Re: [Ltru] Re: Punjabi

["Mark Davis" <[e-mail]>]

It is, in fact, quite relevant to what we are doing.

I have had more and more trepidations about the way in which we are thinking
of using extlang. The one and only purpose of having the more complicated
"zh-cmn" in place of just plain "cmn" is to get matching behavior without
hitting the repository (or a local store of it). That is, if I have access
to repository data, then it is clearly possible for me to do a match so that
"cmn-CN" falls back to "cmn" and then to "zh"; I don't need to burden the
language tag itself with data that is redundant with information in the
repository. It is only if I don't have access to the data that it is an
issue.

My concerns are that we are baking into place a hierarchy that may, in
actual practice, be cumbersome and not that useful. We are already not
making use of the macrolanguages that are in ISO 639-1/2, like no and sh,
and there are other cases where fallback behavior != macrolanguage, either
always or often. We just ran into one: vls (Vlaams) wants to naturally
fallback to nl (Dutch/Flemish) in doing lookup. And part of the issue with
"pa" is that it means Panjabi in ISO 639-1, which for most people
encompasses the items grouped under lah.

This is not in any way meant as a criticism of the ISO 639-3 committee; I
have been quite impressed with how well they have handled a very tough
problem. But what I am concerned about is baking in a fallback that may vary
quite significantly according to application, just to avoid hitting the
repository, when it would be simpler to just use all the ISO 639-3 languages
as the first tag, just like we do for ISO 639-1/2.

Mark


On 3/15/07, Doug Ewell <[e-mail]> wrote:
>
> As someone who's posted off-topic to this list several times, including
> many times when ietf-languages was a more suitable target, I confess not
> knowing how this discussion affects our project to update RFC 4646 and
> 4645.  All of the current and future language subtags in the Registry
> are there because they are defined in ISO 639, and they have the
> primary/extended groupings that ISO 639 gives them (except "sgn", which
> is a special case).  We don't change these.
>
> I can think of a few broad courses of action stemming from this
> discussion, but none involves the LTRU WG:
>
> 1.  The meanings and/or macrolanguage groupings associated with Panjabi
> and/or Lahnda are not adequately representative of reality, and need to
> be changed or clarified.  This should be brought to the ISO 639-3/RA at
> [e-mail].
>
> 2.  The subtags "pa" and/or "lah" should have comments attached to them
> that might help ensure correct usage.  This should be brought to
> [e-mail].
>
> 3.  Something about CLDR needs to be changed or re-examined.  This
> should be brought to the CLDR Technical Committee; see
> http://www.unicode.org/cldr/ for information.
>
> --
> Doug Ewell  *  Fullerton, California, USA  *  RFC 4645  *  UTN #14
> http://users.adelphia.net/~dewell/
> http://www1.ietf.org/html.charters/ltru-charter.html
> http://www.alvestrand.no/mailman/listinfo/ietf-languages
>
>
> _______________________________________________
> Ltru mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/ltru
>



-- 
Mark

_______________________________________________
Ltru mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/ltru