Please add Serbian to the list of supported languages during installation. The locale(s) for Serbian are sr_CS.UTF-8 (default, cyrillic script) sr_CS.UTF-8@Latn (po files not yet available in CVS, still awaiting to be added to ACL) More info on these locales can be found here if needed: http://srpski.org/locale/ The keyboard for sr_CS is "srp" in XKB. The latin locale sr_CS@Latn uses "srp(latin)". The timezone for Serbia and Montenegro is GMT+0100.
Having users select the script that they use is absurd -- most users won't and don't care to know the difference. Given the translation status, I'm okay with adding Serbian, but only one. Which is the one that is going to be preferred by most users and the "best" for them to use?
Cyrillic is the preferred script and main script for Serbian language. However, both are official scripts and historically both are used. You should consider adding both not because Milos or I like it that way, but because whole political situation in last couple of decades had made it such that some people can read only latin script and other only cyrillic ( there is so much to say about having two scripts, but this is really not place for that). I know that this sounds absurd to you, but for us it is necessity for many reasons.
I'm mostly concerned about user confusion. How do they differentiate that they want one vs the other?
The difference is obvious to any Serbian speaking person (i.e. we are all aware of both scripts). Example of two scripts (same sentence) is given below, first cyrillic then latin and then English translation (you must use UTF-8 encoding to see them). Ðво Ñе ÑеÑениÑа пиÑана на ÑÑпÑком. Ovo je reÄenica pisana na srpskom. This sentence is written in Serbian (language). I do not know if my word is worth anything, but there will be no confusion. Only people that will want to use Serbian language (i.e. Serbs, Montenegrians etc.) are aware that we use two scripts and that one looks like English script and other like Russian. It is people of Serbian descent that are born in western countries that would preffer latin (and most likely would only know latin script) while people born in Serbia would use exclusively cyrillic script (in general). Also, regions in Serbia and neighbouring countries that use Serbian language might preffer latin script as well. Well, closest to real life example of this confusion might be Cantonese and Mandarin dialects. They use same script for two dialects. We use two scripts for one language :). If you wonder how to name them, one called Serbian would default to cyrillic, other could be called Serbian(Latin).
Jeremy, as Igor, I also do understand why this is confusing and seems absurd to people not knowing how this situation came about. Igor explained it sufficiently well, but maybe the following reading will help further http://en.wikipedia.org/wiki/Serbian_language http://en.wikipedia.org/wiki/Serbo-Croatian_language Bottom line is, wether we like it or not, that's just how things are on the ground, as with traditional and simplified Chinese for example. We _absolutely_ have to support both scripts as they are equally used by Serbian speakers all over the world. As you can see from the glyphs, the two cannot be confused, and the labels when choosing one can the ones Igor suggested. Before Unicode, supporting latin only was easier so cyrillic was completely neglected (unlike Russians and Bulgarians who worked hard on cyrillic support as that's the only script they use, and I guess our predecessors took the easy way out with latin), but luckily we don't have to do that any more. To support only one (any of the two) will come off very unprofessional and a half-hearted attempt, not to mention fuel already existing political debates between people prefering one or the other (trust me, you don't want to touch that one with a mile long pole), and that is why we do not have a luxury to make a preference. Besides, GNOME supports both, KDE supports both, as do most other projects, even Windoze. There is no reason for Fedora/RHEL not to support both. We have the sr@Latn .po files ready to go (we maintain the translations in cyrillic and simply make the latin ones by means of a script), but we can't commit them to the CVS as our requests on fedora-trans-list have so far been falling on deaf ears: https://www.redhat.com/archives/fedora-trans-list/2005-November/msg00012.html https://www.redhat.com/archives/fedora-trans-list/2005-December/msg00005.html Since I just saw test2 has been pushed back, let's please try to figure out the best way how to handle this? As mentioned, we only need to keep track of the cyrillic translation on the statistics page as the other one is generated automatically and will always have the same status. We just need have both sr and sr@Latn in ACL to be able to commit. Maybe you can check how the GNOME project implemented this duality? Thanks a lot for working with us on this one.
*sigh* I still don't like it, but added them to anaconda's list. As for getting the translation stuff added, if you're not getting a response on fedora-trans, send mail directly to sarahs and bgroh AT redhat DOT com
Just tested installation of fc5test2 in Serbian, it works well for the most part, thanks. A few things need to be polished though: - Language selection: Serbian(Latin) does not appear to be enabled (I guess this is because sr didn't appaer until after the test2 freeze date) - Language selection: string for "Serbian" not translated (I guess not really a problem because anaconda.pot got updated after the test2 translation freeze date, should appear in the next version, same for "Serbian(Latin)" when it gets added to anaconda.pot) - Keyboard selection: Strings for "Serbian" and "Serbian(Latin)" not translated (couldn't find these in anaconda.pot; I guess rhpl.pot is used for this, but these two are not there either, need to be added) - Timezone selection: please set existing "Europe/Belgrade" as the default location - Timezone selection: none of the location strings are translated at all, although system-config-date/timezones translations for sr and sr@Latn exist
Yeah, the po files didn't quite make it in time. I'll add it back now and the strings should show up in the pot files (and then get translated) fine. Also changed the default timezone. For the timezone location strings, can you file a separate bug so that it can get looked at appropriately?