[ale] Re: Vote Today - Secret Ballots

[Joseph A Knapka]

Benjamin Scherrey wrote:
> 11/5/2002 5:59:04 PM, Joseph A Knapka <jknapka at earthlink.net> wrote:
> 
> 
>>Benjamin Scherrey wrote:
>>
>>>The problem with these techniques is that they don't guarantee anonymity of the vote/voter.
>>
>>Sure they do. A secure voting protocol can allow a voter
>>to produce "proof" that he voted any way at all - that is,
>>if someone tries to coerce a voter into demonstrating that
>>they voted for some "approved" candidate, the protocol can
>>be designed in such a way that the voter can produce
>>that proof no matter how she actually voted.
> 
> 
> 	I need to review my "Applied Cryptography" then as I failed to see this. Everything I read 
> protected identity but always provided a way for the initiator to retrieve his vote in an irrefutable 
> manner. This does not serve our purposes! :-) I'd be interested in seeing any such proposal for a 
> protocol that clears this up.

Recall that the means of vote verification is just a published
list associating each voter's random ID with their vote. When
the union rep asks you to point out your vote on the published
results, point out one that corresponds to the vote the union
rep wants to see.

Of course, a smart union rep will ask you for your random ID
and look up the vote himself. That possibility can be
addressed by adding a small change to the "Improved Voting
With A Central Facility" protocol in "AC": when the voter
votes, the CTF gives him the ID of a randomly-chosen voter
that voted the other way (assuming a two-state vote); the
voter can then pretend to be that other voter when the
union rep comes around. Since the voter can't use the
second voter ID to discover *who* it was that cast that
vote, secrecy is maintained. Of course, this depends on
there being a large voting population. The CTF would need
to cast a small number of "primer" votes before the actual
election, in order to have a selection of alternate-vote
voter IDs to give out. (Note that it can't give out
purely random numbers, since the union rep is going to
execute the confirmation procedure on the voter ID
you give him.)

One can imagine voting machines with two sets of radio
buttons: "The candidate you want to vote for", and "The
candidate you want me to *tell people* you voted for".

> 	Computationally expensive but well within reasonable limits. The IDs need not be 
> permanent  I don't think and could be generated at the precincts for that 
specific election.

Yep.

Cheers,

-- Joe


---
This message has been sent through the ALE general discussion list.
See http://www.ale.org/mailing-lists.shtml for more info. Problems should be 
sent to listmaster at ale dot org.