Monday, April 27, 2009

A note about Arab & Hebrew translation

Why there is a specific case for those two languages. In fact Second Life Viewer does'nt support correctly those two languages.
Everone can read the information on JIRA http://jira.secondlife.com/browse/VWR-112.
Thus we need to resolve into our translators what Linden didn't fix for now.
All those fixes are temporary and will need to be removed the day this Second Life Viewer will be fixed.

I was the first to work on this trouble long before others did and I release the first translator really support Arab & Hebrew the 28th September 2008 with the 0.13 Release of Q-Translator.
Actually there are ony two translators than can be used to have a talk with an Arab or Hebrew within SL (translating in both way).
Those two translators are Q-Translator and AF Translator (with some restriction for AF).

Anyone one that would like to test can contact me.
This is why on my comparaison table in world i remove 2 languages from Ferd Translator, MH Translator, UTv12 and X-Lang.
There is a major difference between calling the Google Translator API and make it works with Arab/Hebrew within SL.

Now there are many situations you need to care to fully support Arabic & Hebrew.
Let's make a test everyone can reproduce using the last translators version below.

AF: AF Pro 1.5
FT: Ferd Translator 7.5d
MH: MH Translator 3.82
QT: Q-Translator 0.27
XL: X-Lang 6.0

The correct translation of the English sentence "I have 35 flowers in my box" into Arabic is "عبرم يدلب يف روهزلا 35 يدل"

I SPEAK ENGLISH SOMEONE ELSE SPEAK ARAB

1) Me from English into Arabic
With autdetect setup for My Language
My Language setup to French (fr), Foreign Language setup to Arabic (ar)
I say "I have 35 flowers in my box"
My translator should translate "لدي 35 الزهور في بلدي مربع"

AF: لدي 53 الزهور في بلدي مربع (almost correct)
FT: لدي 53 الزهور في بلدي مربع (almost correct)
MH: لدي 35 الزهور في بلدي مربع (correct) *Autodetect is always off for me
QT: لدي 35 الزهور في بلدي مربع (correct)
XT: عبرم يدلب يف روهزلا 35 يدل (gibberish) *Autodetect is always off for me

2) Me from English into Arabic
Without autdetect setup for My Language
My Language setup to English (en), Foreign Language setup to Arabic (ar)
I say "I have 35 flowers in my box"
My translator should translate "لدي 35 الزهور في بلدي مربع"

AF: لدي 53 الزهور في بلدي مربع (almost correct)
FT: لدي 53 الزهور في بلدي مربع (almost correct) *Autodetect is always on
QT: لدي 35 الزهور في بلدي مربع (correct)
MH: لدي 35 الزهور في بلدي مربع (correct)
XT: عبرم يدلب يف روهزلا 35 يدل (gibberish)

3) Others from Arabic into English
With autodetect setup for others
My Language setup to English (en), Foreign Language setup to French (fr)
Someone say into arabic "لدي 35 الزهور في بلدي مربع"
My translator should translate "I have 35 flowers in my box"

AF: My flowers in the box indicates 35 (gibberish)
FT: My flowers in the box indicates 35 (gibberish)
MH: لدي 53 الزهور في بلدي مربع (gibberish) *Autodetect is always off
QT: I have 35 flowers in my box (correct)
XL: My flowers in the box indicates 35 (gibberish)

4) Others from Arabic into English
Without autodetect setup for others
My Language setup to English (en), Foreign Language setup to Arabic (ar)
Someone say into arabic "لدي 35 الزهور في بلدي مربع"
My translator should translate "I have 35 flowers in my box"

AF: I have 53 flowers in my box (almost correct)
FT: My flowers in the box indicates 35 (gibberish) *Autodetect is always on
MH: 35 setacidni xob eht ni srewolf yM (gibberish)
QT: I have 35 flowers in my box (correct)
XT: My flowers in the box indicates 35 (gibberish)

I SPEAK ARAB SOMEONE ELSE SPEAK ENGLISH

5) Me from Arabic into English
With autodetect setup for My Language
My Language setup to French (fr), Foreign Language setup to English (ar)
I Say "لدي 35 الزهور في بلدي مربع"
My translator should translate "I have 35 flowers in my box"

