Jon Hanna jon at spin.ie
Wed Jun 18 12:38:31 CEST 2003

> Michael Everson wrote on 06/17/2003 06:04:31 AM:
> > It needn't be four letters but for Jon's suggestion.
> An observation: if we made it four letters, or if we followed Jon's
> suggestion, then that would have implications for incorporating ISO 15924
> into RFC 3066bis: it we *don't* do it, then a new RFC could potentially
> contain a statement to the effect, "Any alpha-4 subtag is interpreted
> according to ISO 15924." But if we *do* do it, then it will take a little
> more work to come up with wording describing use of ISO 15924 (any ISO
> 15924 tag could be used productively, but not all alpha-4 subtags could be
> interpreted in terms of ISO 15924).

I favour the position of interpreting any alpha-4 subtag as ISO 15924,
despite it going against my initial suggestion.

Two ways of keeping the intention of my initial suggestion would be any
4-character but not alpha-4 subtag is orthography information, or script
information that does not relate to ISO 15924 (akin to the i-* registrations
not relating to ISO 639) Hence we could use en-0oed (ghastly I know).

The equally ghastly en-0ed (m4d l33t!) could allow us to used digits for
such identification, though that rules out the possibility of using digits
any other way (and it seems likely we would want to).

Finally en-Latn-oed would work, if we assume that everything following a
subtag about script and/or orthographic information is also script or
orthographic information. It is arguably more in line with the recent
registrations as well, and while the Latn could be argued to be redundant,
it could just as well be argued to be a valid and potentially useful piece
of hierarchical information.

en-Latn-oed seems the least kludgy as well.

More information about the Ietf-languages mailing list