The three cited by Philip Newton are historical and are documented in
>> the body of BCP 47).
> 
> I was under the impression that this was a preliminary discussion to gather
> additional feedback to help ISO 639 make an informed decision, rather than
> deciding independently what to do within the scope of BCP 47. At least, that
> was the framework under which I was asking my question. I imagine we'll be in
> a better position to decide what to do with the old and new codes once a
> decision is made in ISO 639.

Yes. And my point is that neither solution suggested by Peter (moving the code to ISO 639-5 or making it into a macrolanguage) results in deprecation in BCP 47. And to point out that BCP 47 *could* deprecate-and-prefer the new code(s) without that, although BCP 47 says that this is something we should not do.

Put another way: BCP 47 is written to follow whatever ISO 639 does except in very selected and well-described cases. Neither of the options Peter presents would make qualify to deprecate the existing code. Specifically, Section 3.4 says:

2.     Values in the fields 'Preferred-Value' and 'Deprecated' MAY be
        added, altered, or removed via the registration process.  These
        changes SHOULD be limited to changes necessary to mirror changes
        in one of the underlying standards (ISO 639, ISO 15924, ISO
        3166-1, or UN M.49) and typically alteration or removal of a
        'Preferred-Value' is limited specifically to region codes.

Addison