Dear Gérard,

There is an ISO/TC 37 standard on alphabetisation, which reflect the
user's point of view in oredering character strings:

ISO 12199:2000 Alphabetical ordering of multilingual terminological and
lexicographical data represented in the Latin alphabet

I may add that there may be other "common ways" of alphabitic ordering,
such as for telephone directories etc., which do not follow a
"standardized" way.

Best regards
Christian



---------------------------------------------------------------------
Dr. Christian Galinski, Director
Infoterm - International Information Centre for Terminology
Mariahilfer Strasse 123/3, A-1060 Vienna, Austria
T: +43-664-344 6181
[log in to unmask] - http://www.infoterm.info
---------------------------------------------------------------------
Founded in 1971 by UNESCO to promote and organize
co-operation in the field of terminology worldwide
__________________________________________________
THIS E-MAIL HAS BEEN SCANNED FOR ALL KNOWN VIRUSES



-----Original Message-----
From: Lang Gerard [mailto:[log in to unmask]]
Sent: 21. april 2006 08:46
To: Havard Hjulstad (E-mail)
Cc: Lang Gerard
Subject: TR: 2-letter and 3-letter codes



 
 
 Dear Havard,

I send you a copy of my exchange with Michael Everson.

All the best.

Gerard LANG

-----Message d'origine-----
De : Lang Gerard
Envoye : vendredi 21 avril 2006 08:37
A : 'Michael Everson'; Lang Gerard
Cc : ISO 639 RA-JAC
Objet : RE: 2-letter and 3-letter codes


Dear Michael,

Thank you for your message.

 I respect your opinion, but it remains only an opinion (even if, as
pratically everybody, I am pretty sure that only my personnal opinions
are the bests in the world).
 My point about ISO 639 is that the normative clauses, the ISO
639/RA-JAC is entitled to manage with, are in contradiction with your
opinion, and that the ISO 639/RA-JAC does not have the right to "have a
policy" permitting to do something against what is written in clause 4.2
of ISO 639-1, that has been adopted in 2002. So I am obliged to remark
that your opinion (that should not be recent, and so was the same before
2002) has not at all been taken in charge by ISO 639 (otherwise the
point that you support and that is important would have been explicitely
included in the text), and is not normative. So we both are not members
of ISO 639/RA-JAC, but both have opposite opinions about what they
should do concerning my propositions. But my point is that, concerning
this question, the first (and perhaps only) thing they have to do is to
follow what is written in the ISO 639 standard they voted and accepted,
as it is mandatory in the ISO world.

Tres cordialement.

Gerard LANG

N.B.:This exchange gives me the occasion of a question concerning ISO
15924 "Code pour la representation des noms d'ecritures". Do you
consider that the four-letter scripts codes given by ISO 15924 to
represent "scripts", defined as "set of graphic characters used for the
written form of one or more languages", are representing ordered set of
graphic characters or do not consider any privilegied order concerning
the graphic characters included in the considered set ?
:
..

 
-----Message d'origine-----
De : Michael Everson [mailto:[log in to unmask]]
Envoye : jeudi 20 avril 2006 21:35
A : [log in to unmask]
Cc : ISO 639 RA-JAC
Objet : 2-letter and 3-letter codes


Dear Mr Lang,

I am one of the people who, years ago, insisted on an ISO 639 policy
regarding 2- and 3-letter codes for languages.

The reason the rule exists is to insure stability and uniqueness in
the use of these tags on for instance the internet. Tok Pisin, Tetum,
and Palauan have the ADVANTAGE of having only one code to worry about
right now. The official language of my country, Ireland, have to
contend with two possible taggings, ga/gle and en/eng respectively.

I am not a voting member of the ISO 639 JAC, but I support their
decision. "tpi", "tet", and "pau" are unique identifiers for these
languages, and they are not disadvantaged in any way for not also
having 2-letter codes.

In my personal opinion, there is no reason that ANY more 2-letter
codes should be added.
--
Michael Everson * http://www.evertype.com