Commonwealth Club event...

From: Joseph Lorenzo Hall <joehall_at_gmail_dot_com>
Date: Mon Oct 11 2004 - 15:42:55 CDT

Please correct me if you were at the event and I got something wrong.

Sorry this comes so late, it's really only notes (it will be broadcast
at some point)... the event had the following in attendance:

DAVID L. DILL, Ph.D., Founder and Board Director, VerifiedVoting.org
DANIEL TOKAJI, Professor of Law, Ohio University; Author, "Equal Vote" Blog
MARC CARREL, Assistant Secretary of State, California
DAN BURK, Registrar of Voters, Washoe County, Nevada
HENRY BRADY, Professor of Public Policy, UC Berkeley
KIM ZETTER, Senior Reporter, Wired News; Moderator

Dan Tokaji mentioned that he comes at this debate from the perspective
of a civil rights lawyer. In that respect, DREs seem to reduce
residual votes.

Dan Burk mentioned that residual votes includes undervotes as well as
overvotes, and DREs *don't allow overvoting* so to compare the two in
terms of "residual votes", isn't that simple. He said that the only
problems he had in the recent Nevada primary (in Washoe County... the
largest county to use the VeriVote printers other than Clark County
(Las Vegas) which only had one VVPAT-enabled DRE per precinct)) were
either the pollworkers being overly nervous and anxious with the new
system, misguided voters who had though that they could walk out of
the polling place with the receipt and people who were very upset at
having to vote on a machine at all.

David Dill mentioned that he saw two problems with the VeriVote
system... that it doesn't allow non-visual verification of what's
printed on the paper (which he said, as an engineer, he could easily
think of a few ways of solving) and that the VVPATs are stored on a
roll of paper in the cast-vote order which could compromise ballot
privacy. (not to mention that they're on flimsy thermal paper. -jlh)

Marc Carrel was asked a question, "How did uncertified code get on
some machines in the first place?" He said, essentially, that the
system had been broken... that it wasn't prepared for electronic
voting. They've put into place measures to address some of the
problems but he emphasized how different elections with these machines
are. He mentioned that in the 2000 election very few localities were
using e-voting and in the 1996 presidential election none of them
were.

Carrel was then asked (my question, sorry for no plug OVC), "In the
decertification order of April 30, the Secretary of State mentioned a
list of conditions that vendors had to meet for their systems to be
used in Nov. How much pushback have you received from the vendors and
[what's the plan]?" Carrel simply said, "It's been tough to get them
to agree... really tough. But they have all agreed and we're
confident that everything will be in place for Nov. 2." I talked to
him after the event and he said that, indeed, all 4 vendors (Sequoia,
Diebold, ES&S and Hart) had all agreed to the 4 conditions in the
SoS's order and that they would soon be issuing a quick RFP/RFB for
the source code analysis work (unfortunately, only certain
pre-qualified firms will be allowed to make proposals/bids).
(However, talking to someone else after Carrel revealed that yes all 4
vendors had agreed to meet the 4 terms of the order but that only one
vendor had agreed to the source code transfer and analysis agreement.)

Henry brady defined residual votes and explained that only 0.5% of the
population intentionally undervotes. He also explained where his
funding comes from (Kim Z. asked him directly to do so) and emphasized
that his funding is independent from his results per agreement with
the funders. He also mentioned a few things like how much the
tabulation part of the machines-in-elections monster was being ignored
(my wording to keep this interesting -jlh) and that DREs tend to do a
very good job of counting votes (I'm still a skeptic... lots of
examples of miscounts -jlh).

David Dill answered a question, "Dr. Dill, is it true that you cannot
for certain tell that the properly qualified and certified software is
running on a given machine?" Dill said yes... even if the voting
application software was bit-for-bit the same as the software binary
qualified and certified, the machine has various layers of software,
the operating system, which is not tested as it is COTS (commercial
off the shelf), is still a large uncertainty and cause a slew of
problems that could even be undetectable.

I can't remember anything else... this will be broadcast at some point
so listen to KALW in SF[1]Tuesdays at 1pm. It may show up here[2].

[1] http://www.kalw.org/listen.html
[2] http://www.commonwealthclub.org/broadcast/index.html

-- 
Joseph Lorenzo Hall
UC Berkeley, SIMS PhD Student
http://pobox.com/~joehall/
blog: http://pobox.com/~joehall/nqb2/
==================================================================
= The content of this message, with the exception of any external 
= quotations under fair use, are released to the Public Domain    
==================================================================
Received on Mon Nov 1 15:28:44 2004

This archive was generated by hypermail 2.1.8 : Mon Nov 01 2004 - 15:28:58 CST