Picture of the author
Topic :
Thread Created at Invalid date | Started by
Number of Post in this thread: 1Please Sign In to comment on this Thread
Brent_Allsop replied 14 years ago (Apr 11th 2009, 11:06:39 pm)
Hi Jim, Thanks for submitting and joining this new camp. Now we can move our conversation to this public forum for the camp. I've always been interested in the Ranking Runoff voting process. It sounds like this is a two pass algorithm? One must use some type of scoring to determine what the top few choices are (certainly, one's 3rd choice doesn't get as much of a vote in this first pass than a first choice?) Then once the top (how many?) choices are selected, everyone gets one full vote at this level, even if it is not their first choice. It also seems like such a 'runoff' system is only to determine one top choice, where as current implemented algorithms give ranking values to all choices so they can all be compared, rather than just determining the top choice? Also, I'm not sure how this would be much different than what is currently being used and what additional value it would have. The canonizer algorithms are all implemented in PERL. Perhaps you can provide some pseudo code for this Ranking runoff algorithm? For example, currently, after a '$value' is determined from the selected canonizer (i.e. the PhD canonizer gives one vote for PhDs and zero $value for everyone else), it is run through this algorithm: if ($multi_supporter) { return($value / (2 ** ($support_order + 1))); } else { return($value); } where $multi_supporter is a flag indicating more than one camp supported, and $support_order is an indicator if this is your 2nd, 3rd, and so on choice. All the canonizer code has been developed in an open source way. So anyone can enhance the system any way they want. Brent