Re: core components analysis

[Martin Bryan <mtbryan@sgml.u-net.com>]

Bill

> Though Martin Bryan thinks ISO 639 is incomplete, and instead advocates
> IETF RFC 1736, I think the deficiencies of the language code list have
> been addressed by ISO 639-2 which is an extension of the ISO 639-1 code
> set, and takes care of every oddball and obsolete language out there;
> see http://lcweb.loc.gov/standards/iso639-2/englangn.html.  I think the
> Library of Congress as the RA for ISO 639-2, and a language authority in
> its own right, trumps the IANA. The Dublin Core makes allowance for
> using either technique, though.

That's interesting - I was always told that ISO 639-2 failed to take into
account the one case I specifically mentioned, that of American Indians.
(Just in case you think this is a fanciful requirement, I mention that my
son bought a tape on Black Feet and Sioux tribal songs when we were last in
the US. This had no word of English on it. If I wanted to export such tapes
to the UK to sell to the Scottish clans, who have a very similar musical
tradition, what ISO 639-2 language code would I use to describe its
contents?)

The other difference that I was led to believe that ISO 639-2 does not
address is that of the differences between Canadian French, Mali French and
Official French, which the more flexible RFC 1736 does address.

Incidentally, don't take me too seriously - I'm playing devil's advocate to
make sure people properly think through all the implications, especially
those associated with presuming that ISO 639 equates to Part 1, a commonly
made mistake. (Please note that the reference to ISO 639 in the Core
Components form is specifically dated at 1988, which means it does not
include Part 2, which was not published until 1998.) If we do adopt 639 we
need to be specific that both parts are relevant.

Martin Bryan