Comment Paper is not perfect either (Score 1) 193
As someone working directly on the "Linux" piece of software used in Flanders and part of Brussels, I would like to add a few points to the discussion on electronic-voting (I will not comment on the bug, as I do not work for that company and do not know much about the specifics of the issue). Keep in mind electoral systems vary quite a bit country to country, and what you think you know for your country is not necessarily true for other countries.
It has been said that electronic voting adds nothing compared to paper voting. That is maybe true in some elections, but it is not the case in Belgium. In Belgium, the electoral system is extremely complex. This Sunday, people voting on paper ballots actually had 3 ballots, the largest one of which, in Brussels, was big enough to physically hide behind. Manipulating this piece of paper is very hard, filling it out as well (keep in mind this is being done inside a voting booth, which is relatively small). The reason for this is that voters can select as many candidates as they want in one list among the many lists. It's a logistics nightmare, even just printing the thing.
Due to this complexity, actually counting the ballots, when done on paper, is extremely unreliable. In fact, experiences were made here, where groups of people actually totalized the ballots of an urn. They never found similar results from one group to another, not by a long shot. In "small" elections, a few votes can change who the actual winner is - this would be the case in local elections, such as in 2012 in Belgium, less so this year. On the other hand, the electronic counting of ballots consistently gave the same result, of course. Moreover, it goes MUCH faster to recount. Where paper counts could take up to a few hours, polling place results are instantly available.
Countries which publish results immediately at polling place closing times or prior to a complete count are usually using a mix of exit polls, partial results, etc. This is fine for large elections and when a clear margin is accepted. Is this always the case ? Do the media ever publish results in an anticipated rush for the exclusive report ? Electronic voting gives absolute results, not projections, which are definitive. In tight races, little margin situations, this is actually a huge benefit... when everything works (not the case this Sunday in Belgium, where one of the contractors was unable to totalize some results for a while). In many countries where the results are "known" on election day, the official results are only published a few days later (I don't know about Canada which one slashdotter mentioned however). Electronic voting can give final results on the same day.
I would love to bring up the argument for voters with disability, but for Belgium-specific reasons, this was not implemented here yet. The idea is that electronic voting would allow blind people to have audio voting; there are ways for people who cannot hold a pen or use a touchscreen to vote with large buttons; even some mechanisms to control the voting using a breath-driven controller. This allows a lot of people to vote independently, where paper ballots force breaking the privacy of the vote... But again, not implemented yet here. The touch screen here actually does allow larger fonts to be used than on paper, this helps a lot for people with low vision, so it's still an improvement over paper.
People have brought up transparency and reliability. The system here is very different than the one in the US for instance. Here voters go to a voting machine which prints a paper ballot with a QR code containing their selection of parties / candidates, as well as clear-text of this. The voting machine holds no record of what is voted for. The paper can be scanned by the voter on a separate machine so he can check the QR code actually contains who he thinks he voted for. Once he is satisfied, he goes to the urn, scans the QR code and places the paper ballot in the urn. This final scanning is when the ballot is saved to disk. The actual code of all the software used in the polling place is published on the federal government's website. This code is audited before the election by a college of experts (university professors,...) and an audit cabinet. After the election, random sampling is done by the college of experts, and they recount the ballots with their own independent software, and compare with the official results. They also make sure the contents of the QR codes match the clear-text which is printed.
Now everything is not perfect, and there are many vocal opponents to electronic voting even here in Belgium (yes, despite all the above! :) ). This gives areas in which things can be improved. But to think paper ballot voting is impossible to fraud is a fallacy which needs to be also addressed. Urn stuffing can be detected with evoting, not paper. In paper voting, there are always scandals of removed ballots, ballot boxes, etc. In France: an incumbent was caught with many ballots for his opponents stuffed in his socks; in the Netherlands, a ballot box was found dumped in a canal; the stories go on and on.
