[Localization] Vietnamese input not working fully on the XO

[Clytie Siddall]

To: OLPC localization list
Cc: Marina and Nancie, OLPC Vietnam

Marina sent this:

Begin forwarded message [the numbers refer to the dead key which applies the accent]:

> I checked all letters and there are other letters that do not work:
> 1. So the "hook above" = 6 does not work at all with any letters, not just letter "a".
> 2. 7 [tilde above] does not work with "e".
> 3. 9 [dot under] does not work with any letters.
> 4. I still can not type numbers or dollar sign. "=" Sign shows wiered not readable in any language. 
> Please forward it to the local team. Thanks.

For background:

Vietnamese keyboard layouts use "dead" keys (pressed before or after the vowel) to add accents to vowels: the standard vowels, including "y", and six vowels already bearing accents. 12 base vowels × 6 accent states means Vietnamese has 72 vowels. To use a Vietnamese keyboard layout (including Unicode hex refs):

1 = ă (0xC483)
2 = â (0xC3A2)
3 = ê (0xC3AA)
4 = ô (0xC3B4)
5 = added grave accent (dấu huyền)
6 = added hook above (dấu hỏi)
7 = added tilde above (dấu ngã)
8 = added acute accent (dấu sắc)
9 = added dot under (dấu nặng)
10 = đ (0xE282AB)
) = Đ (0xC490)
= = ₫ (0xE282AB)
[ = ư (0xC6B0)
] = ơ (0xC6A1)

Marina's message is not good news. However, it's not entirely unexpected. :S

Vietnamese is a test case for Unicode support, because of its combined diacritics and its widely scattered range on the Unicode plane. Also, it's an ideographic CJK language roughly translated into a Roman alphabet. If your software supports Vietnamese, you have very robust Unicode support: the real deal.

So I have become accustomed to introducing Vietnamese translations to a piece of software or platform, and then having these sorts of problems. Unfortunately, I can't be of much help in solving them with the XO, because I don't have one. (I use a Mac.) We really need someone testing the language on the platform before it goes out in the field. I'm deeply embarrassed that people can't use it properly when they need to. :(

In the problem described by Marina, it looks like only the acute (8) and grave (5) accents are being applied. This sounds like a Latin-1 (Western European languages) limitation to me. The acute and grave accent are available in the Latin-1 charset and thus many commonly-used but limited fonts. None of the other Vietnamese accents are. I don't know how you're defining your charsets and default Unicode fonts, but for Vietnamese you need UTF-8 and a pan-Unicode font. For a default pan-Unicode font, on Debian we use Dejavu, to which I contributed persistently to get the Vietnamese characters displaying correctly, if not beautifully. You can also use URWVN [2] which is designed for Vietnamese. Both font packages are GPL (freely available).

The number and symbol characters should be accessible with the use of a control key. On the Mac it's "option". I guessed "control" for the XO, but I could well be wrong. How about "alt"? Marina, did you try each of the control keys? If you're getting odd symbols, that may indicate that you're looking at a higher "layer" of the font. You want the second layer, however you access that on the XO. Do you have a Keyboard Viewer on the XO? That kind of tool shows you a keyboard on the screen, then shows different characters depending on what control keys you press.

I'm hoping the accent-application is just a default-font issue. Sayamindu, what can you tell me about this? And does any OLPC person in Australia have the opportunity to ship me a temporary XO so I can help sort things out? (I live in Renmark, in South Australia: if you travel down the Sturt Highway between Adelaide and the Eastern states, you go right past my house.)

I'd like to use this handy space at the end of my email to put in a plug for Project Gutenberg:

http://www.gutenberg.org/wiki/Main_Page

the first and best truly free ebook repository online. If you have access to out-of-copyright books in any language, or can assign copyright for a text, please contribute the text to Project Gutenberg. Marina reminded me today just how much we need more free, translated or native-language ebooks. Free it, scan it, distribute it!

from Clytie 

Vietnamese Free Software Translation Team

[1] http://dejavu.sourceforge.net/wiki/index.php/Main_Page

[2] http://freshmeat.net/projects/urwvn/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/localization/attachments/20100316/7e7d3042/attachment.htm