And parentheses. And spaces (that may, or may not, be trimmed). And all kind of unicode equivalent characters, that might have to be canonicalized. Why not treat it as a byte buffer anyway.