>I guess the bottom line is that a complete multibyte character >interface is part of the LSB. This is a perfectly fair problem statement, and is something I'm afraid we believed we had (via inputs from the OpenI18N workgroup) and so have not been focusing on until now.