[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Serbian locale codes

From: Danilo Segan
Subject: Serbian locale codes
Date: Tue, 03 Jun 2003 18:38:42 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20030210


There are several issues to be resolved in regards to Serbian language
and country code assignment.

First and foremost, some "facts":

- Serbian language is official language in "Serbia and Montenegro",
  written in cyrillic script

- Latin script is allowed where explicitely prescribed (in minority
  regions), but in practice, it's used throughout the region, and
  especially on computers (it's substantially easier to get a working
  "Serbian-Latin" than "Serbian-Cyrillic", or at least it was in the
  past). This was good argument in the past, and it isn't anymore.

- The name of the country has changed from "Yugoslavia" to "Serbia and
  Montenegro" at the start of February 2003; ISO 3166 code has not yet
  been assigned[1].

- There's an ISO standard[2] which prescribes script identifiers for
  bibliographic and other usage: in it, interesting codes are "Latn"
  and "Cyrl"

- Currently, "sr" code is used mainly for latin transliteration of
  Serbian language, and "address@hidden" or "sp" (non-existent ISO 639
  code) is used for cyrillic

There are several issues to be fixed:

- When one sets "sr" as the language, official expectations are to get
  a Serbian cyrillic if it is available

- It's better to use "standardized" script identifiers of "Latn" and
  "Cyrl" instead of free-form "cyrillic" and "latin"

- It appears that the LC_DATE definition in Serbian locales in GNU
  libc is incorrect (it probably just "inherits" some international

My suggestion is that all of these get fixed when ISO publishes a new
country code in 3166. This should not take longer than a month
(see [1]). Some other languages could also profit from consistent
usage of [2].

Some of these is based on discussions led on gnome-i18n list, and the
conclusions and the discussions[3] are given for reference. In short,
it was decided to use "sr" for Serbian in cyrillic script, and
"address@hidden" for Serbian in latin script.

Since Serbian Gnome translation project has currently gone farther
than any other project concerned with Serbian language usage on
computers, it seems appropriate to break the backward compatibility
sooner, rather than later. If one must keep compatibility, I'd suggest
on renaming "sr" to "address@hidden", "address@hidden" to "address@hidden", and
creating appropriate aliases, with the *default* alias for "sr" being
"address@hidden" (this would break compatibility, but would be easily fixable
by those who insist on having it "the old way").

Serbian KDE translation team has also agreed with these choices, though their translation is not yet ready.



[1]ISO news about country code assignment:


[3]Discussion on Gnome-i18n list on these issues

First suggestion on what codes to use, similar to the one already in
use in Glibc (address@hidden and address@hidden):

Some clarifications:

And finally, a mention of ISO standard for script names:

Conclusion reached in private e-mail between Christian Rose and myself:

>>>> As we discussed earlier on the gnome-i18n mailing list, there is a
>>>> Serbian cyrillic translation (code: sr) and Serbian latin (I'm not sure >>>> on the code: you may use the ISO four-letter code for the script "Latn"
>>>> to construct the code, or "latin", so you get either "address@hidden" or
>>>> "address@hidden").
>>>Hmm, haven't really decided yet. Which one do you prefer?
>> Since this seems to be the only language with this situation at the
>> moment, I'd go for "internationally recognized" ISO standard of "Latn",
>> and so "address@hidden".
>Yes, probably the best choice. I updated the w.g.o/i18n page

(my words are prefixed with an even number of >'s)

reply via email to

[Prev in Thread] Current Thread [Next in Thread]