Ted, I was wondering why you assume both not equal (x,a) and x>a, since the second implies the first? Have you found problematic cases? If so, those should be reported as bugs. -s