How about a proportional proxy? If I'm unsure about my proxies, I'll diversify and hedge my bets. I'll put 60% on A, 30% on B, 10% on C.
I think if A doesn't vote, 60% of my vote gets split between A's proxies.
What do you think? Useful model? Unnecessary complication?
Also, I think there is a Depth First traversal that makes this reasonably fast.
(Unified proxy distribution would be searched Breadth First: Have my proxies (in order) voted? have theirs? ...)
Brian Olson http://bolson.org/
---- Election-methods mailing list - see http://electorama.com/em for list info
