Proposed general model for Serbo-Croatian continuum

Doug Ewell doug at
Fri Nov 22 02:59:47 CET 2013

Michael Everson <everson at evertype dot com> wrote:

>> The prefix should be specified as 'sr', or possibly all of 'sr’,
>> 'sr-Latn', 'sr-Cyrl'.
> All would be better.

Disagree. We only need 'sr'.

If you have 'sr' alone, you don't need the additional prefixes with 
script subtags. Prefix fields for variants aren't necessarily the 
complete language tag that precedes the variant. Section 3.1.8 makes a 
point of this:

> Although the 'Prefix' for 'rozaj' is "sl", other subtags might appear
> between them.  For example, the tag "sl-IT-rozaj" (Slovenian, Italy,
> Resian) matches the 'Prefix' "sl".

The use case for script subtags is exactly the same as that shown for 
region subtags.

And you wouldn't want to have only 'sr-Latn' and 'sr-Cyrl', because 
these subtags (which denote varieties in pronunciation) very clearly 
could be used for spoken content, or content where the script is 

The place where you want a Prefix field to include a script subtag is 
where you want to emphasize that the script subtag must be present. I 
believe that was the reason for giving, for example, 'hepburn' a Prefix 
of "ja-Latn" but not "ja" -- to show a strong preference for prefixing 
'hepburn' with "ja-Latn" including the 'Latn' script subtag. That is not 
desirable for 'ekavsk' and 'ijekavsk'.

There is also the case of a variant whose Prefix field includes another 
variant, as, for example, 'biske' with a Prefix of "sl-rozaj". The 
intent here is to discourage taggers from creating "sl-biske", or worse 
yet, "sl-biske-rozaj". This is not the situation here, however.

The only thing "all" would accomplish for these new variant records is 
to add unneeded redundancy.

Doug Ewell | Thornton, CO, USA | @DougEwell ­ 

More information about the Ietf-languages mailing list