On Fri, 27 Feb 2004, Radovan Garabik wrote: > Don't forget, that is you want enough generality, you have to allow > 1->N scheme (e.g. one input char can be transliterated as a string > of more characters).
Alexander R. Pruss: > Yes, but then the searching will be significantly slower, or we > will have to have more than one search function. > There is no harm doing 1->1. It just might mean a transliteration > scheme that the user will have to actually learn and that is non-obvious. > (E.g., "q" = "Iu", etc.) Could you be more specific with the example? For instance, if I'm reading something in French, and I want to look for the word student, that could be represented as eleve (assuming you have the same font) e'l`eve (a common transcription) eleve (another common transcription, though less so now) Are you saying that I have to know in advance which one the document uses? So the purpose of the extension is to let me enter "e" instead of having to trick graffiti into taking an "e"? Or are you saying that it will make "eleve", "eleve", and even "Eleve" the same, but it just won't be able to handle "e'l`eve" because of the change in length? -jJ _______________________________________________ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
