[E-voting] Question on optical scan

[Colm MacCarthaigh]

On Wed, Jan 19, 2005 at 11:20:28AM -0800, Justin Mason wrote:
> yep, that's getting right into punch-card (ie. "butterfly ballot")
> territory; and I can see stickers being iffy too -- accidentally
> scratching off, sticking ballots together, etc.  Too many moving
> parts, so to speak.
> 
> I'd have to cast my vote for the "matrix" camp.

Personally I think simple ballot-printers, and Optical scanners are the
easiest, most practical and most securable solution.

Just some simple machines, with a ballot-paper-looking interface (though
a better UI than the machines already purchased would be ideal), have it
produce a printed copy of your ballot, in a nice-big anonymous very
legible computer-font, put it in a ballot-box when you've made sure it's
correct. Have it torn up if it's not.

The ballot-box storage and transmission as normal. Have the ballot boxes
emptied in front of observers, who can do a tally if they want, get a
rough idea what the answer should be. Feed into optical scanners, put
through a few counting implementations - if they all agree, that's
the result.

It'd take maybe 4 hours to count a constituency and declare a result.
There would still be a physical inspection of ballots. In every
constituency there'd be an impression as to the accuracy. In any
contested results, or where people suspect compromised scanners the
count can be repeated by hand or fed into different scanners.

It's easier for everyone.

-- 
Colm MacCárthaigh                        Public Key: colm+pgp at stdlib.net