Re: [Ltru] Re: Macrolanguage and extlang

Hi -

(as a technical contributor)

Perhaps it would be helpful to take a step back and consider
some goals (and non-goals) here.  For example:

Goal 1a: do not trigger retagging of existing ar-* and zh-* materials.
Goal 1b: do not trigger retagging of existing tagged material in other languages
Goal 2: avoid introducing "synonym" tags - there should be a single tag for a given language
Goal 3: minimize changes to matching specification

Non-goal 1: using the registry to solve the general problem of fallback
Non-goal 2: identification of all language varieties in a single subtag

I also think it should be a non-goal that the tag for a language being introduced to
the registry for the first time would have to look like existing tags.  Since
by definition there's no legacy of existing tagged data to worry about (goals
1a and 1b don't apply) we should strive to uphold goal 1c in whatever scheme
we decide to use.

I think the key issue underneath this is whether there are any cases similar to
ar-* or zh-* in the data to be added, and whether supporting those cases in
a similar manner is worthwhile.  Rather than a theoretical discussion of fallbacks
in general, I'd like to see identification of the specific worthwhile cases among
the languages slated for addition.  If there are such beasts and they are deemed
worthwhile, we should make provision in the registry format and procedures to
handle them.  If not, then we should go with the simplest approach that will
let us attain our other goals.

Randy

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