Re: Secret ballot and the Colorado constitution

[Paul E Condon <pecondon@xxxxxxxxxxxxxxxx>]

Paul T.,

Thanks. I think this is valuable. There are problems, but maybe not
unsolvable technical problems. 

Concerning Linda Salas and ballots in random order. I was an election
judge in 2004 in both primary and general elections. I was a 'supply'
judge (who had the extra job of picking up the polling place
materials, and leading the group) I'm sure no one ever told me to
shuffle ballots for either election. Also, in the primary, the end of
day checks that we were supposed to do, could not have been done
according to the instructions that we had been given if we had
shuffled the ballots. In the general election, the instructions were
fixed, but still we were not told to shuffle. Only people who read the
newspapers carefully would have known about Linda's shuffle, and none
of the people working with me had read that carefully. 

Concerning the problem of ballots being printed at multiple locations
and how to ensure that printed ID is unique: Assume that there are
less than 10000 locations at which printing takes place. Assign a
unique, non-random 4 digit ID to each printing location. For unique ID
on the ballot use a random number and the 4 digits of location
ID. Other schemes can be invented, but this is probably good enough to
show that its a solvable problem. I agree that implementing this
solution is well beyond the capabilities of BCClerk office staff. That
is a problem that must be solved.

Concerning it being unconstitutional, I think it probably is, but if
one pushes that point too much without suggesting ways to really solve
it, one is likely to find that the opponents of verifiable elections
will write an draft amendment that will actually make the situation
worse than the current one. Or a draft amendment that is just so goofy
that no one, including you and I, is willing to vote for it. 
 

On Tue, Feb 21, 2006 at 11:44:16PM -0700, Some Guy wrote:
> Paul (et al),
> 
> Paul C makes a good point. While I was at odds with the unique numbers on
> the ballot at the time of the suit, I had already had enough interaction
> with the Hart system to understand that it wouldn't be able to operated
> without them. And in retrospect, myself and the other ERC members quickly
> concluded that we would have had a difficult to impossible time trying to
> resolve the errors without them.
> 
> Here's some review that might help people understand how unique numbers on a
> ballot could be problematic if created on-the-fly.
> During *early voting* the ballots were printed for the voter at the time
> they showed up. Each location for early voting was given a range of numbers
> to serialize the ballots. In this way it was not likely that the Louisville
> location would print the same numbers as the Longmont location.
> If randomization was done, then one of two things would have to occur.
> Either the 'seed' number or formulae used at the various locations would
> have to create random numbers in a range OR the computers creating the
> numbers would have to be networked to be sure that no two numbers were the
> same.
> I don't know enough about the mathematics to perform the first idea, but I'm
> sure that the byte geeks could figure it out. The second method would bring
> about a flurry of activists who think that networking any part of the
> elections systems would be an issue. But a real problem for Boulder County
> is that IT support for elections is REALLY BAD. I have serious doubts that
> BC's IT group could keep a connection solid for more than a few hours at a
> time (sometimes minutes at a time). Believe me, I have had insider
> experience with BC IT and elections. [ in 03 while posting elections results
> the DHCP server renewed my IP lease at 9pm (two hours into tally). The
> webserver no longer recognized my desktop as the official poster of tallies,
> and an emergency page to BC IT was met with a 3 hour response time. ]
> 
> As for pre-printed ballots used on election day or for absentee voting -
> they were all sequentially serialize when printed. These could have been
> randomized.
> Clerk Salas instructed polling place judges to shuffle the ballots and hand
> them out out-of-order. Many of us know that this did not happen. I heard
> from a number of CVVers that it wasn't happening at their polls, and found
> that to be true in my precinct. I was able to see that the person in front
> of me and behind me in line had ballots with numbers one above and one below
> mine. And of course I heard their names called out, so I knew who had what
> ballot number. The place was a zoo and I could have hung out for awhile
> keeping notes.
> Randomization for pre-printed ballots is doable.
> 
> So in part my response is about computer technology, but really it is about
> people and BAD support.
> All this aside, I still find unique marks on ballots to be unconstitutional.
> Yes, there could be an amendment, but there hasn't been one, so we have to
> follow the laws that we have until they are changed.
> 
> Paul Tiger
> 
> -----Original Message-----
> From: Paul E Condon [mailto:pecondon@xxxxxxxxxxxxxxxx]
> Sent: Tuesday, February 21, 2006 11:14 AM
> To: cvv-discuss@xxxxxxxxxxxxxxxxx
> Subject: Re: Secret ballot and the Colorado constitution
> 
> Paul,
> 
> I have a somewhat detailed issue with your opinion about
> serialization.  I can see how it is used in tracking data records,
> such as images, through a computer system, so I appreciate why it is
> desired by implementors of computer-based vote counting systems. BUT
> wouldn't that use be just as well served by having a unique number on
> each ballot? Unique number is something like a random 256 bit public
> encryption key. If they are put on the ballots in the order that they
> are generated, they are not *sequence* numbers, and can't be used to
> put ballots in the order in which they were cast, but they can be used
> to check if a ballot has already been scanned into the system, and if
> a ballot is a forgery. So, my question: could random unique
> identifiers satisfy the techy system developers? And, just as
> important, would random unique identifiers satify your concerns about
> voter privacy? Please, no flip answer.  Think about it.
> 
> Others, please also express opinions on this issue, but if your beef is
> with computers per se, don't pretend it is a response to this post.
> 
> Of course, they may very well not satisfy the letter of the state
> constitution, but state constitutions have been amended in the past,
> often for trivial or ill considered reasons.
> 
> 
> 

-- 
Paul E Condon           
pecondon@xxxxxxxxxxxxxxxx