Following #153 and r982, the calculations are now a lot fairer and the changes are smoother.
But, we should take into account which window is causing most of the damage pixels (and even number of bytes) that clog up the system and penalize this window more than those with fewer or smaller updates.
r1191 does this (as without it, we ended up just calculating wrong)
This will do for now.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/155