> The jury is still out on harder scenarios. The only thing for certain is, we > will need better tools for harder upgrades. The question is, can tools good > enough be made at all? what is the current thinking of which theory is appropriate for modeling upgrade?