Message-Id: <6.1.2.0.2.20050103042650.031a7980@mail.jefsey.com>
Date: Mon, 03 Jan 2005 05:29:23 +0100
To: iesg@ietf.org
From: "JFC (Jefsey) Morfin" <jefsey@jefsey.com>
In-Reply-To: <PNEHIBAMBMLHDMJDDFLHCEAJJAAA.aphillips@webmethods.com>
References: <200501021348.35516.blilly@erols.com>
	<PNEHIBAMBMLHDMJDDFLHCEAJJAAA.aphillips@webmethods.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed;
	x-avg-checked=avg-ok-58D62E44
Subject: last call - draft-phillips-langtags-08.txt
Precedence: list
Sender: iesg-bounces@ietf.org
Errors-To: iesg-bounces@ietf.org

I devoted extensive time to a debate of the quoted draft on the 
ietf-languages@alvestrand.no mailing list. I adopted an IETF deliverables 
user's point of view. After this review, I support this draft with the 
provision that it is declared:
- a way to improve and replace the RFC 3066 in the existing areas of 
application of RFC 3066
- and that it extension to other areas should be subject to the acceptation 
of all the concerned parties with possible IAB guidance.

The authors and many on the mailing list have fully documented that some 
application (in particular the W3C) needed this update and documented many 
cases. However there are many reasons why this whole issue should be 
reconsidered, towards:

- a more systematic, less subjective and neutral approach of the value of 
the tag informations.

- a clear explanation of the intended applications of the language tagging 
which seem much larger, wider in scope, diverse, complex, calling for 
various authoritative information sources, etc. that what is permitted by 
the documented rigid and too simple language/country code system.

- a consistency of the used country codes with the ccTLD country codes (as 
an application of RFC 1591 rather than a direct parallel application of ISO 
3166 which may lead to two conflicting country grids, what seems hardly 
acceptable for the IANA, but that most of the members of the mailing list 
accept, showing their lack of experience of International network 
intergovernance issues).

- the research of a consensus with other areas of application of language 
tags such as the Internet governance, SMTP, MIME, OPES, DNS, WIPO/WTO, the 
digital convergence with other standards and technologies, works on 
dialects, on cultural groups, etc.. There is no interest of an Internet 
language tagging which would differ from others technologies and cultural, 
political and societal usages. It would delay the Internet development and 
possibly to be rebuild differently if another consensus of usage prevailed.

I do not necessarily call for a WG-languages to be created. I have 
pre-drafted some thinking on the way the current Internet standard process 
can support the documentation of a multilingual Internet. This might be a 
too important issue with deep architectural implications for it to be left 
to a single working group as I initially suggested. It might be better to 
make it a general issue to be considered by every Internet document. In 
both scenarios (a dedicated WG or an IETF culture evolution) several IAB 
guidances seem necessary and should be called upon after serious and open 
debates.

My first conclusions are that a language tag with (for easy recognition and 
simple sorting) 3 letters language code plus 2 letters country code, should 
be extended to include a four letters ISO scripting code and a "kind of 
application" (probably 1 letter) code (web, IANA, IDN, Governmental, IP, 
e-commerce, legal, etc.) and to be added a free format authority field (in 
addition to the current and to be extended full list of sub tags). It is to 
be noted that 2 letters country codes could use a 2+n sequence for regional 
and cities/business areas, the same for 4+n scripting tags, to take into 
account languages tables, 3+n for dialects, etc. To be fully multilingual 
this scheme should also document the way to use ISO numeric values to 
obtain an all-numeric version of the language tagging.

The work I currently carry on Common Reference Centers (centers for 
information risk containment, contexts support (understood as 
"network"locales") lead us to consider universal language tags such as the 
simple "ssss-xx.ggg-yy.cc-zz.a.authority"structure, where the possible 
meanings could be supported.

- "ssss" is an ISO language scripting
- "xx" a legal variation or a police
- "ggg" a language ISO code (ISO 639-1) with "yy" possible dialect extensions.
- "cc" as a ccTLD country code - plus "zz" possible extensions
- "a" a kind of applications
- free format description of the authoritative source of the presented 
information.
The interest of this format is to be order proofed. Sub-tags can be 
presented in any order to be sorted.

JFC Morfin