-
Website
http://techliberation.com/ -
Original page
http://techliberation.com/2006/11/29/is-venezuela-ahead-of-us-on-e-voting/ -
Subscribe
All Comments -
Community
-
Top Commenters
-
MikeRT
184 comments · 6 points
-
eee_eff
800 comments · 8 points
-
mwendy
73 comments · 2 points
-
Ryan Radia
176 comments · 5 points
-
Richard Bennett
612 comments · 1 points
-
-
Popular Threads
-
The Ugliness of Privacy Notices
3 days ago · 4 comments
-
Google’s Privacy Dashboard: Another Major Step Forward in User Empowerment & Transparency
3 days ago · 2 comments
-
Open Source is Not the Enemy
4 days ago · 3 comments
-
Broadband as a Human Right (and a short list of other things I am entitled to on your dime)
3 weeks ago · 18 comments
-
“Internet Freedom”: How Statists Corrupt Our Language
1 week ago · 7 comments
-
The Ugliness of Privacy Notices
Imagine that 100 voters submit their votes into a photocopying machine that redistributes votes to each of the 99 other voters. We have 100 people with 100 votes and 100 counts. If all counts are the same, it would seem that the count is accurate.
Magnify 100 to 100,000,000 and the photocopying machine to the Internet...
The trouble is, proprietary solutions are immune to any and all detractors picking holes in their inevitably fundamentally flawed sec-via-obsc e-voting solutions, whereas this, public solution, has every geek in Christendom taking pot shots at it as though it were an alien invader intent on destroying civilisation as we know it.
Let the people count their own votes for Christ's sake.
Votes are stored in discrete encrypted files that use public key encryption in various ways.
People can vote anywhere, at any web terminal, using a piece of free software that creates their vote file. People can vote any number of times - each vote supercedes the previous.
All files are redundantly replicated to as many who want them. This enables any voter to later check at any other station that their vote is correct.
Until a key is released to reveal all votes' selections (but not the voters' human identities), the only counting that can occur is how many votes have been made (including multiple votes).
It is nevertheless possible for a voter in conjunction with the software to generate a pattern that will enable the retrieval of their vote (possibly along with a few other similar votes).
Once the count key is published (which also specifies a time code to invalidate subsequent votes), all votes' selections can be decrypted and counted (with overrides of multiple votes being prioritised appropriately).
Thus all people in possession of all votes can perform a count. But, the identities of the voters remains anonymous - though a voter can still retrieve their vote.
So, you have a pre-count period in which it's up to each voter to convince themselves that their vote has been registered correctly (by checking at multiple sites), and a post-count period in which this checking is superfluous (nothing can be done for those voters convinced everyone has deleted/changed their vote).
And for prevention of vote selling, we also have a pin number that enables a voter to issue a higher priority vote. All pins appear to operate the same, but only the correct pin actually achieves the desired effect.
This system also requires mutual/social human identity accreditation, i.e. we all certify each other as bonafide humans entitled to vote. You could probably start the ball rolling with a part-centralised/part-distributed system, but ideally you'd have to distribute the trusted authority among the people.
So, yes, apparently simple, but technically exotic.
You'd only need 1% of the population to continuously challenge and verify the software.