AF: My flowers in the box indicates 35 (gibberish)
FT: My flowers in the box indicates 35 (gibberish)
MH: عبرم يدلب يف روهزلا 35 يدل (useless) *Autodetect for me is always off
QT: I have 35 flowers in my box (correct)
XT: (keep silent) (useless) *Autodetect for me is always off

6) Me from Arabic into English
Without autodetect setup for My Language
My Language setup to Arabic (ar), Foreign Language setup to English (ar)
I Say "لدي 35 الزهور في بلدي مربع"
My translator should tranlate "I have 35 flowers in my box"

AF: I have 53 flowers in my box (almost correct)
FT: My flowers in the box indicates 35 (gibberish)*autodetect for me is always on
MH: My flowers in the box indicates 35 (gibberish)
QT: I have 35 flowers in my box (correct)
XT: My flowers in the box indicates 35 (gibberish)

7) Others from English into Arabic
With autodetect for others
My Language setup to Arabic (ar), Foreign Language setup to French (ar)
Someone say "I have 35 flowers in my box"
My translator should translate "لدي 35 الزهور في بلدي مربع"

QT: لدي 35 الزهور في بلدي مربع(correct)
AF: عبرم يدلب يف روهزلا 35 يد (gibberish)
MH: عبرم يدلب يف روهزلا 35 يد(gibberish)*Autodetect for others is always off
XT: عبرم يدلب يف روهزلا 35 يد(gibberish)
FT: لدي 53 الزهور في بلدي مربع (almost correct)

8) Others from English into Arabic
Without autodetect for others
My Language setup to Arabic (ar), Foreign Language setup to English (en)
Someone say "I have 35 flowers in my box"
My translator should translate "لدي 35 الزهور في بلدي مربع"

QT: لدي 35 الزهور في بلدي مربع(correct)
AF: عبرم يدلب يف روهزلا 35 يد (gibberish)
MH: عبرم يدلب يف روهزلا 35 يد (gibberish)
XT: عبرم يدلب يف روهزلا 35 يد (gibberish)
FT: لدي 53 الزهور في بلدي مربع (almost correct) *Auto detect always on

Conclusion XT don't support Arabic at all
FT, MH supports it only when translating from foreign language into Arabic (thus you cannot having any discussion)
AF can work in both way (you need to desactivate Auto Detect)
Q-Translator is the only translator that support Arab/Hebrew in all conditions.

A note about Taiwanese translation

There is actually no translator within SL supporting Taiwanese.
The words "simplified" and "traditional" refer to the characters used, not the language that's written in them.
Google translates into Mandarin and Mandarin Only.

There are seven major Chinese dialect
Mandarin (850 millions), Wu 90M, Yue (Cantonese) 80M, Min 50M (including Taiwanese), Xiang 35M, Hakka 35M, Gan 25M

The simplified character set is mostly use in Republic of China, Singapore, Malaysia, the traditionnal set in Taiwan, Hong Kong and Macau.
Debat on traditional and simplified Chinese characters

From Wiki pages


About 70% of the people in Taiwan belong to the Hoklo ethnic group and speak both Standard Mandarin (officially recognized by the ROC as the National Language) and Taiwanese Minnan (commonly known as "Taiwanese"; a variant of Min Nan spoken in Fujian province). Standard Mandarin is the primary language of instruction in schools; however, most spoken media is split between Mandarin and Taiwanese. The Hakka, about 15% of the population, have a distinct Hakka dialect. Aboriginal minority groups still speak their native languages, although most also speak Mandarin. English is a common second language, with some large private schools providing English instruction. English is compulsory in students' curriculum once they enter middle school. English as a school subject is also featured on Taiwan's education exams.
...

About Taiwanese Min-nan


Taiwanese Min-nan is a variant of the Min Nan Chinese language spoken by about 70% of the population of Taiwan.
This language is absolutely not supported by Google.
Min is the only branch of Chinese that cannot be directly derived from Middle Chinese.
This may account for the difficulty in finding the appropriate Chinese characters for some Min Nan vocabulary.
This is maybe also part of the reasons why it is almost totally mutually unintelligible with Mandarin or other Chinese dialects.
Mandarin and Min-nan are 62% phonetically similar and 15.1% lexically similar. In comparison, German and English are 60% lexically similar

Everyone interrested into languages can find detailled informations on chinese languages here:
Chinese Language
Taiwanese Min-Nan
Min-Nan

Q-Translator 0.27 supports both "Simplified" & "Traditionnal" characters as target character set for Mandarin.