I've done a preliminary analysis of the Florida results based on data available from
http://election.dos.state.fl.us/index.html and I've not seen a hint of evidence of foul-play. The Republicans just plain brought out more voters. Read below the fold for pictures and details.
I think this suggests that we should let the BlackBoxVoting.org people follow up on this, but as a community Kossians need to move on to why liberalism itself is so weak right now.
Using data avaiable from the Florida Elections website, I calculated the difference between the fraction of votes bush and kerry received per county compared to the fraction of dem or rep registered voters in each county. The sequoia e-vote machines are highlighted in yellow. There is no correlation between bush advantage and e-vote counties. This is not evidence that there was nothing wrong, but it is enough to say that if there were manipulations they were not obvious and without precinct level data a strong enough statistical correlation cannot be made.
Higher numbers indicate that the candidate overperformed relative to his registration numbers. Kerry did best in some of the e-vote counties. Without very solid exit poll data this avenue of spatially matching evote counties with results can go no further.
ES&S Touchscreen
ES&S Touchscreen
Update: Here is a picture of the Diebold Machine distribution with bush vote percentage - republican registration percentage highlighted. Here there are a few outliers, but they do not really suggest voting machine problems. Also, see the plot in comments below for a comparison of machine, voter deviation and county population. It further suggests that the machines perform similarly, or at least not detectably differently.
Diebold GEMS Optical Scan