Improving Thai Braille: the Problem of No Word Endings

Inkprint files containing Thai (and Lao) text do not usually have spaces between words. In the braille generated from such a document, the user ends up with the correct braille characters but no word divisions. The result is braille output that causes line divisions to occur at the wrong places. In these two languages, Duxbury DBT needs to be told where the division points are between the words.

In the search for a way to insert word endings into Thai text, we came across software named LaoScript8. It is an add-on for Microsoft Word (or Open Office) to improve functionality for Lao and Thai. It is distributed by Tavultisoft, a company located in Tasmania. There are three purchase levels of LaoScript8, but only the highest is sufficient:

The feature that helps Thai braille production is slated for the next version (as of this writing). Until the official release, an advance copy can be downloaded for those with the Gold Edition from The developer, John Durdin, warns there may be bugs in this build, but claims it has worked well for him.

After obtaining, installing and licensing the LaoScript8 software, when you use Microsoft Word you find a new menu choice: LaoScript8.

When you are finished with a document and are ready to produce braille, you should click on the LaoScript menu. This displays a ribbon in Word. To use the new function, go to the LaoScript Menu, click on Settings, then More Settings, and set the Thai to Thai Word selection option to "None (wrap Thai with ZWSP)".

Next, in the Word document, use LaoScript8 - Thai to Thai (Automatic) to insert zero-width spaces (ZWSP) into the Thai text at word and syllable boundaries. While the current menu setting suggests translation into Lao, this does not occur. The result is a Thai file with ZWSP characters.

To be technical, LaoScript8 inserts Unicode U+200B, the zero-width space, as the end of word marker. A zero-width space between two printable characters makes them appear side by side without a gap, but it guides the inkprint rendering to potentially divide lines at these positions.

DBT looks for these U+200B characters, and divides braille lines based on where they are placed.