Peter Constable petercon at
Fri Feb 18 16:37:17 CET 2005

> From: ietf-languages-bounces at [mailto:ietf-languages-
> bounces at] On Behalf Of Michael Everson

> >  >>I have given you process reasons why deciding when a country ID
> >  >>or isn't warranted is not scalable.
> So... you have given me "process reasons" why deciding when a country
> ID is or isn't warranted is not able to be expanded to meet future
> needs.
> I'm not a programmer, and I don't think like one, and I don't live
> and work in an environment where this kind of vocabulary use makes
> sense or where anyone uses it this way. (I don't know what a
> heuristic is, either.)
> Sorry, Peter, but you've lost me.

As platform developers, we're trying to supply a tag that reflects a
generic* tag for the language of each locale in our system so that, if
users create content or some other language resource using that locale
and need to tag the content (e.g. with xml:lang) then they've got a tag
they can use.

*Generic in the sense that e.g. for the "Arabic (Saudia Arabia)" locale
we'll return the tag "ar-SA" when asked, though depending on the
situation the best choice may be "ar" (when we are asked, we have no way
of knowing let alone evaluating the situation in which it will be used).

Note that this is not using the tag as a locale identifier. There will
be cases in which the same tag is returned for two or more locales since
the differences between the locales are out-of-scope for language

So, with that in mind, consider the exercise John and Tex did, trying to
determine when it did or did not make sense to include a country ID in a
tag. It was far from obvious for a group of people to come up with
answers. Plus, they really were not considering the point I subsequently
made that there may be valid scenarios in which e.g. fr-CI vs. fr-GH are
needed, something that this group could not possibly predict. There is a
limit to which we can predict when country distinctions will or will not
be useful to users.

So, with *that* in mind, put yourself in the place of someone deciding
for each locale in a platform whether the appropriate language tag to
return should or should not include a country ID. Facing a future in
which you may be doing this for hundreds of cases, it would take a huge
amount of resources to investigate such issues, with no guarantee that
there will be one definite answer in any case. I think you can see that
such a process ----- of determining when country IDs should or should
not be included -- cannot work when you expand to hundreds of cases. It
a problem even in single cases of fr-CC to know which CC values may be
needed. (That's what I mean by a process that does not scale.) So, the
solution is to say, we simply include the country ID for all cases.

> >I gave you an explanation on Feb 12 of why something like mn-Mong-MN
> >vs. mn-Mong-CN or fr-CI vs. fr-GH may well be needed by a user AS
> >LANGUAGE TAGS, NOT AS LOCALE IDs. ISO 639-1, ISO 639-2 and RFC 3066
> >all make reference to combining country IDs with language IDs,
> >without limitation on combinations, and they are not discussing
> >locales.
> May well be needed? If I understand the RFC, it requires some level
> of proof that the entity exists,

Where does it say that?

Here's what I find it does say:

   2.2 Language tag sources:

   Tags constructed wholly from the codes that are assigned
   interpretations by this chapter do not need to be registered with
   IANA before use.

   This document leaves the decision on what tags are appropriate or not
   to the registration process described in section 3.

   2.4 Meaning of the language tag:
   The relationship between the tag and the information it relates to is
   defined by the standard describing the context in which it appears.
   Accordingly, this section can only give possible examples of its

   3. IANA registration procedure for language tags

   The procedure given here MUST be used by anyone who wants to use a
   language tag not given an interpretation in chapter 2.2 of this
   document or previously registered with IANA.

   Note: The purpose of the "published description" is intended as an
   aid to people trying to verify whether a language is registered, or
   what language a particular tag refers to.

> and I wasn't aware that we were
> supposed to be registering aliases of tags for entities.

The tags mn-Mong-CN and mn-Mong-MN are not aliases.

Peter Constable

More information about the Ietf-languages mailing list