Re: Can we talk details about adding Humboldt-style scanning to the OVC project?

From: Arthur Keller <voting_at_kellers_dot_org>
Date: Thu May 28 2009 - 02:29:51 CDT

At 11:20 PM -0700 5/27/09, Edward Cherlin wrote:
>Is duplication more likely with our system than with the current
>systems for mail-in ballots? Is the concern that ballots printed on
>plain paper are less secure than ballots printed on paper chosen for
>the purpose and provided to voters individually?

In Santa Clara County, California, a voter who received a
Vote-by-Mail ballot may surrender it at the voter's home precinct and
vote a regular in-precinct ballot. With the OVC approach, there may
or may not be something to surrender, and so the ballot may have to
be provisional. If both pre-printed ballot for voter marking and
ballot printer architecture ballots are supported for Vote-by-Mail,
then the processes need to accommodate accordingly.

Current card stock ballots (folded along defined creases) can
withstand high speed scanners, while print-on-a-home-PC folded in
random places on paper of varying quality probably cannot withstand
high speed scanners. With 2/3 of the votes in Santa Clara County
being Vote-by-Mail, hand scanning could be a problem and potentially
compromise the anonymity of the ballots. The potential requirement
to duplicate some ballots to enable scanning is increased,
particularly on inkjet printed ballots that smear from getting wet.

>It is necessary in either case to verify that only one envelope per
>voter is accepted, in accordance with some legal standard. This has to
>be done before opening the envelopes. Then the approved pile gets
>opened and separated in a procedure designed to prevent matching them
>up again, just as we are supposed to do now. All envelopes, duplicate
>ballots, rejected ballots, etc. have to be preserved separately from
>ballots being counted, at least until all possible legal challenges
>and audits are completed.

The problem is the interaction between in-precinct voting and
Vote-by-Mail voting.

>Can we add further security elsewhere in a system for on-line ballot
>selection and printing, without linking the voters to their votes?

They key to counting Vote-by-Mail ballots printed by a home computer
is the envelope. All the legally required information needs to be on
the outside of the envelope, so that the decision to count the ballot
occurs prior to opening the envelope and reviewing the ballot. I
take credit for getting Santa Clara County, California adopting a
similar approach to provisional ballot envelopes. If we can't solve
the problem of proper envelopes, I expect that many Vote-by-Mail
envelopes not supplied by the Registrar of Voters will result in
ballots that cannot legally be counted.

> > I think that this approach can reduce
>> the delays for overseas and military voters, but in a safe manner.
>It deals with the problem of ballots lost or delayed in the mail, or
>never sent, and with addressing errors.

I don't know of a way to deal with ballots not sent, do you?

The problem of addressing errors will hopefully be addressed (pun
intended) in the solution to the problem of ensuring that the
envelope has the legally mandated information (like voter's name,
registration address, signature, date and statement).

> > I do
>> believe all such ballots must be received by the close of the polls in
> > election day, as is the case for any ballot in California.

Some jurisdictions allow certain extension for late arrival of
military votes. There's also potential for abuse in that process.

> > At 10:30 AM -0700 5/25/09, Jim March wrote:
>>> Arthur, what's your thoughts on precinct-scanning (graphical scans
>>> that is) versus taking precinct ballots back to be centrally counted?
>They should be counted both ways.

That's easy to say but hard to implement. You don't want to double
count the ballots. Ballots should be scanned/counted once for the
canvas, but extra times for auditing. And Vote-by-Mail ballots, even
those dropped off at the polls on Election Day, as well as
provisional ballots, are not scanned at the precinct but only
centrally. So I suggest that regular ballots cast at the precinct be
scanned there (including early vote centers), but all other ballots
be scanned centrally.

> >> There's obviously merit both ways. Chain of custody is better
>>> protected doing it at the precinct, at a moderate (sub-$300) hardware
>>> cost and (arguably) a bit more pollworker complexity.
>>> Doing it centrally, we can get ID stamps put on them as they're scanned.
>What happened to the unique ID numbers we were going to assign, bar
>code, and print in each machine?

Unique ID numbers affixed to the ballot prior to scanning are
problematic in terms of voter privacy and anonymity (and increase the
potential for vote selling and coercion). And that does not solve
the need for ballot tracking numbers on hand-marked paper ballots
that preserve voter privacy and anonymity.

I think the current idea is there is no longer an intent to reconcile
the ballots from the Electronic Ballot Printer and those scanned
later. I think that was a useful simplification.

About barcodes: unless there is a way for the voter to ascertain that
he precinct-count optical scanner has correctly interpreted the
voters ballot, having the scanner rely on barcodes that aren't human
readable is problematic. Barcodes could be a vector for nefarious
voting machine activity, and auditing is necessary (but may or may
not be sufficient) to detect such activity. In a recount, does the
barcode or the printed text prevail?

Best regards,

Experienced advisor to leading edge startups and
accomplished expert witness on patent infringement cases.
Arthur M. Keller, Ph.D., 3881 Corina Way, Palo Alto, CA  94303-4507
tel +1(650)424-0202, fax +1(650)424-0424
OVC-discuss mailing list
By sending email to the OVC-discuss  list, you thereby agree to release the content of your posts to the Public Domain--with the exception of copyrighted material quoted according to fair use, including publicly archiving at
= The content of this message, with the exception of any external 
= quotations under fair use, are released to the Public Domain    
Received on Sun May 31 23:17:07 2009

This archive was generated by hypermail 2.1.8 : Sun May 31 2009 - 23:17:07 CDT