Re: TED SELKER: US Election Assistance Commission--Questions

From: Douglas W. Jones <jones_at_cs_dot_uiowa_dot_edu>
Date: Sun May 09 2004 - 13:48:27 CDT

On May 9, 2004, at 12:26 PM, Alan Dechert wrote:

> Doug,
>
>> This has real potential as an alternative to a paper record. I think
>> it should be taken seriously. ....
>
> If so, I think the danger of this system is that you'd train voters in
> such
> a way that they always hear the vote read back correctly and over time
> they
> would stop checking.

It's hard to "stop checking what you hear." Much harder than to
skip over reading something.

Also, in general, training is almost irrelevant for voting. People
will forget almost all trained reponses from one voting session to
the next. That's because voting is something we do once in a blue
moon, infrequent enough that training is almost irrelevant.

> Eventually, few, if any, normally sighted voters would
> take the time to listen.

It's not a matter of taking time to listen. Ted's idea is that
as you push the button for Ogelthorpe, it says "Ogelthorpe for
Drain Commissioner." You don't take time to listen, it reads your
selections to you in real time, with immediate feedback.

In this regard, Chaum's current voter-verified and provable paper
scheme is similar. He wants to have manufactured computer screen
enclosures that include a printer behind optics, so that, as you
select Ogelthorpe, it prints "Drain commission: Ogelthorpe" and
projects it on a one-line projection screen below the display
screen (as close as possible to the display screen and in the
same focal plane). So, as you vote, you get to examine, immediately,
the printed record of your selection, with no time delay between
making the selection and verifying it.

(Both proposals face problems when a voter changes their selection.
In both cases, you simply record all selections as part of that
voter's "ballot image" and have the rule that the final selection
recorded or printed in the ballot image is the final one.)

Note that Chaum's voter-verified proposal includes cryptographic
components that allow the voter, after the fact, to verify that
their selection has been included in the totals. This is really a
separate element of his proposal not the topic of this thread.

> As always, trials are needed. The ballot printing system we are
> proposing
> would most likely find virtually all of the normally sighted voters
> checking
> the paper and this would not diminish over time.

Well, I'll agree that most people would check the top-of-the-ticket
race, but I wonder how frequently people will check the mid-ticket
and bottom-of-the-ticket races. It's the mid ticket races that
probably represent the most corrupt practices in American politics
today. County board, sherriff, offices like that where corruption
has a pretty decent payback and where the jurisdiction, the county,
is vulnerable to fixes you can implement entirely in the local
election office.

> This scheme still has the "trust us" factor. The average person would
> have
> no idea if this really means anything.

I agree, all solutions that involve recording on write-once memory
technology such as CD-R involve computers that we have to trust, but
proof that a recorder does nothing but record is far simpler than
proof that a tabulator tabulates correctly. In this regard, these
solutions do represent genuine progress over DRE.

I happen to agree, though, that paper is better. I do think, however,
that these alternative schemes shouldn't be written off out-of-hand.

                Doug Jones
                jones@cs.uiowa.edu
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Mon May 31 23:17:27 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT