Referencing an alias resource file in 3066:bis

Doug Ewell dewell at
Sat Mar 20 09:00:14 CET 2004

Mike Ksar wrote:
(note that I had to get this second-hand from Addison's response,
because Mike's messages are all in HTML and thus get stripped out of the

>   I understand that existing implementations that use alpha3 codes are
> not in conformance with RFC3066 but the alpha3 code is conformant to
> ISO 3166.  I am aware of the issues within the ISO 3166 and its
> stability and that is why the alpha3 codes should be put back in
> RFC3066bis as an additional option per the 2nd draft.

One thing that is worth noting is that ISO 3166 alpha-3 codes have their
own instability.  Witness the recent recoding of Romania from ROM to
ROU, in response to a vaguely explained "request from the government"
that was rumored to have something to do with misinterpreting the
country code ROM on passports as a code denoting the ethnic group Rom

In fact, the ISO 3166/MA stated at the time that if such a request had
been made against the alpha-2 code, it probably would have been denied,
because of the comparatively greater use of alpha-2 codes.

The alpha-3 codes are administered by the same MA as the alpha-2 codes,
using the same policies and taking the same pride in "revising" the
standard over the past 30 years.  The only stability advantage I see in
the alpha-3 codes is that there are 16,458 possible codes instead of
633, so there is less chance of having to recycle a code.  (633 = 26 *
26 minus AA, OO, QM through QZ, XA through XZ, and ZZ; 16,458 = 633 *

-Doug Ewell
 Fullerton, California

More information about the Ietf-languages mailing list