> Thinking about this more, it might make sense to add this > behavior onto the existing approxMatch stuff. Hehe. That's why I started thinking about it, trying to see why Gerard ~= Gérard was returning false. ('cos Gerard ~= Grard is not true).