5 ms·

I am a red-blooded technologist, but I think voting should be done on paper ballots. Call me a luddite, but it's just too easy to manipulate votes--either at th

I am a red-blooded technologist, but I think voting should be done on paper ballots. Call me a luddite, but it's just too easy to manipulate votes--either at the time of voting or in post-processing--with an electronic voting machine. That being said, the only conceivable way to have a secure electronic voting process is to use a completely open source system. Open source hardware and software, with publicly viewable results.

cpeterso 14y ago
What problem do electronic voting machines actually solve? The US didn't seem to have problems counting millions of paper ballots (save for some hanging chads) until the "solution" of electronic voting machines appeared.
- michaelt 14y ago
  As I understand it the intention is to solve: 1. Allowing people with disabilities to vote. For example, headphone voice prompts for blind people, large print/high contrast for the elderly etc. 2. Making sure voters understand the voting procedure - particularly if there are multiple offices and issues on the ballot, multiple ballot papers and so on. 3. Ensuring voters' intent is recorded unambiguously. For example, you can't put marks in multiple boxes (intentionally or accidentally) and you can't leave a dimpled or hanging chad. 4. Perfect counting (assuming the data is entered correctly and the software works right) as opposed to manual counts where papers sometimes end up in the wrong pile. 5. You can offer voting in multiple languages. 6. You can randomise the order of candidates on the ballot, to evenly distribute any impact being at the top might have. 7. Ability to take backups for transit or deliver results electronically, avoiding boxes of ballots going missing. 8. Reduced counting costs and faster results, allowing more polling stations to be operated and allowing the polls to stay open later for the same result delivery deadline. Whether it succeeds at offering these benefits, and whether they're worth the obvious cost in trust is debatable.
danielweber 14y ago
If your plan to secure the machines is to have the source be open, you have already failed. You need a system that works well even if the source has been altered. Paper ballots have issues, too. They are different issues, and largely unseen by the general public until you have a really close election and have to use human judgment to decide whether that mark counts as a vote or not.
- revelation 14y ago
  And here I thought in the Turing year of 2012 people would finally start to understand what it means that any electronical voting machine contains a Turing equivalent machine. No magic can make the halting problem decidable.
  - EvilTerran 14y ago
    An electronic voting machine doesn't have to be Turing-complete, surely? It would require especially-fabricated non-Turing-complete chips / ICBs, and some very simple interface like an LED bank and good old-fashioned physical buttons (as a touch-screen LCD unit alone is probably turing-complete), but I'm pretty sure it'd be possible to devise a specialised electronic voting system that couldn't have its behaviour altered without physical modification. [edit PS] Of course, this is purely academic. You might as well fantasise about a voting system made of Babbage-esque clockwork.
- Peaker 14y ago
  In really close elections, the outcome matters much less than the potential for systematic rigging of non-close elections.

cpeterso 14y ago

What problem do electronic voting machines actually solve? The US didn't seem to have problems counting millions of paper ballots (save for some hanging chads) until the "solution" of electronic voting machines appeared.

michaelt 14y ago
As I understand it the intention is to solve: 1. Allowing people with disabilities to vote. For example, headphone voice prompts for blind people, large print/high contrast for the elderly etc. 2. Making sure voters understand the voting procedure - particularly if there are multiple offices and issues on the ballot, multiple ballot papers and so on. 3. Ensuring voters' intent is recorded unambiguously. For example, you can't put marks in multiple boxes (intentionally or accidentally) and you can't leave a dimpled or hanging chad. 4. Perfect counting (assuming the data is entered correctly and the software works right) as opposed to manual counts where papers sometimes end up in the wrong pile. 5. You can offer voting in multiple languages. 6. You can randomise the order of candidates on the ballot, to evenly distribute any impact being at the top might have. 7. Ability to take backups for transit or deliver results electronically, avoiding boxes of ballots going missing. 8. Reduced counting costs and faster results, allowing more polling stations to be operated and allowing the polls to stay open later for the same result delivery deadline. Whether it succeeds at offering these benefits, and whether they're worth the obvious cost in trust is debatable.

danielweber 14y ago

If your plan to secure the machines is to have the source be open, you have already failed. You need a system that works well even if the source has been altered. Paper ballots have issues, too. They are different issues, and largely unseen by the general public until you have a really close election and have to use human judgment to decide whether that mark counts as a vote or not.

revelation 14y ago
And here I thought in the Turing year of 2012 people would finally start to understand what it means that any electronical voting machine contains a Turing equivalent machine. No magic can make the halting problem decidable.
- EvilTerran 14y ago
  An electronic voting machine doesn't have to be Turing-complete, surely? It would require especially-fabricated non-Turing-complete chips / ICBs, and some very simple interface like an LED bank and good old-fashioned physical buttons (as a touch-screen LCD unit alone is probably turing-complete), but I'm pretty sure it'd be possible to devise a specialised electronic voting system that couldn't have its behaviour altered without physical modification. [edit PS] Of course, this is purely academic. You might as well fantasise about a voting system made of Babbage-esque clockwork.
Peaker 14y ago
In really close elections, the outcome matters much less than the potential for systematic rigging of non-close elections.