[Best and Brightest]

Ok, are you saying that no one could alter the code and/or the data translation? How can anyone know they’re vote was counted as they wished?

[zeugma]

Ok, are you saying that no one could alter the code and/or the data translation? How can anyone know they’re vote was counted as they wished?

This stuff has been pretty thoroughly discussed in the past. Here's the quick run down on a secure, properly designed electronic voting booth:

1. All source code for the system is completely open to anyone who wants to look at it. Binaries on the systems can be validated as good by printing shasums of the binaries with all compiler options publicly disclosed. That way, you could compile the program yourself, and verify that it matches the published shasums.
2. In addition to the electronic tabulation, the voting machine should also print out a computer generated ballot, (i.e., no stray marks or hanging chads). This is critical, as you have to be able to re-run the election without relying absolutely on the electronics. Without offline auditability, the system is a fail.
   
3. This paper ballot may or may not have a digital checksum of the vote hashed with its own serial number. There are different schools of thought on this. It might provide an opportunity to correlate a particular ballot to a particular voter for unscrupulous poll workers, however simple protocols can be put in place to eliminate much opportunity for this. Once the voter is validated as being legitimate, he/she would pull a ticket (which is used to tell the machine you have a new voter) randomly from a stack of them. Only one vote per ticket is allowed.
4. The paper ballot can be validated by the voter before it is deposited in the box with other ballots.

 

There have been a number of papers published which go over all of this. Google Bruce Schneier and electronic voting.