3-letter tags galore (Re: Question on ISO-639:1988)

Jeremy Carroll jjc at hplb.hpl.hp.com
Fri Jun 11 13:13:28 CEST 2004

Harald Tveit Alvestrand wrote:

> [apologies for the lateness of this curmudgeonly response]
> --On 2. juni 2004 07:43 -0400 John Cowan <cowan at ccil.org> wrote:
>> Let me see if I can summarize the new state of affairs:
>> 639-3 provides 3-letter identifiers for languages.
>> 639-5 provides non-overlapping 3-letter identifiers for collections of
>>       languages.
>> 639-2 provides a subset of the codes in 639-3 and 639-5 representing the
>>       "important" languages and collections.
>> 639-1 provides a subset of the codes in 693-2 representing the "very
>>       important" languages (and collections?) and giving them 2-letter
>>       identifiers as well.
> Sigh. If I understand this right, you have a 3-letter namespace, and 
> have 3 different standards allocating names out of that namespace. Who 
> administers the namespace?

I guess I should let someone who knows the ISO process answer, but it 
seems that we are talking one namespace the ISO-639 namespace, which for 
ISO administrative purposes is divided up into IS0-639-X as above. That 
looks to me like a case of 'not our problem'.

>> RFC 3066 successors should allow any 639-3 or 639-5 code except those
>> which have 639-1 equivalents.
> IMHO, if my understanding is correct, RFC 3066 should refer to the 
> standard defining the 3-letter namespace, and ignore the difference 
> between 639-2, 639-3 and 639-5. A code is a code is a code is a code.


More information about the Ietf-languages mailing list