Re: [tlug] Somewhat OT- open source software for US voting machines

["Edmund Edgar" <lists@example.com>]

Curt Sampson wrote:
> ...the security depends on the
> entire system, not just the source code.
>
> With voting, verification is the hard problem, and the easiest solution
> is audited manual counts, mainly because they're so transparent and
> understandable to everyone involved.

All of the above is true, but what Edward is suggesting is that you should have a voter-verified paper trail.

So the procedure can be something like:
1) Choose your candidate on the computer
2) The computer prints a ballot paper with your choice on it
3) You check the paper says who you want, walk over to the box and put the paper in it

You can then count the votes either by using the computer's records or by counting the bits of paper.

(Of course, if you decided that the bits of paper were the final word on the matter rather than the computer's records of what it printed, you could say that you're not doing e-voting at all; You're just doing traditional paper voting with a more sophisticated pencil.)

Strikes me that it should be possible to create a better audit trail with two methods of counting the votes - especially if you divide responsibility properly so that the people responsible for managing the computer systems are different from the people responsible for managing the ballot boxes.

Whether or not it's worth going to the trouble of making all these machines and teaching people how to use them I don't know, but a setup  like that doesn't strike me as obviously useless in all cases.

Edmund Edgar