Valencian Language Tag registration request
gerard.lang at insee.fr
Fri Jun 19 08:41:19 CEST 2009
Not exactly so, because the title is "Code for the representation of names of languages", so that:
(i)-the only veritable name of the considered language is the autonym (original name inside ISO 639:1988; vernacular form inside ISO 639-2:1998; indigenous form inside ISO 639-1:2002; and should be reference name inside ISO 639-3: 2007), and this is the object that should whose representation should be coded inside ISO 639.
(ii)-in the case that this autonym can be written inside the considered language by a variant of the Latin alphabet, the representation should be this written language name (eventually simplified by elimination of some diacritics) and the corresponding code element should be built so as to have a "visual association" with this representation.
(iii)-in the case that this autonym cannot be written by a variant of the Latin alphabet, we have to build a representation of this autonym written in a variant of the Latin alphabet, that should preferably be a romanization (i.e.:a transcription or a translitteration of this word within a variant of the latin alphabet, or else a phonetization in the case there is no written form of this autonym in the considered language, further eventually simplified by elimination of some diacritics) but could also sometimes be translation in english or french language giving the english or french form of this language name. And the corresponding code element should be built so as to have a visual association with thie considered representation.
(iv)-so that every coded language name should have only ONE original language name, transformed into only ONE representation serving as basis for the only ONE code element identifying the considered language name.
Considering the particular case of Ainu (China) and Ainu (Japan/Russia), I think I already wrote that it would have been easy to build two distinct representations like "Ainu" and "Aynu" by romanization or phonetization, so as to build two distinct code elements for two distinct language names.
Concerning the language name whose autonym is "english", english is a very correct direct representation so that "en" and "eng" are very correct code elements.
Concerning the language name whose autonym is "français", francais is a very correct quasi-direct representation so that "fr" and "fra" are very correct code elements, when "fre" is not so correct.
In my opinion, this description has been the constant practice of ISO 639, ISO 639-2 and ISO 639-1 until 2002.
After that, maybe ISO 639/RA-JAC (eventually in conformance with BCP 47 mind) changed his minds and registered "languages" (without definition) and no more "language names" (with definition), but as the normative texts of these standards have been left unchanged this practice is not in conformity with the letter of the considered standards
De : John Cowan [mailto:cowan at ccil.org]
Envoyé : jeudi 18 juin 2009 20:20
À : Lang Gérard
Cc : Marion Gunn; ietf-languages at alvestrand.no; vmbenet at gmail.com; ietf-languages-request at alvestrand.no
Objet : Re: Valencian Language Tag registration request
Lang Gérard scripsit:
> And, for me, the answer is clearly given by the title of the ISO 639
> series that is "Code for the representation of NAMES of languages".
Clearly that does not correspond with the practice of ISO 639/RA-JAC.
Otherwise, we'd have a code element for "English", another code element for "Anglais", another for "Englisch", another for "gliban.", and so on through the world's 7000+ languages.
> So that the coded entries of ISO 639 should be NAMES of LANGUAGES
> (ideally the language's autonym, that is the name given by the
> considered named language inside this considered language; so that
> there should ideally be only one language name for every entry of ISO
> 639, because this entry is exactly this name that is being coded).
So you are proposing that because Ainu (Japan) and Ainu (China), two unrelated languages, share a name, they should also share a code?
It's time to put the "names of languages" canard to rest once and for all.
There may be a case that ISO 3166 encodes names of countries, but no case at all for ISO 639.
John Cowan cowan at ccil.org http://www.ccil.org/~cowan
Historians aren't constantly confronted with people who carry on self-confidently about the rule against adultery in the sixth amendment to the Declamation of Independence, as written by Benjamin Hamilton. Computer scientists aren't always having to correct people who make bold assertions about the value of Objectivist Programming, as examplified in the HCNL entities stored in Relaxational Databases. --Mark Liberman
More information about the Ietf-languages