Wading through the recent apportionment debate (and flame war) I didn't get a clear sense of what it is we're trying to optimize with various solutions.
representation = representatives / people so, right now the target representation should be about 435/300000000 (and I intend to get my full .00000145 share of my congresswoman!) Possible Apportionment Goals, as concise as I can make them: least over-representation - the solution shall have the least possible per-capita representation in the highest representation state. No state shall be the one everyone is jealous of. least under-representation - the solution shall have the highest possible per-capita representation in the state with the lowest representation. No state shall be unduly ripped off. minimize representation spread - find the solution with the smallest per-capita representation difference between least and most represented people. minimize standard deviation of per-capita representation - there may be some odd highs and lows, but on average it works out. On any of those I think it would be a valid variation to ignore as outliers states with less than one whole representative worth of population (this should filter out measurement-distorting effects of the one seat minimum). So, maybe this is a philosophical question, which of these is most fair? Brian Olson http://bolson.org/ ---- election-methods mailing list - see http://electorama.com/em for list info
