Preferred-Value for region CS

Doug Ewell dewell at
Wed Sep 27 18:35:00 CEST 2006

I've asked Michael to contact IANA and put a hold on the proposal to 
deprecate region subtag CS without a Preferred-value, because it has 
generated some (unfortunately private) questions.

Stephane Bortzmeyer <bortzmeyer at nic dot fr> wrote:

>> It should be evident that neither "Serbia" nor "Montenegro" is a 
>> suitable replacement for "Serbia and Montenegro."
> It is not evident at all. Serbia-and-Montenegro, for various reasons 
> (location of the capital, military and political strength, etc) took 
> over many of the yugoslav assets such as the top-level domain name 
> ".yu" (still alive, since ".cs" was never delegated by ICANN).
> So, giving this habit, and the relative sizes of Serbia and 
> Montenegro, it is not evident that CS should not have a 
> preferred-value RS. I do not think that ISO 3166 gives any guidance 
> here?

This was analogous to what was done (retroactively) with the subtag SU 
for Soviet Union.  Russia was the dominant country out of the 15 Soviet 
successor states, politically and in terms of population, and notably 
they inherited the Soviet nuclear weapons arsenal.  Despite that, it 
seemed clear that for language identification purposes, saying that a 
language was "as spoken in the Soviet Union" did not seem equivalent to 
"as spoken in Russia" considering the diversity of the area and 
potential for erasing important distinctions.

For example, the tag "ug-SU" for "Uighur as spoken in the Soviet Union" 
would have been reasonable, to distinguish from "ug-AF" or "ug-CN".  But 
it would be inappropriate to codify an assumption that "ug-SU" was 
equivalent to "ug-RU" through a Preferred-Value mapping, because most 
Uighur in the former Soviet Union is spoken in Kazakhstan, not in 

On a much smaller and probably less linguistically important scale, 
there is a subtag "NT" for "Neutral Zone," a small no-man's-land between 
Saudi Arabia and Iraq that was divided more or less equally between the 
two in the early 1990s.  (There's more to the story, but that's the 
idea.)  It would not have made sense for programs or humans to assume 
that old "Neutral Zone" tagged content should be interpreted either as 
Saudi Arabia or as Iraq.

The same danger would exist by mapping "CS" to "RS"; important 
distinctions related to Montenegrin usage might be inappropriately 
assumed to be Serbian.

I agree that ISO 3166 gives no guidance here.  UN M.49 did assign two 
new codes to replace the one code for Serbia and Montenegro, but that is 
not really "guidance" about usage going forward.

Another list member wrote:

> Plus comment of See RS for Serbia or ME for Montenegro?

I actually like this idea, and everyone knows how conservative I am 
about the Comments field.  There will probably be people years from now 
who still think Serbia and Montenegro are one country (just like the one 
I talked to six months ago who thought Czechoslovakia still existed), 
and this sort of comment would help them make the right call.

On the contrary...

> So, multiple Preferred-Value fields?

No, no.  That would be crossing the streams.  Preferred-Value tells the 
program or human, "Instead of tag or subtag A, you should use B 
instead."  It wouldn't make sense to say, "... use B or C instead, it's 
up to you."

The opposite situation might make sense some day, though: two or more 
deprecated subtags with the same Preferred-Value.  "Instead of A or B, 
use C instead."

Remember that Preferred-Value doesn't imply that existing content should 
be retagged.

If anyone else has an opinion on Preferred-Value or Comments fields for 
region CS, or anything related to it, please send them to the list where 
all 120-odd list members can see them and respond.

Doug Ewell  *  Fullerton, California, USA  *  RFC 4645  *  UTN #14 

More information about the Ietf-languages mailing list