> This WG should focus on issues of AVERAGE length of ACE lables. why? do you have some evidence that this has a significant effect on performance of ANY protocol?