RE: [Ltru] RE: ISO 639-2 decision: "mis"

From: Randy Presuhn [mailto:randy_presuhn@mindspring.com]

>> but that is part of the documented contract around 'mis'
>
> Which "documented contract"?

I expect the ISO 639 RAs will be providing a usage note wrt 'mis' explaining that its extensional semantic must be defined in relation to an application context. This was one of the action items listed if the proposal to change the name were approved. I will follow up with the RAs on getting that done.

>> -- and why users need to understand its shortcomings, so also why users
>> would probably be warned against exporting data.
>
> This is making the case for using "MUST NOT" rather than "SHOULD NOT".

IMO, this is in the same line as people using a tag like "en-Latn": we say "the script subtag SHOULD be omitted" in such cases, but not that they MUST NOT include it in such cases. We expect users to make informed choices; we're not going to police them on certain things they *could* do but that wouldn't be wise to do in general.

>> The only issue here is the inherent instability and shortcomings of 'mis'.
>> I don't see any abuse of BCP 47. There is nothing whatsoever like using
>> "sw" to mean 'Swiss German'.
>
> Yes, it is in that permitting the tag to be used for languages for which tags
> already exist is in direct contradiction to its description.

The issue here is evaluating "languages for which tags already exist". You're assuming that every application of BCP 47 must determine that based on the complete inventory of the LSTR. I'm assuming that a higher-level protocol can restrict the contents of the LSTR it permits and that it can also determine the extension of 'mis' on the basis of its restricted set. I see these as two options we can permit; I don't see one of these excluded a priori. It's a choice we should discuss and probably clarify in documenting 'mis' (unless we decide to deprecate it).

Peter

_______________________________________________
Ltru mailing list
Ltru@ietf.org
https://www1.ietf.org/mailman/listinfo/ltru