Why didn't they switch to using UTF-8 internally in the 3.0 switch instead of sticking with the archaic UCS crap?<p>I'd even give them a pass if their UTF-8 implementation didn't handle combining characters correctly, since noone else seems to bother!