Slashdot Mirror
Florida Election Ballots to be Printed On-Demand
davidwr writes "The St. Petersburg, FL, Times reports that Florida is going back to paper ballots, but with a twist. They are printing the ballots on-demand, right there at the polling booth. This isn't machine-assisted voting where a touch-screen fills in your printed ballot for you. It's just a way to save printing costs and reduce paper waste. 'Without ballot on demand, poll workers at 13 early Hillsborough voting sites would need to stockpile stacks of every possible ballot type. With ballot on demand, poll workers can print out a person's distinct ballot type when he or she arrives to vote.'"
These machines will jam or run out of ink with no geeks around to fix it.
Welcome to good ideas which don't stand up to the reality of 5-6 old people monitoring a station.
I like muppets.
Call in a mathematician and get them to figure out how many ballots should be needed to keep costs to a minimum, assuming you leave open the option of printing more ballots, in case the 5/1/0.01% probability comes back to bite you---whether printing it off with a printer on-site, or keeping a large-scale printer on standby in the event that it looks like you are to run out.
The maths isn't exactly difficult---with sufficient historical data, one learns all that's necessary in high school, at least down my way.
That said, we have compulsory voting down our way (Australia), so it's not really an issue that comes up. For that matter, does the risk of printing ~600 sheets of paper too many matter that much? It shouldn't be a problem.
The more you regulate a company, the worse its products become.
An individual's pompous pronouncements on internet fora should be proportional to that person's ability to use the local language.
Upon completing my liberal education, the real learning began.
Of course the p.o.d. ballots are an accident waiting to happen. The whole point of printing the ballots ahead of time is to ensure to the extent possible *ahead of time*, i.e., with time for corrective action to be taken, that there will be no systemic failures. I am impressed by the spectacular cheapness displayed in the decision to go to a "just-in-time" solution for a system which should not accept delays.
Don't trust anyone under thirty.
"What is wrong with paper in the first place?"
Here in Belgium we have electronic vote for more than ten years. I've seen recently a study comparing paper and electronic machine costs.
I don't remember the figures precisly but it was something like:
The cost per vote on paper 2 US$
The cost per electronic vote 5 US$
I always been extremely suspicious about these electronic voting machine. Especially those running Windows (Desktop PC) with accessible serial ports like those we have here.
The good news is that the government plans to get rid of it (at least for a part of the country) and go back to the much safer (and cheaper) paper.
I don't get. We have had, in theory, the protocols to make cryptographically secure verifiable & anonymous e-voting for years now, and yet it hasn't been implemented.
A bunch of hungover CS undergrads with 24 hours till their deadline, would come up with a better e-voting implementation than the hopelessly naive excuses spewed up by diebold et al.
Only difference is, sometimes the vehicle, which is supposed to bring the ballot papers on demand from paper factory, is also out of order...
Pre-printed ballots have a security code on them (otherwise anyone with a decent photocopier could make 100 of them).
It's theoretically possible to link the ballot number to the person but quite hard.
The worst are postal ballots are 100% traceable, and 0% verifiable. In the UK they forced postal ballots on us for a couple of years (closed the polling stations) - you had to fill in your vote then sign and date the form!! So much for anonymous ballots... (only ref. I can find these days is an old blog: http://postalvoting.blogspot.com/)
The practice was stopped, luckily. It was found people were stealing/buying unused ballot forms and sending them in bunches to influence the vote (the whole husband/wife thing came out.. with no anonymous voting the pressure on one person to vote the same as their spouse was extremely strong).
So you're talking about a replay attack, where someone reuses a challenge (blank ballot) to stuff a box with multiple responses (filled-in ballots). Here, we thwart replay attacks with a session token.
Each polling place has a pair of public and private keys, such as RSA or ECC. Each ballot is printed with a barcode containing a session token. The token includes a code representing the polling place, a ballot serial number, and possibly some other information, along with an encrypted hash of this information. The counter looks for ballots whose decrypted hash does not match the hash of the cleartext and tosses those out as spoiled. Then it looks for ballots with duplicate serial numbers and chooses one randomly from each set.