Re: "dumb scanners"

From: charlie strauss <cems_at_earthlink_dot_net>
Date: Thu Dec 07 2006 - 13:51:35 CST

-----Original Message-----
>From: Marc Baber <>
>Sent: Dec 7, 2006 1:50 PM
>To: charlie strauss <>
>Cc: EI Legislation List <>,, Open Voting Consortium discussion list <>
>Subject: Re: [OVC-discuss] "dumb scanners"
>Thank you for keeping the two aspects of my proposal separate and for
>taking time to formulate a workable solution that uses the dumb scanner
>idea in a scenario where ballot secrecy is 100% preserved and vote
>selling is eliminated. I think different places will want to make
>different trade-offs, so I'm somewhat flexible about how each part of my
>proposal gets used (or not) in different places, as long as the election
>system as a whole has strong integrity and the scenario you describe is
>one such good plan given the requirements choices you've outlined. Good

The key attribute being preserved in the dumbscanner concept, in case it got lost, was that the machines would not need data input, would not be programmable, and might possibly even use discrete logic,(if CPU's they would not need RAM memory (maybe a few registers)), yet it would be flexible enough to accomodate almost any ballot except ranked preference.

>Personally, I'd be sorry to lose the possibility of ranked voting, but I
>don't see a way to break up ballots into pieces, as you suggest, and
>still keep it. I suppose one could keep the ranking columns together as
>a unit, but that could (especially in Calif Gov. Recall race with 130
>rankable candidates) give voters a way of encoding the tail end of the
>rank choices as a "signature" for vote-buying purposes. Perhaps, if the
>IRV was limited to top-3 or top-5 choices the risk would be acceptable?

This is a digression, but i'd go further and say that a strong argument can be made that it is very healthy to limit ranked preference depth to two or three. It's a long winded discussion, but in a nutshell, you can show two things 1) pathologies that deviate from Arrows axioms of fairness can be expected in most ranked preference methods if the vote is split too evenly among many candidates, 2) that deep ranking implies a study of more candidates than most voters have the time for resulting in noise 3) that if an eleciton is undecided past rank-3 or so that it would be better to have a run-off of the remaining few viable candidates who the voters could study with care.

At the risk of adding complexity, I'd propose the following "fix" to the dumbscanner to incorporate ranked preference voting. One places a tick mark on the ballot for consecutive sets of columns that must be "kept together". In such a case scanner, outputs records race by race, instead of oval by oval. Thus raw races, including ranked preferences, would be published. This weakly compromises the secret ballot since it allows a tiny amount of characteristic encoding to be done, but is much better than publishing raw entire ballots. by Weakly, I mean that the amount of encoding that would be possible (as long as the rankings were not deep) would not allow selling of many votes before the encoding was redundant and thus not personally identifiable, making it less useful to a buyer (not entirely however especially if the threat was coercion versus simply a payoff--still it would remove the wholesale market for remote vote-buying and bring it back to retail-one-on-one coercio!

In any case this "fix" has the virtue that once again it does not imply a ballot description data. It's all fixed invariably in the scanner. The problem with this add on is that stray marks made by voters might trigger the mechanism. However if one were also using an ancillirary overvote protection scanner as well, then this could also watch for stry marks as well. Complications. Still much less complicated than say the three-ballot system proposed by Rivest. And as always the dumber the better.


>Best regards,
>Marc Baber
>charlie strauss wrote:
>> Al large part of our discussion comes down to just this:
>> Should we be dismissive of the secret ballot?
>> You would argue yes while other's like me hold it sacrosanct.
>Fair enough. I don't like being characterizes as "dismissive". I would
>prefer to say that I think 100% verifiability trumps 100% secrecy, but
>that's mainly rhetorical.
>OVC-discuss mailing list

OVC-discuss mailing list
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Sun Dec 31 23:17:08 2006

This archive was generated by hypermail 2.1.8 : Sun Dec 31 2006 - 23:17:16 CST