Re: [Ietf-languages] "Black ASL"

[Doug Ewell <[e-mail]>]

Hugh Paterson III wrote:

> Are there other cases where a sub tag is generated prior to a language
> tag via iso 639-3? It just seems cleaner if 639-3 has a chance of
> being generated prior to implementation in the sub tag registry. Less
> chances of two tags representing the same conceptual entity. 
>
> I thought that in the past we had referred people to apply to the ISO
> 639-3 first.  Otherwise don’t we create a race condition to become the
> easiest entity to apply to?

The only situation in which a requester MUST contact ISO 639-3/RA first is if they are requesting a language subtag, per Section 3.6, first bullet point. No one has yet requested such a subtag, and it seems unlikely, given the strong wording of RFC 5646 on this topic, that one will ever be approved or even requested.

We have registered more than 100 variant subtags, all of which (by definition) represent some variety of an existing language. Many of these are for orthographies or transcription standards, but some are for dialects. These are generally not referred to 639-3 unless the question of “language vs. dialect” reaches a certain point, which has started to become the case here.

Determining whether an entity is a language on its own or a variety of some other language (or even exists) is one of the major responsibilities of the Reviewer, and one reason we have a two-week review period with input from list participants.

There are some BCP 47 variant subtags which could arguably represent a true language. When that is discovered, ISO 639-3 may create a code element for that language. This happened, for example, with Western Armenian. As Michael said, we then register the ISO 639-3 code element in the Registry (as usual) and deprecate the variant. This process has worked well in the past and no rivalry or significant confusion has arisen.

The BCP 47 process is absolutely not to be viewed as an “easier” approach to registering an actual language, one which should be in ISO 639-3 instead, and I doubt many people see it that way. Indeed, there is still a perception in the field that ISO 639 code elements are more desirable or “legitimate” than BCP 47 variant subtags.

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