The C functions towupper and towlower can only be counted on to transform the English characters. Any other behavior is implementation dependent and influenced by locale, which brings in messy environment variables and portability issues. Txr's behavior should not be influenced by such committee-designed garbage.

We need upper and lower casing functions that simply work for the entire Unicode, across the scripts of all languages in which there is such a concept of case.

(For instance, perhaps Japanese hiragana might be mapped to katakana and vice versa, since katakana serves as a sort of alternate case, much like "block letters" in Latin-derived writing systems.)