New variant subtags for Serbian language

[Doug Ewell]

Goran Rakic <grakic at devbase dot net> replied to Philip Newton:

>> I wonder whether it might not be better to either omit a language
>> prefix or to provide for additional prefixes such as "hr", "sh",
>> and/or "bs", and to modify the description so that it is not
>> restricted to Serbian only.
>
> I think that prefix is required as two new variants are mutually
> exclusive.
>
> Without a common prefix one could form the sr-ekavn-ijekavn language
> tag that makes no sense.

One can always use variants to form a language tag that makes no sense,
because the Prefix field is a strong suggestion, not a law. If these two
variants are approved, I could write "zh-ekavn" and it would be
perfectly valid, though almost perfectly meaningless and not at all
recommended.

Even observing the Prefix field, a user could still combine multiple
variants that are mutually exclusive. Adding prefixes for languages
other than Serbian doesn't add to this problem in any way.

Section 2.2.5 points out that one should not write "de-1901-1996", which
is very much analogous to "sr-ekavn-ijekavn". Although this is not
recommended, it is still not invalid. Rather than making senseless
combinations invalid, the guiding principle in BCP 47 is "tag content
wisely," as described in detail in Section 4.1.

The opposite problem, strongly *encouraging* the use of variant B only
when paired with variant A, would be solved by setting "xx-A" as a
prefix of B.

Incidentally, these requests need to be separated into two registration
forms, one for each proposed variant.

--
Doug Ewell | Thornton, CO, USA
http://ewellic.org | @DougEwell ­