Scottish English

Doug Ewell dewell at
Wed Aug 22 07:23:40 CEST 2007

<Karen underscore Broome at spe dot sony dot com> wrote:

> I share Addison's concerns with using the 3166-2 codes. Many of these 
> will be redundant with the ISO 639-3 codes, though this one is not.

In fact, I've been carrying around a summary of "Why ISO 3166-2 won't 
work with RFC 4646" for some time, ready to be invoked the next time 
this topic came up:

1.  The ISO 3166-2 code list is not freely available, unlike other code
    lists used to create subtags.

2.  The ISO 3166-2 code list is not stable.  ISO 3166/MA can change the
    assignment of code elements at any time to reflect changes in
    subdivisions or names as reported by governments.  This is similar
    to the situation with ISO 3166-1, but country subdivisions and
    coding systems change much more frequently than countries.

3.  ISO 3166-2 code elements may be from 1 to 3 letters and/or digits.
    This puts them into conflict with the RFC 4646 syntax, since they
    could not be distinguished from other types of subtags or from
    singletons used to introduce extension or private-use subtags.

    Examples of countries that have different ISO 3166-2 formats
    (A = alpha, N = numeric):

        A   - Argentina
        AA  - United States
        AAA - Mexico (this format is used when ISO 3166/MA assigns
        the codes)
        N   - Austria
        NN  - Japan
        NNN - Slovenia

    Code elements with mixed letters and digits:

        FR-2A (Corse-du-Sud)
        FR-2B (Haute-Corse)
        GR-A1 (Attiki)

    Code elements consisting of the letter "X" (thus conflicting with
    private-use subtags):

        AR-X (Córdoba)
        EC-X (Cotopaxi)
        SE-X (Gävleborgs län)
        VE-X (Vargas)

(One of my first involvement with this list was when I made a totally 
false, newbie assumption that ISO 3166-2 codes could be tacked onto the 
end of RFC 1766 tags, like "en-US-NY".  John Cowan quickly and 
diplomatically divested me of that idea, and I vowed to learn more about 
how language tags really worked.)

Debbie Garside <debbie at ictmarketing dot co dot uk> wrote:

> On another note, when I spoke to the 3166 MA on this very subject 
> about 18 months ago he assured me that the codes could not be 
> allocated as the trio make up a geopolitical entity in its own right - 
> UK (along with NI).  We then get on to the codes for Guernsey, Jersey 
> etc. and the Falklands.  It would seem that if there is water between 
> the geopolitical/devolved entities they can have their own code and if 
> there isn't they can't.  Strange but true I fear!

But UNSD apparently doesn't have the same policy; they still have 830 
for Channel Islands despite the new codes for Guernsey and Jersey.

> As to Scottish English.  Bad move in my humble opinion.  I can 
> understand Edinburgh Standard and not Glaswegian Standard.  Where is 
> the use in such a code?  Better would be: en-standard_glasgow, 
> en-standard_edinburgh etc.

Unfortunately, that would introduce a great deal much more subjectivity 
than "Scottish."  If there are any "in-between" variations of note, how 
would they be indicated?

Doug Ewell  *  Fullerton, California, USA  *  RFC 4645  *  UTN #14

More information about the Ietf-languages mailing list