voting

[Ed Gerck]

Yeoh Yiu wrote:
> 
> Ed Gerck <egerck at nma.com> writes:
> 
> > The 'second law' also takes precedence: ballots are always secret, only
> > vote totals are known and are known only after the election ends.
> 
> You get totals per nation, per state, per county, per riding,
> per precinct, per polling stion and maybe per ballot box.

The lowest possible totals are per race, per ballot box. The 
'second law' allows you to have such totals -- which are 
the election results for that race in that ballot box. For 
example, if there are two candidates (X and Y) in race A ,
two candidates (Z and W) in race B, and only one vote per 
candidate is allowed in each race, the election results for 
ballot box K might be:

Vote totals for race A in ballot box K:
  Votes for candidate X: 		  5
  Votes for candidate Y:   		 60
  Blank votes: 	         		 50

Vote totals for race B in ballot box K:
  Votes for candidate Z: 		 45
  Votes for candidate W:   		 50
  Blank votes: 	         		 20

Total ballots in ballot box K: 	   	115

Because only the vote totals are known for each race, a 
voter cannot be identified by recognizing a pre-defined, 
unlikely voting pattern in each race of a ballot. This 
exemplifies one reason why we need the 'second law' -- to 
preserve unlinkability between ballots and voters.

> So there's a need to design the system to have more voters
> than ballot boxes to conform to your second law.

No. All you need is that there should be more than one voter
per ballot box. This is a rather trivial requirement to meet.

Cheers,
Ed Gerck

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com