I was commenting on this question, and was wondering what would be the best to find similar words to build an index of word suggestion in case of a misspelled word, but all this regardless of the current locale (perhaps using a dictionary to find the word language, then use the proper method to find the metaphone, soundex, etc. value?)
I know that languages differ, and character sets also, but regardless of the language, any word is composed of "sounds" and this can be processed to find other words that "sounds" the same in any other or the same language.
Perhaps some project already exist for this, but since I'm mostly programming in PHP these days, I would like to know how this could be possible with that language. Thanks!
(BTW: I know it's somewhat a complex subject, I've already read some books about phonetics on this already, but am no language expert.)