Re: [Ietf-languages] BCP47 violation in the recent extlang ajp change

[Doug Ewell <[e-mail]>]

Hugh Paterson III wrote:

> 1. Any larger BCP-47 effort for revision should likely include
> incorporating the Unicode Extensions, RFC6067 and RFC6497:
>
> https://cldr.unicode.org/index/bcp47-extension
> http://www.unicode.org/reports/tr35/#36-unicode-bcp-47-u-extension

The T and U extensions are quite properly described in separate RFCs, as specified in Section 3.7 of RFC 5646. Folding them into an RFC 5646bis would place them on a different definitional level from any future extensions that might be created, which would be a Bad Thing™.

I assume the purpose in doing so would be to call better attention to the extensions, since RFC 5646 itself predates them and so cannot refer to them. The Wikipedia page on BCP 47 already describes the extensions and links to the extension RFCs, as does my page; perhaps the maintainers of other sites that link to RFC 5646 could be persuaded to link to those documents as well (and also to RFC 4647).

> 2. The issue with using Glottolog codes is specifically that ISO 639-3
> is scoped at the language level, while the informatic model of the
> Glottolog is specifically scoped to the document level, rather than
> the language level. That is, Glottolog codes can span the classes of
> macro-language, idiolect, dialect, or hypothetical reconstructed
> language.  The two scopes are inconsistent with regard to purpose. If
> the purpose of the positions of the constructed BCP-47 tags is to stay
> consistent, I don't see how Glottolog codes become a possibility
> except after -x- or some other yet-to-be defined identifier.

I agree wholeheartedly, but one does occasionally hear suggestions to do this. It’s almost as if 8,000 language subtags, plus variants and the ability to register more, plus private-use sequences, weren’t enough.

--
Doug Ewell, CC, ALB | Lakewood, CO, US | ewellic.org