[Ietf-languages] Re: Proposed forms for "Bihari languages"

[Doug Ewell <[e-mail]>]

Winston Sung wrote:

> If I remembered correctly, to achieve the compatibility and stability,
> we (language subtag registry) don't follow ISO 639 deprecations?

We do follow them, by deprecating — not removing — subtags that correspond to code elements that have been deprecated (some standards call it “withdrawn”) in the core standards.

Most recently, we deprecated 'yol' just a month ago, 'dek and 'nte' in 2024, and many others before then (e.g. 16 in 2023 and 10 in 2022).

> Like, if we did that, that means, every system depends on BCP 47
> subtag registry can no longer treat BCP 47 as a "long-term stable"
> system, and need to continuously replacing newly-introduced deprecated
> language tags.

Our stability consists of keeping deprecated subtags in the Registry, with a Deprecated value showing the date of deprecation, and (crucially) the Preferred-Value and/or comments indicating what tag or subtag should be used instead of the deprecated one. The Preferred-Value can be used by processes to automatically link old and new tagging. Deprecated subtags can continue to be used for tagging and searching, although it is not recommended to do so.

This is in stark contrast to the way some of the core standards operate, or at least used to operate, in which withdrawn code elements would simply disappear. Stability is a primary purpose of BCP 47, and one of the main reasons that, instead of relying on the varying, often unstated stability policies of the core standards, we created a Registry that would include any subtag ever valid in a BCP 47 tag.

> If we're replacing bh with bih, that means we need to do the same
> thing for sh → hbs,

That’s actually a great point, and another one I missed. [sh] was deprecated 26 years ago, and the Set 3 macrolanguage [hbs] exists as a suitable Preferred-Value. We need to do that as well.

> and every other single ISO 639 deprecations, if I understand
> correctly.

We should have all of them, except [bh] and [sh], if I’m reading the Set 2 list correctly: [mo] in 2008, [jw] in 2001, and [in], [iw], and [ji] in 1989. If I’ve missed any, please let me know.

> Please correct me if I misunderstood something and clarify whether we
> are going to follow all the ISO 639 deprecations over the
> compatibility and stability.

Our approach to handling the concept of deprecation reflects the best ongoing expertise of the core standards, while simultaneously ensuring that old subtags can continue to be used, and never lose their meaning. We were extremely careful not to sacrifice one for the other.

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