draft-phillips-langtags-08, process, specifications, "stability", and extensions

[Bruce Lilly]

>  Date: 2004-12-30 10:46
>  From: "JFC (Jefsey) Morfin" <jefsey at jefsey.com>
>  To: "Peter Constable" <petercon at microsoft.com>, ietf-languages at alvestrand.no, ietf at ietf.org

> 1. OSI 3166 is refered to. RFC 1591 should. RFC 1591 introduces differences 
> (we all live with) with OSI 3166 which is taken as a reference to know what 
> is a country.

RFC 1591's scope is the DNS, not language tags.  I see no
relevance to the current discussion. Domain names and
language tags are different types of names, used for
different purposes, and with different scope (largely
non-overlapping, though one might legitimately ask how
one is supposed to determine the language of an
"internationalized" domain name...)

> 2. OSI 639 scripting fr-FR is used while RFC 1958 leads to fr-fr or FR-FR 
> or FR-fr indifferently and calls for fra-fr to avoid confusion.

RFC 1958 states (and RFC 2277 reinforces) that fact that
"names" (as that term is used in both documents, meaning
tags) are case-insensitive.  That is consistent with the
case-insensitivity aspects of the draft under discussion.
I see nothing in RFC 1958 that would call for "fra-fr"
rather than "fr-FR" (specifically, w.r.t. 1958 sect. 4.2
the 2- and 3-letter ISO 639 codes are part of a single
list of codes; w.r.t. sect. 4.4 "fr-FR" (in any combination
of letter case) is unambiguous as a language tag).