User:Actor9

I made three major changes in Counting Single Transferable Votes:

1. I re-stated the Process A Process B part to be general enough to encompass Hare, Meek's, and any other STV method. The original description was not general enough.

2. I removed the subjective opinion that transferring surpluses from the most recent transfer to candidate X only, rather than from the whole set of votes for X, was defective and/or unfair. A neutral point of view is as follows: If a surplus of 30 votes from Y gives X a surplus of 20, because X only needed 10 to be elected, then either (a) the 30 who put X over should get to decide who gets their other 20 votes (i.e., help their own next preferences), or (b) the 30 who put X over should join with the 190 who previously voted for X to decide who gets the other 20 votes (i.e., help the next preferences of the whole 220 in proportion to the number of votes they contributed to X). Hare assumes (a) and Meek's assumes (b). Either is a reasonable position to hold.

3. The various descriptions of methods just gave some salient facts about them and were not anywhere near sufficient. I provided a sufficient description of Hare-Clark, roughly as used in Australia, that people could actually write a computer program from the description. It would be tiresome to do so for every STV method, but one fulsome example seemed like a good idea to me. So I chose what is perhaps the simplest reasonable non-random method.

Finally, I apologize for having so many separate changes. I know now that I should have got it all worked out and then posted it in just 2 or 3 steps, not a dozen or so. I won't do that again.