] Fully electronic systems do not provide any way that the
] voter can truly verify that the ballot cast corresponds
] to that being recorded, transmitted, or tabulated. 

Lots of links to resources on electronic voting systems on this site if you dig around.

Voting systems are easy to manipulate, whether you are building misleading ballots, stuffing ballot boxes, flyering the inner city with notices about the election that include the wrong date, or simply paying off some friends in hollywood to run against you so that the serious candidates are forced to back down. :)

Its very difficult to build a ballot that cannot be stuffed. Tying things to drivers licenses is not enough. Creating all kinds of fake drivers in a district is really easy to do, and the people in a position to do it also run the election. Answers here are really hard to come by. You could publish a list of every voter's address. People could verify that the number of voters is the same as the number of votes. People could also attempt to audit this data independently. Run, for example, some software that correlates the addresses with physical space using a mapquest database in order to make sure that the people live at real addresses that don't overlap and that the population densities aren't out of whack. This might be a resource for spammers, but then again, so is the phonebook. In many states the voter registration lists are already available to political parties. Its just a matter of putting it online.

As for the votes themselves, all the security experts here argue for paper validated audit trails. Now, versus some questionably designed electronic system, yeah, I can see that being useful, but I don't share their agreement that electronic voting systems are bad, for one reason that I don't think they are considering.

Why not publish all the votes on the Internet? Why is the counting process always something that happens in a back room of a high school by a bunch of "bipartisan" old ladies?

When I vote, I get a random number. I can go on the website later and verify that my random number got tabulated correctly. I can count all the votes on the website using my own software and decide for myself who won the election.

The subject of voter coercion comes up. It always does. This is probably the least common and most difficult way of manipulating an election, and yet people always raise it. So, we have systems that are secure against voter coercion but not secure against ballot stuffing. Sigh...

If you want to protect against that in this system, you need only publish some of the votes immediately for people in the voting booths. If they don't want to tell the person coercing them what their random number was, they can simply hand out another random number that has the vote the coercor is looking for.

So, in sum, I think that electronic voting systems can be much more secure then any existing paper or electronic system, simply by making the whole process very transparent, and opening the market up to people who want to audit the whole thing for fun or profit. Seems like an endless source of college CS term papers that will not only educate students, but will keep our political system honest.

So, I say, don't just open the source for the voting machines, Open the Votes!

(I now find myself in the extremely unsettling position of having just publically disagreed with Bruce Schneier. I think thats the first time I've ever done that. I have to wonder if it means I'm missing something...)

(BTW, Rebecca Mercuri was recently kicked out of a conference on electronic voting systems along with David Chaum! I don't really know much about Mercuri, but if someone like Chaum says your democracy is not secure... well... kicking him out on the street for it is not just childish, its dangerously foolish.)

U: Of course, I post something like this, and then the problem with this scheme is immediately obvious. There is no way for me to do anything about a mistabulated vote. Simple solutions break the coercion protection. I think there is a solution in here, but its 2AM. Will repost if I find an answer...

UU: The answer is, of course, exactly what the experts advocate. You verify with a paper ballot before you submit your vote. Having some to this conclusion by my own means I'm now very much in support of it.

The advantage of publishing the votes online and self-auditing your vote is that it may be less expensive then hand auditing serveral thousand paper ballots, and making all of this information public makes it immune to certain kinds of ballot stuff attacks that may involve collusion from people running the election.

The core problem with it is that it assumes that everyone voting has internet access. Its also a little more complex for users... Shrug... Might be useful for a college election.

Rebecca Mercuri's Statement on Electronic Voting