Language Variant subtags

Doug Ewell doug at
Mon Jul 5 18:13:30 CEST 2010

I agree with everything Philip Newton said about these proposals by 
Peter Scharf.  Some additions:

> Variant subtags that begin with a letter must be between five and 
> eight (inclusive) characters long, so "ve, cl, ep, bu" are not valid 
> variant subtags. (They would be valid language subtags, and indeed ve 
> = Venda.) See RFC 5646 section 2.2.5 subsection 4 paragraph 1 (minimum 
> length 5 for variant subtags beginning with a letter), RFC 5646 
> section 2.1 (maximum length 8 for all subtags).

In the position in which Peter proposed them, they would be valid region 
subtags (though not canonically cased), and indeed three of the four are 
already assigned as such:

sa-Deva-VE = Sanskrit in Devanagari script, as used in Venezuela.
sa-Deva-CL = Sanskrit in Devanagari script, as used in Chile.
sa-Deva-BU = Sanskrit in Devanagari script, as used in Burma (a 
non-preferred synonym for MM = Myanmar).

In fact, 'vedic' and 'epic' and 'buddhist' might work as subtag values 
for three of these proposals.  "Classical" will require a bit more work: 
not only is the word nine letters long, but the term "classical" is very 
open-ended and could apply to a large number of languages, whereas 
variant subtags in the Registry should have the same meaning for any 
language to which they are applied.  This works for "written in IPA" or 
"written in the Unified Turkic Alphabet," but doesn't work for, say, 
"classical Sanskrit" versus "classical Latin."

Doug Ewell | Thornton, Colorado, USA |
RFC 5645, 4645, UTN #14 | ietf-languages @ is dot gd slash 2kf0s ­

More information about the Ietf-languages mailing list