RE: voter verification on the possible target system

["Paul Tiger - LPBC - Outreach" <LPBC-O@xxxxxxxxxxxxx>]

The Hart E-Slate, which I figure the county will buy by 2006, is essentially
a dumb-terminal. It has brains, but on the order of comparison of my PDA to
my desktop. Hell, I can write C programs on my PDA and there are great apps,
but the human interface sort of sucks for a computer geek. So for all
intents and purposes, the E-Slate is a dumb DRE.
There's a central (server) brain box that holds the ballot styles and
records the votes. When we looked at it last year it could print a
transaction log and some other functions. The idea was tossed out that it
could also print the ballots, but that's not workable from the voter's
standpoint. So the pseudo dumb term would get a printer added on. Probably a
thermal one like the one on the brain box.

Hart's system when shown to the committee had a feature that most of the
other vendors had standardized on. The voter or judge would get from the
brain box a time-stamped slip that told a code number to put in at the DRE.
Within ten minutes they would have to enter that on an E-Slate that would
then get the right ballot style from the brain box and show it on screen.
If the brain box had been programmed to give a receipt, then the voter would
leave with two slips of paper. Beginning and end. You were there and you
voted, and we know what E-Slate you voted on.

With the advent of VVPB, the E-Slate would print an optoscan ready ballot.
(Bob McGrath seems to know quite a bit about the paper from earlier
discussions).
The ballot can have the unique ID printed on it; and on one of the two slips
given to the voter.

Like Nick I'm thinking about the number being unique as created by a random
seed and the computer clock. Though I suggest that it be created on the
E-Slate and not the brain box. However this hardly matters in the case of
securing the ID# from the brain box. Since the E-Slates would all be
electrically connected to the brain box, and any data created on either can
be exchanged through the manipulation of software, a dishonest programmer
could store the ID# just as easily as any other datapoint.

That's a look at what the possible future might hold for BC.

But here's the question that applies to all of these systems - integrated or
stand alone - how to keep the ID# from being stored somewhere? If the
computer generates it, then the computer can store it. In point of fact it
does temporarily store it while moving it from one register to another while
being assembled and then to the printer's buffer.

paul

-----Original Message-----
From: Nicholas Bernstein [mailto:nicholas.bernstein@xxxxxxxxxxxx]
Sent: Tuesday, November 16, 2004 3:49 PM
To: Paul E Condon
Cc: cvv-discuss@xxxxxxxxxxxxxxxxx
Subject: Re: Hand count or open source - voter verification - more...

Yes. I agree. You have to write the ID on the ballot while the voter is
the in polling place.
I was saying that it is possible to have a computer print out a paper
ballot with a unique ID. This can happen either when the user enters the
booth (the ballot can then be filled out by hand), or after the choices
have been make on a computer, or even right before the ballot is dropped
into a ballot box (something like a punch clock but with random times).
[|>]