> > > > Have you tried the existing charset support found in libc? > > You just need to convert cases? > No I need to know what characters are in a word. > In english a word is made up of [A-Za-z0-9] > things like .,!" are not part of words. > > Of course there is a fine line often in this sort of thing. You may use setlocale(LC_CTYPE,"es_AR"); isalnum('ñ'); It works here...