So I started thinking about how to patch these election flaws.
The redistricting is hard. There's a project that is trying to figure out cultural regions by voting. (for instance a large swath of southern california considers itself to be part of Los Angeles). It might be possible to use something like that to group people into common "neighborhoods" and aggregate those for elected official districts. (the really simple things like grids or "at-large" are problematic).
As for election machine software fraud, that was a lot easier to come up with ideas for. Just baning "electronic voting machines" isn't truly enough as questionable entities also make the vote counting machines.
Currently I'm thinking the most tamper resistant system would have the following features
- not be reprogrammable (burn the program into ROM)
- all source and all binaries (used to build the source and running the vote machine must be publicly available (and verified)
- ideally the system would be fairly small and it would be possible to read all the software used to build the system. (stop whining and just write it assembly, how hard is it to read a name and add a number)
- burning the roms must be publicly observed.
- since hacks can be installed into the hardware, all schematics must be available and use only common commerical components, and a sufficiently random selection of vote machines should examined under (at least) electron microscopes for rogue components
- all machines that process votes, once verified should remain in locked glass rooms with the opportunity for both video and local observation
- the transfer of votes from the voting stations to the tallying sites must also be under constant observation
- you might also want to have multiple organizations conduct exit polling as a secondary statistical test
I'm wondering if I should try to talk my local green party county council to start inviting people to form some committees to work out those ideas and write them up as initiatives.