Hi, I have been trying out rules for Hindi in the phonetic code table, by adding hi_phonet.dat, appropriately modifying the hi.dat file, and remaking the dictionary. Using UTF-8 in this file is OK, is it not? Simple rules seem to work, like Devanagari vowel sign i being equivalent to Devanagari vowel sign ii. However, I am getting mixed results with another simple example, a rule that a consonant sounds similar to the same consonant, plus vowel sign a. Here is just one example that I have added to the file ह हा By adding this rule, I would have expected that any word would be zero edit distance away from another, if they differed only in that one used ह, and the other हा However, I am not seeing that. The way I am testing is by mispelling a word such that it is an edit distance of two away from a known word in the dictionary, assuming that the rule above makes the edit distance of ह and हा zero. I would then expect the correct word to show up in the list of suggestions, and indeed close to the top. However, I am not seeing that. Am I missing something?
Regards, Gora _______________________________________________ Aspell-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/aspell-devel
