Comment 7 for bug 669102

Braille is a script. There are many /encoding/ used with Braille; different languages (Braille Music, French Braille, English Braille, Japanese Braille, ...), some whole words in those languages, punctuation, aliasing of numerals.

It is not possible to "translate to Braille" simply by aliasing certain Braille characters to existing Latin positions, just as it is not possible to translate Latin French to Latin British English just by aliasing some glyphs.

To store Braille script in a computer, two main mappings of Braille to codepoints are used; an older method ("Braille ASCII", codepoints BA+0020..BA+005F, six-dots, non-linear mapping) and a newer one (Unicode "Braille Patterns", codepoints U+2800..U+28FF, eight-dots, linear mapping). Neither of these have any bearing on the languages written the script (and one must know this, just as in Latin). Since we are a Unicode font, the latter encoding is what we should be targetting.

It is possible to transliterate between various Latin scripts and Braille to /some/ extent, just as some Latin languages share similar phrases ('colour' in English/Français) but to think that it is possible by simple mapping is misguided. For example, capital letters (in some Braille languages) can be written with two Braille symbols consecutively, while in others whole words might be written with a single character based on context. Each of those must remain separate, in order that it can be fed to a Braille display.