Anomaly in upcoming registry

Doug Ewell doug at
Mon Jul 6 05:56:39 CEST 2009

CE Whitehead <cewcathar at hotmail dot com> wrote:

> Hi, I am not sure the code should have been 'retired'; that was done I 
> assume since it was a macrolanguage and a three-letter code was 
> preferred??, but I really do not know why it was 'retired'.  However, 
> as far as I can tell, a mixture of three-letter and two-letter codes 
> ([ar], [zh]) are used for macrolanguages too!

This is total speculation, and totally wrong.

As has been stated already on this list, "sh" was deprecated/withdrawn 
from ISO 639-1 back in 2000 because the individual code elements "sr" 
and "hr" overlapped it, and were considered preferable by the RA.  ISO 
639-3 and its concept of macrolanguages had not been created yet.

See for the 
real explanation.  Search about 3/4 of the way down the page for the 
term "Serbo-Croatian."

Two-letter code elements are assigned in ISO 639-1; three-letter code 
elements are assigned in ISO 639-2.  All two-letter code elements 
assigned in 639-1 have a corresponding three-letter code element in 
639-2, except for "sh".  There would be no intrinsic reason to "prefer" 
a three-letter code element over a two-letter one.

Macrolanguages in draft-4646bis and draft-4645bis are taken directly 
from ISO 639-3, which assigns three-letter code elements compatible with 
those in 639-2, but also cross-references the corresponding two-letter 
code elements.  Extended language subtags (extlangs) are taken from a 
small subset of these, plus "sgn".  Each of the languages involved was 
chosen irrespective of whether a two-letter code element exists for that 

Sorry to say it, but with all of these errors in interpretation and 
understanding, it might have been better if your post had stayed 

Doug Ewell  *  Thornton, Colorado, USA  *  RFC 4645  *  UTN #14  ˆ

More information about the Ietf-languages mailing list