nepali

appaji


Morsels of life!

Mostly boring, sometimes geeky


Previous Entry Add to Memories Share Next Entry
te_IN translation for Debian Installer
nepali
appaji

My parents found it strange that I spend considerable time in the night doing "some Linux thing", so the other day I explained to them what Free software is and we talked about copyright and licensing. I was very happy that they were able to appreciate it. I also gave a Debian Installer demo.

While there are a bunch of Indian languages supported in the Debian Installer, తెలుగు (Telugu) isn't one of them. They noticed it and motivated me to work on te_IN translation for d-i and promised to help me. So the three of us have been discussing translation strings. I have been committing changes slowly and intend to complete this activity in a few months.

Lack of interest in translation meant that I never gave much thought to it but I see that translating software is rather difficult and it is an activity that would benefit a lot from two or three people doing it together. I found a glossary at swecha.org but there are a bunch of problems

  • Computing terminology is context sensitive and one word doesn't fit all uses.
  • A lot of words in the glossary are rather contrived.
  • Most words can be understood only by the elite of the Telugu speaking world.

I am using these simple guidelines for translation:

  • Use language that is used in the newspapers today. Most people who can read would be used to this and understand it best. e.g password would be రహస్యపదం rather than సంకేతపదము.
  • Do not invent words. If a certain computing term or a word doesn't have an equivalent colloquial Telugu word, use the English word written in Telugu script. Include the English word in paranthesis next to it. e.g. Registering modules... would be మాడ్యూల్స్ (modules) నమోదు చేయబడుతున్నాయి...
  • If a translated Telugu word sounds complex or ambiguous in the translation context, include the corresponding English word in Telugu as well as English in parentheses next to it. e.g Partitioning scheme: would be విభజన (పార్టీషనింగ్ [partitioning]) ప్రణాలిక:

A few resources that I have been using:

  • lekhini - a browser based tool for transliteration based on RTS.
  • en2te and uchicago.edu's Charles Philip Brown Telugu to English dictionaries.
  • J. P. L. Gwynn's Telugu to English dictionary hosted by uchicago.edu

I will have to find a good quality comprehensive Telugu general knowledge book or an official A.P. govt. document to translate ISO 3166 codes etc. The Manorama yearbook would've been great, unfortunately Telugu is not one of the few languages it is published in. I will also bring my copy of Sankaranarayan's dictionary from my next month Hyderabad visit.

If you have suggestions on any of these (pointers to guidelines, resources that are DFSG free and I could copy from etc.) that would make doing this easy for me, please do let me know.

Naidu does good one-off Telugu string translations in Launchpad (so does praveenkumarg, btw), and Naidu has promised to do some of the work as well as review. If all goes well, I will try and create a debian-10n-telugu sometime in the future.


Thanks!

(Anonymous)

2008-10-12 07:05 am (UTC)

Thanks for helping make Debian the truly Universal Operating System!

My pleasure!

hat off

(Anonymous)

2008-10-12 09:41 am (UTC)

What struck me immediately when reading this was your excellent command of Shakespeare's language. I'm not a native speaker myself, but I did notice how you crafted this blog post very carefully. I am confident Telugu support in the Debian installer will be great. As a previous poster already said, Debian becomes a bit more universal. Thank you for that!

Mark

Thank you for the kind words, Mark.

Wow! That's some cool work. Good luck!

sorry, I don't have any helpful tips.

Hi, I would like to help. I've been doing Telugu localizations in BetaWiki (http://translatewiki.net/), Launchpad (https://translations.launchpad.net/+languages/te) and some other places (http://crossroads.koodali.org/2008/09/26/localizing-into-telugu-on-the-web/).I am trying to make it a habit to be in #debian-in, whenever possible.

Couple more resources:
* FUEL project (https://fedorahosted.org/fuel/wiki/fuel-telugu)
* TeluguPadam (http://telugupadam.org/) and its discussion group (http://groups.google.com/group/telugupadam) (probably these are not ready for mainstream yet)

I'll get in touch with you on e-mail and IRC veeven.

Fuel looks promising, and one of your links led me to Google Indic transliteration page that has good quality spell-checking too.

[...]
I am using these simple guidelines for translation:

* Use language that is used in the newspapers today.
* Do not invent words.
* If a translated Telugu word sounds complex or ambiguous in the translation context, include the corresponding English word in Telugu as well as English in parentheses next to it.

[...]

This is an extremely practical approach at localizing the User interface messages. I remember reusing/referencing terminology from automated telephone services, mobile phone user manuals, airport security announcements etc. Reason being, familiar instructions of operation via graphical interface can be replicated and users may find it easier to acclamatise themselves with the new interface. :-)

All the best...

That is a nice idea -- text from announcements and manuals would be very familiar. I'll see if I can find some. Thanks for the tip.

Nice idea!

Btw, just curious why can't you just call password as password? While a common translation exists, I think for some terms sticking with their transliterated scheme would help.

I could do that. But it is important that people understand what a "password" (or some other directly-from-English word) means first. Something like rahasyapadaM is easy to comprehend, however, something like an equivalent of "proxy server" is difficult to. Where it is easy to comprehend, I'll try and stick to the Telugu equivalent.

Like you said, I will stick to the borrowed-from-English word spelled out in Telugu for words like modules, proxy server, kernel etc.

Regarding "password" or similar words, there is one small little test which I have often found helpful. Imagine the word being used in general conversation (eg. wrt - email id password, bank locker's password, STD lock on telephone etc.) by different people (a middle-aged bank employee, a student familiar with user interface messages, a homemaker who uses the word in non-computer related context, the office boy who needs to access some password protected work-area etc.). If clears pictures emerge for each of these situations, it is easy to infer the level of penetration of that word in the common parlance of these people and the probable acceptance of it (or its new localized alternative) on the installer interface.