W3C home > Mailing lists > Public > www-international@w3.org > October to December 2006

Re: What to do with Gaulish (was Re: rfc4646. Some analysis code)

From: John Cowan <cowan@ccil.org>
Date: Wed, 8 Nov 2006 12:22:58 -0500
To: Chris Lilley <chris@w3.org>
Cc: Frank Ellermann <nobody@xyzzy.claranet.de>, www-international@w3.org
Message-ID: <20061108172258.GA22425@ccil.org>

Chris Lilley scripsit:

> I'm becoming more convinced that Gaulish should have been registered as
> a language by itself without the cel- prefix (I was talked into the
> latter, by Michael Everson as I recall).

At the time, it was not possible to register "gaulish" as a tag; the
alternatives were "i-gaulish" and "cel-gaulish".  The "cel" at least
had the advantage that it told naive interpreters that this was a Celtic

The tag "cel", like many ISO 639-2 code elements, represents a language
collection, namely the Celtic languages.  When used as an IETF language
tag, it effectively means "some unspecified language in the Celtic
family".  It so happens that all the modern Celtic languages have their
own 639-2 code elements, so Gaulish is about the only one left.

> Its a language, not a variant; in the same way that Irish and Welsh are not
> 'variants' and English is not a variant of German.

Which is exactly why ietf-languages almost certainly won't register
"gaulish" as a variant subtag; it would be category abuse.

> I plan to, once I can figure out how to do it. Which is not at all
> clear. Can anyone point to the procedure?

I recommend that you try to get it registered with the ISO 639-2
Registration Authority, which is no longer nearly as constipated as it
used to be.  The ietf-languages list does have a residual authority to
register language subtags, but we wouldn't use it unless the ISO 639
process were clearly broken or had miscarried in some particular way.

To get a registration, you need to be able to claim that there exist at
least fifty documents in or about the language (primary or secondary, in
any medium), and that these fifty documents are held by not more than five
organizations (libraries, archives, museums, whatever).  The second half
of the rule ensures that the documents are not so dispersed that no entity
has a real need to catalog them.  (Of course there is no rule against
there being more than fifty documents altogether!  The simplest and
most common case is to show that there is one organization that has at
least fifty documents.)

Once you are convinced of that, go fill out the online form at
http://www.loc.gov/standards/iso639-2/php/iso639-2form.php .  You should
note that you intend Gaulish to be a macrolanguage encompassing the
draft 639-3 registrations for Transalpine Gaulish and Cisalpine Gaulish,
for circumstances in which there is insufficient evidence to make the
distinction (or whatever).

For information on just what that distinction is, you can contact Anthony
Aristar at Linguist List, as they are the sub-registrars for ancient
languages under ISO 639-3.

In due course there will emerge a proper 3-letter ISO 639-2 code element,
which (after ietf-languages rubber-stamps it) will become valid for use
as a language subtag in an RFC 4646 tag.

John Cowan      cowan@ccil.org      http://www.ccil.org/~cowan
Received on Wednesday, 8 November 2006 17:23:08 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:27 UTC