Unfortunately, it doesn't matter whether there's a good way to do redristicting or not, because the US Constitution says that a state legislature can do it however they want, and left to themselves the party in control of the legislature will avoid at all costs the "best" way to do it (see Texas, where the incumbent Democrat's district in the heart of urban Dallas was divided by the Republican legislature into four districsts, some of which included voters over 300 kilometers removed from the center of the "district" to ensure that suburban white republicans in San Antonio could defeat the favorite of urban Dallas.
The simplest way to fix the "gerrymandering" problem would be to pass a constitutional amendmant that requires congressional districts to be composed entirely of contiguous "census tracts." Census tracts are constructed by subdivisions of zipcodes that are defined in such a way as to optimize the constitutionally-mandated collection of census data. Now, the "gerrymander-ers" use this cencus data to split the opposition's districts in such a way that the opposition can never win, and the Supreme Court uses the same data to determine whether a class of voters is being institutionally dis-enfranchised. A constitutional amendment to require congressional districts to be composed of contiguious whole census tracts would still allow Texas Replucans to stretch a district from Dallas to San Antonio, but it would not allow them to split the Dallas voters up into 4 different districts. > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] > ] On Behalf Of Steven Barney > Sent: Tuesday, February 22, 2005 3:13 PM > To: [EMAIL PROTECTED] > Cc: [email protected] > Subject: Re: [EM] Theoretical Gerrymandering Solution > > Matt?: > > This sounds somewhat promising. Can you cite any sources on > the mathematically unsolvable nature of this problem, or > expand on that a little bit? > > > Thank you very much, > SB > > ----- Original Message -- > > Date: Mon, 21 Feb 2005 21:31:40 -0500 (EST) > From: [EMAIL PROTECTED] > Subject: [EM] Theoretical Gerrymandering Solution > > >If we had to redistrict a rectangular state, such as > Wyoming, it seems to > me >that might be a reasonable possible solution. However, I > am concerned > about >irregular states, such as, say, Wisconsin, which > includes a lot of > squiggly >border lines and even some islands. I don't even > know if this > is a >mathematically solvable problem. What do you think? > > >Steve Barney > > It is not mathematically solvable and it arguably doesn't need to be. > There are optimization heuristics that will find reasonably > "good" results > in a reasonable amount of time. Furthermore, since > re-districting occurs > (or should occur) only once every ten years, there can be a > public contest > open to universities, business, individuals, etc. that gives > a reward to > whoever obtains the smallest sum (ties could be broken by > giving the award > to the contestant who first submitted the smallest sum). > > --END-- > > > ---- > Election-methods mailing list - see http://electorama.com/em > for list info > ---- Election-methods mailing list - see http://electorama.com/em for list info