A final remark on the complexity of electronic voting: people usually think it is either trivial to do or overly complex. Claims that a voting machine can be coded over a week end on a R-Pi by an enthusiast are not new. A fine Belgian hacker made that same claim (not on an R-Pi, but that he could do it over a weekend). He was given several *weeks* to produce a system, and failed miserably. Access restriction, different voter types, crypto and security, redundancy, fault-tolerance is not hacked over a couple of days. When the software needs to run for millions of voters coming on a single day, many of whom do not know how to use a computer, and where every delay gets you crucified by the press, and you need to account for things like weather (affecting ESD discharges in the smartcard readers), you better have more than a week end to come up with your software.
It has been said that electronic voting adds nothing compared to paper voting. That is maybe true in some elections, but it is not the case in Belgium. In Belgium, the electoral system is extremely complex. This Sunday, people voting on paper ballots actually had 3 ballots, the largest one of which, in Brussels, was big enough to physically hide behind. Manipulating this piece of paper is very hard, filling it out as well (keep in mind this is being done inside a voting booth, which is relatively small). The reason for this is that voters can select as many candidates as they want in one list among the many lists. It's a logistics nightmare, even just printing the thing.
Due to this complexity, actually counting the ballots, when done on paper, is extremely unreliable. In fact, experiences were made here, where groups of people actually totalized the ballots of an urn. They never found similar results from one group to another, not by a long shot. In "small" elections, a few votes can change who the actual winner is - this would be the case in local elections, such as in 2012 in Belgium, less so this year. On the other hand, the electronic counting of ballots consistently gave the same result, of course. Moreover, it goes MUCH faster to recount. Where paper counts could take up to a few hours, polling place results are instantly available.
Countries which publish results immediately at polling place closing times or prior to a complete count are usually using a mix of exit polls, partial results, etc. This is fine for large elections and when a clear margin is accepted. Is this always the case ? Do the media ever publish results in an anticipated rush for the exclusive report ? Electronic voting gives absolute results, not projections, which are definitive. In tight races, little margin situations, this is actually a huge benefit... when everything works (not the case this Sunday in Belgium, where one of the contractors was unable to totalize some results for a while). In many countries where the results are "known" on election day, the official results are only published a few days later (I don't know about Canada which one slashdotter mentioned however). Electronic voting can give final results on the same day.
I would love to bring up the argument for voters with disability, but for Belgium-specific reasons, this was not implemented here yet. The idea is that electronic voting would allow blind people to have audio voting; there are ways for people who cannot hold a pen or use a touchscreen to vote with large buttons; even some mechanisms to control the voting using a breath-driven controller. This allows a lot of people to vote independently, where paper ballots force breaking the privacy of the vote... But again, not implemented yet here. The touch screen here actually does allow larger fonts to be used than on paper, this helps a lot for people with low vision, so it's still an improvement over paper.
People have brought up transparency and reliability. The system here is very different than the one in the US for instance. Here voters go to a voting machine which prints a paper ballot with a QR code containing their selection of parties / candidates, as well as clear-text of this. The voting machine holds no record of what is voted for. The paper can be scanned by the voter on a separate machine so he can check the QR code actually contains who he thinks he voted for. Once he is satisfied, he goes to the urn, scans the QR code and places the paper ballot in the urn. This final scanning is when the ballot is saved to disk. The actual code of all the software used in the polling place is published on the federal government's website. This code is audited before the election by a college of experts (university professors,...) and an audit cabinet. After the election, random sampling is done by the college of experts, and they recount the ballots with their own independent software, and compare with the official results. They also make sure the contents of the QR codes match the clear-text which is printed.
Now everything is not perfect, and there are many vocal opponents to electronic voting even here in Belgium (yes, despite all the above!
A final remark on the complexity of electronic voting: people usually think it is either trivial to do or overly complex. Claims that a voting machine can be coded over a week end on a R-Pi by an enthusiast are not new. A fine Belgian hacker made that same claim (not on an R-Pi, but that he could do it over a weekend). He was given several *weeks* to produce a system, and failed miserably. Access restriction, different voter types, crypto and security, redundancy, fault-tolerance is not hacked over a couple of days. When the software needs to run for millions of voters coming on a single day, many of whom do not know how to use a computer, and where every delay gets you crucified by the press, and you need to account for things like weather (affecting ESD discharges in the smartcard readers), you better have more than a week end to come up with your software.
