On Mon, 26 Nov 2001, John J. Boyer wrote: > Dave, > This all sounds fine to me. Now about the format of the translation table. > Since we want to accomodate many languages, I think it would be best to put > each rule on a line by itself and to use a simple space between the parts of > the rule. As I see it, each rule would consist of a source string, a space, > the destination string, space, and a number that specifies whether the > source string is a whole word, should be translated only if it is in the > middle of a word, etc. If either the source or destination string includes a > space, we can arrange for an escape sequence. This would be necessary no > matter what separation character we used. First, let me tell you that English braille has a _much_ simpler ruleset when it comes to abreviate it than say French. In French the rules are much more convoluted and complex than a simple pattern replacement. Therefore I suggest having a look at already existing braille conversion packages and see if those can be used instead of duplicating work. For one I think nfbtrans might do the job and could be GPL already. Nicolas