New variant subtags for Serbian language

Doug Ewell doug at
Wed Nov 20 18:11:36 CET 2013

Based on John's explanation, I think I understand better now. We are
talking about one or more languages/dialects which are not adequately
represented at present by *any* language tag.

I had become sidetracked by some of the other issues in CE's messages,
such as assigning BCP 47 prefixes on the basis of a Wikipedia
classification scheme, and ill-formed tags like "Chakavian as
sh-hr-chakavsk, Serbian as sh-rs."

I had also confused this issue with Goran's original, simpler requests
for variants for "Serbian with Ekavian/Ijekavian pronunciation," a
comparatively minor difference which seems well accommodated with
variants against 'sr' as he proposed.

Obviously my bit about tagging Serbian content as Serbian, Croatian as
Croatian, etc. doesn't hold for languages/dialects which aren't Serbian,
or Croatian, or Slovenian, or anything else that already has a language
subtag. The problem is how fuzzy this is with a language continuum that
has already been broken out by nation (whether that was a good idea is
left as an exercise).

If none of these languages/dialects is clearly Slovenian or Croatian, in
the sense that the original requests clearly applied to Serbian, then it
does seem sensible to me after all for the additional subtags to have
"sh" as their prefix, instead of a specific language that they're not.
Allowing, say, both "sh" and "hr" would be the worst scenario, an
invitation for people to tag the same content differently.

It looks like Michael's preference (November 17) is to hold off on
Goran's requests until the rest of the problem is solved. Earlier I had
said that some subtags could be approved now and more later, as long as
the model stayed consistent, but it looks like there is uncertainty
about the model.

Doug Ewell | Thornton, CO, USA | @DougEwell ­

More information about the Ietf-languages mailing list