BUNG SANG COMPUTER PROJECT

Reading Vietnamese on the Braille display

1. Vietnamese in the computer
Vietnamese is a six tones monosyllabic language.
The alphabet consists of 33 letters (12 vowels and 21 consonants), with upper and lower case forms.

Table 1. The vowel characters


Three accent symbols are used to represent different sounds of the same root vowel.

Table 2. The consonant characters


One additional symbol is used for consonant no.4.

Table 3. The accent symbols



Table 4. The tones


In writing, the tone is attached to the vowel sound of the syllable.
For example, vowel no.3 will need three elements when a falling tone is applied to it.

The computer tool used to represent the characters of all languages based on phonetic alphabets is a table of 256 cells called ASCII.
People refer to a cell in the table as an ASCII code. So ASCII code 65 is everywhere taken by the vowel "A". In this way the first 128 cells constitute the standard ASCII table used for characters and symbols common to many languages written by using the Roman alphabet.
Some of the first 128 and the others 128 (=extended ASCII) can be used to customize a particular alphabet or a "font" (=set of characters).
In Viet Nam the PC is used with two languages at the same time: English and Vietnamese. The first one for the operating system and for reading and writing in this language, the other for reading and writing in Vietnamese. The most common software is in English but data are entered in Vietnamese by using additional software that converts some keys to generate accents and tones symbols applied to the alphabet letter.
If characters in the ASCII table take 1 byte, we say that the table is a "8 bit ASCII table".
If characters in the ASCII table take 2 bytes, we say that the table is a "2 bytes ASCII table".

In the history of computers in Viet Nam many solutions have been created to code Vietnamese letters using 8 bit ASCII tables. To show Vietnamese letters (with or without accent and/or tone), we must use the extended part of ASCII.
Two popular solutions are TCVN3 and VNI.
The TCVN3 solution maps one Vietnamese letter to one ASCII code.
The VNI solution combines 2 ASCII characters to show one (accented and/or toned) Vietnamese letter, one byte for the root letter and another for the accent and/or tone. Because of art requirement, VNI also uses one-to-one mapping for 12 letters.

2. Vietnamese Braille

Vietnamese Braille Grade 1 uses 6 dots for each letter. The accented or toned letter is represented by 2 Braille characters, the first for the tone and the other for the alphabet letter. As a result, the 5 tones have their dot-combinations. The standard alphabet letters use the same dot-combinations of the International Braille.
Seven accented Vietnamese alphabet letters ( vowel no.2, 3, 5, 8, 9, 11 and consonant no.4 ) have also their own dot-combinations in one Braille cell.

3. Braille Display

The Braille Display can only be used with a 8-bit code table.
It means one 8-bit character in the computer is mapped to one dotted character in the Braille Display cell.
Since we use Jaws as the screen reader, the mapping was defined in a JBT (JAWS Braille Table) file. Each JBT file contains 256 mappings for the 256 characters in a 8-bit code table. Each mapping defines a dot-combination for one 8-bit character. So for every 6 or 8-dot Braille character you can redefine the dot-combination output of an ASCII table.

4. Solution for reading Vietnamese on the Braille Display

Because of the similarity between the VNI font and Vietnamese Braille in showing letters with tones (=2 cells, one for the root letter and the other for accent), we decided to use the VNI font to translate Vietnamese on the Braille Display and created a JBT file to show Vietnamese Braille on the Display, with a few necessary changes in the Braille code dot combinations.

The user can make copies of the available JBT files and change them to define new translation tables.
We made a copy of VISUADE.JBT and saved it as BUNGSANG.JBT, then we made the neccesary changes in BUNGSANG.JBT.
For reading Vietnamese on the Braille Display, it is necessary to set the user defined Braille Table (in our case bungsang) as the default braille table in the Configuration Manager of Jaws.
By simply changing the default table it is possible to read many languages on the Display, and the blind has a tool to study and read Vietnamese and other languages that may not be voice-supported by the screen reader.

Table 5. Differences between VNI and Vietnamese Braille grade 1


In customizing the JBT file and creating the Vietnamese Braille Code for the Braille Display, we made the following changes:

In the Braille Display an upper case letter is shown by dot 7.

Modify the dot-combination of vowel no.8, 11 and consonant no.4, identified by the listed ASCII code in the VNI font, to obtain the same dot-combination used in Vietnamese Braille grade 1.

Define new dot-combinations for combinations between letter i and five tones, and letter y and tone no.5.
The five dot-combinations between letter i and five tones have been defined equal to the dot-combinations of the five tones.
This has never caused any confusion because the tone never stands alone; it is always applied to a vowel.
So, when the dot combination for a tone stands alone, it means an accented (i) letter matching the sequence above.
For the last letter, combined between y and tone no.5, which is rare in Vietnamese, we used all of the dots of 6-dot Braille: 123456.

Define dot-combinations for the VNI tones symbols.
For the five marked tones, we used the same dot-combination of Vietnamese Braille grade 1.
In our Braille Display - Braille code the dot combination for the tone
comes after the vowel.

The two accent symbols are translated in this way:
dot 7 for accent symbol no.1 and dot 8 for accent symbol no.2.

For example:

 

5. Conclusion

With this solution, the Vietnamese blind can read a Braille code which is very similar to their grade 1, because only a few changes were made.
They can also define new dot-combinations for scientific signs and other symbols. This solution takes less than one hour of practice for learning the variations and then the blind will be able to read fluently any Vietnamese VNI source.
The limitation of this solution is that there are less Vietnamese web pages using VNI than TCVN3. However, there are many texts and documents written in VNI in southern Viet Nam.
For the Vietnamese blind it is therefore a simple and innovative solution (although expensive) for accessing information in digital format for many different purposes.

Vu Huy Phuong
Ho Chi Minh, August 2001

 



Back to top | Home