[LINK] Fwd: vip-l: Electronic votiing

[Stewart Fist]

Craig wrote:
> not exactly. unless they can read the printed ballot (which, of course,
> they cant) they have no way of knowing whether what was printed actually
> matches their vote. they'd have to have a sighted assistant to verify it
> for them.

Good point, but wrong conclusion about sighted assistant.

There's no reason why a scanning machine can't read-back the filled in
ballot form. So if you want even greater security, then have a separate
checking machine that scans and plays back through headphones.

The key point is that you don't ask the computer to perform tabulation
functions; just act as an assistant.

But actually, I think David Rose's solution is along the right lines:

> Headphones and a large "stepped" knob (ie with gradations in turning).
> Turn the knob right to move forward, a candidate is read out by the
> machine.  Left, the previous candidate is read out.  A button is next to
> the knob, when you press it you get a confirmation prompt asking if this
> is who you want, press it again to confirm, or move the knob to choose
> someone else.

Simple soldered electrical selection devices like stepped knobs with audio
links don't have one-millionth the potential for manipulation that we all
know exists with computerised devices.  And they can easily be checked by
individual booth scrutineers before voting begins by running some dummy
ballots under supervision.

In fact, if you combined this with a paper-punch, you'd have easy
readability and checkability (just by feel).  You wouldn't have hanging chad
problems, if the machine was able to read-back the selections before the
paper was stuffed in the ballot box.


I think the solution is pretty obvious.  Use computerised, punch-tape,
punch-card or any other type of devices you can think of as assistants to
aid ballot selection.

But never let these machines do any form of tabulation.

Speed of counting is a trivial issue anyway.