Some new effectivity bounds in computational geometry