[EM] Dual Dropping method and "Preference Approval" ballot ideas

[matt at tidalwave.net]

The following changes (version 2.1) were made:
1) Display of approval percent for each winner.
2) Display of total dropping cost.
3) Recommendation to use SSD and RP separately for multi-winner 
elections. Then DD can be used as the tiebreaker (when SSD and RP 
results are different but their dropping costs are the same).
4) Margin as default DD dropping cost (sum of margin and magnitude 
remains default for SSD and RP).
5) Documentation changes to reflect all of above.

Because of the following helpfull contributions:
1)  Comments and suggestion by Forest Simmons about dark horse 
candidates.
2) Stephane Rouillon's observation that minimizing single election round 
dropping cost doesn't minimize multi-election round dropping cost.
3) Markus Schulze's argument that DD fails several fairness criteria that 
SSD and RP pass and 2).
4) My conclusion that margin as dropping cost will reduce Monotonicity 
violation frequency.
5) 1-4.

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em