Translating / localising Linux to a new language
 

Hi folks,

I may have the opportunity to work with people who can translate linux into a language which is not presently catered for, namely a Central Australian indigenous language. Before European contact, such languages have been transmitted by oral forms only. Their written forms (esp. spelling) can vary according to which "school" or branch of linguistics / anthropology coded the written language (and their own guidelines in doing so), and these variations can look significantly different to each other on paper. E.g. "Yeperenye" and "Ayiparinya" are two different spellings of exactly the same word. So, some questions:

1) Are there examples of other such oral languages which have been translated to a Linux OS? If so, which languages and OS's?
2) How have such issues in spelling conventions been addressed / overcome?
3) If I were to get help in creating such a translation, either from someone who speaks the language as a mother tongue, or from an anthropologist / linguist, and with the purpose of getting the widest use and value from the translation, would it be best to begin translating Debian, with the view that it could be easily transferred to other downstream OS's, or simply translate the OS (say, Ubuntu) which the intended audience are most likely to use? Or maybe begin with just an app (say, Firefox)?
4) More generally, lots of the basic words seem common to menus in both OS's and individual programs (like "open", "save", "close", "maximise", etc). Is there a way to, say, make a translation of a the menu docstring "open" once on the OS, then make the translated docstring automatically available to any program which also happens to use "open"?

Many thanks,
Snapper.

Well it seems to me nobody had such an experience, so I'll jump in with something similar.

1) I have been part of the cp6linux team. cp6 is slightly changed Ubuntu localized to Serbian language. Problem with Serbian language is that by Serbian constitution it had two alphabets. So, any translation had to be done with support for both Cyrillic and Latin alphabet. What made our job easier comparing to yours, is that there are strict rule for transliteration. For instance a->a, ц->c, џ->dž, њ->nj, љ->lj, etc. In order to transcribe this, guys made script that could do it. We translated English to Serbian Cyrillic and script did the other half of the job, more or less automatically. What I may notice is that you don't have such a commodity with your language. By examples you provided I presume there are no strict rules for translation between those two variations.

2) Those two alphabetical versions of Serbian language are among the language options as they were totally different languages.

3) cp6 team started directly with Ubuntu. During localization, you will translate .po files. Far as I know, they are highly portable - once you translate the package it will be reusable no matter which distro this software is running on, so there is no need to begin with some kind of "root" distro. You may aim your goal from the beginning.

4) We couldn't make it automatically in such a manner, because words are not always mentioned in their basic meaning, so in different sentences they could be translated using different words. For instance, "close" might be used for closing(затворити) window, and also to indicate that two objects are close(близу) to each other, which as you may notice, are significantly different words in Serbian. Or even worse - sometimes phonetically and semantically same word, might have different form after translation, depending on the situation in which it is being used.

For more details that could help you, you may contact me when you get to work on that project. I'm not in Belgrade right now. Probably won't be there for a month. When I get back there, I'll try to help you as much as my work let me and, as well, try to connect you with people who had much more responsibility on their back, thus much much more knowledge :)

Kind regards

Hi Hogar,

Thanks for your considered and generous reply! I appreciate getting a better sense of the localisation process, and the experience you and your team have. At this stage, I'm mainly gathering info to find out what's actually involved in organising a translation, but I'll need to do some investigating at this end to see if I can find anyone who'd be interested in putting the effort in to providing a translation. One question that might help me on this front would be to find out how much time/effort is involved in producing a localised version of an OS (say, Ubuntu).

So even though I'm not at a stage yet of proceeding with the plan, I'm very keen to hear more of your experiences and insights...

Many thanks,
Snapper.

Ffuf! Well, unfortunately, this is something we can't really help you about. Problem is, that nobody knows how many people were actually involved in localization(how come that we don't know those numbers is a long and irrelevant story). But let's say that a little army of people have been doing it for months :(

Good luck to each of you :).