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

Frank Ellermann nobody at xyzzy.claranet.de
Thu Dec 30 06:55:21 CET 2004

Addison Phillips [wM] wrote:

> http://www.inter-locale.com/ID/draft-phillips-langtags-09.html
> Your comments on that would be appreciated.

The source of en-NH is still unclear, I don't have a copy of
ISO 3166 3rd ed. 1988.  Even if somebody is willing to pay for
it he may have a problem to get exactly this edition (?)

Maybe use the existing IANA ccTLD registry instead of ISO 3166:

Garbage in, garbage out, but at least I know where to find the
IANA garbage ;-).  With your 1988 edition you have a problem
for TP (now TL), DD (now a part of DE), PS (introduced after
1988 IIRC), etc.  I don't check Yemen and Zaire.  If you really
must use an old ISO 3166 edition use the 5th edition 1999 (they
commited net suicide with CS in 2003).

ISO 3166 has numerous non(sense)-country codes, not only clear
cases like FX.  Try ccTLDs, there you'd also get weird stuff
like SJ or AQ, but at least nobody plans to recycle old ccTLDs.

> Frank has raised some good issues: I believe I responded to
> his message.

Sorry, I somehow missed it, or it was very recently on the IETF
"language" list (I only look into "general" regularly, skipping
the admin stuff it's a quiet list).

> The draft specifies ONE mechanism, just like RFC 3066, and
> notes that more specialized processing is possible.

Okay.  Maybe use en-boont as a Caveat, where en-US-boont would
be missed with your algorithm.  That also covers my de-CH-1996
problem.  You could add se-Latn-AX as second example, because
it's not only a potential problem for country codes, it's also
a potential problem for scripts, when the script is more or
less irrelevant, because it's the default for the language.

And while you're at it maybe add some words about i-default,
if I got it right you would expect IANA to mark i-default as
"obsoleted by UND" (?)  It's one of these odd cases, probably
you also expect IANA to mark i-klingon as "obsoleted by TLH".

Actually that's already the case with RfC 3066.  That's strange,
Harald registered i-default (December 2001), and he also wrote
RfC 3066 (January 2001), and RfC 3066 says that all i-whatever
should be deprecated as soon as an ISO 639 code is available...
Maybe TLH and UND were introduced later.

> The current draft REPLACES RFC 3066.

AFAIK the Unicode consortium plans a registry of locales, stuff
like de-DE etc.  I hope that your ideas are compatible with
whatever they do (I've no idea, sorry)

                       Bye, Frank

More information about the Ietf-languages mailing list