Recently, Peter Duniho posted:

"Jose" wrote in message
[...]
An electronic voting machine whose software OTOH is open, public, and
whose compiling and loading into standard interchangable chips and
media is properly supervised is much more difficult to rig. I would
have more confidence in such a machine.

I would not.
[...]
I do agree that an open source software voting machine is preferable.
But IMHO, the more important aspects are for the voting machine to
provide a paper record of the vote, and for the voting results to be
audited.

Specifically, electronic voting machines ought to spit out a paper
ballot very similar to what is used today. The voter should inspect
the ballot to verify it has recorded their vote accurately. Then,
some small percentage of voting machines should be selected
(randomly, of course) for their output votes to be compared to
manually counted paper ballots from those machines.

I agree with you. Further, the percentage of sampled machines should not
be "small", as in 1 or 2%, but significant, as in at least one machine
from each precinct. The paper proofs should be printed at the same time,
with the voter inspecting both for accuracy, and then give one copy to the
registrar (or designated official). That copy would be used to verify the
electronic tally. The question becomes, what to do if there is a
discrepancy?

It really angers me that such basic and simple methodology is not even
being discussed, much less that Diebold is pawning off an approach that is
completely unverifiable, and that politicians are buying into it.

Neil