We'll be stuck with UTF-16 as long as we have Unicode, it would seem. The character set itself has to have a hole in the middle of it for the sole purpose of working around UTF-16 limitations.