3 ms·
> This is why they are also slow and expensive. They're not all that expensive - data is inconclusive, but looks like they're anywhere from twice as expensive
by deogeo 7y ago
> This is why they are also slow and expensive.
They're not all that expensive - data is inconclusive, but looks like they're anywhere from twice as expensive as e-voting [1] to cheaper [2].
And they're not all that slow either - only takes about two days to get results, about a week to make them official. Which is much, much faster than necessary, since the time it takes to form a coalition, or take over governing, is longer.
So it's entirely unclear what problem e-voting is meant to solve, other than enriching voting machine vendors.
[1] https://www.researchgate.net/publication/327977453_How_Much_Does_an_e-Vote_Cost_Cost_Comparison_per_Vote_in_Multichannel_Elections_in_Estonia https://www.researchgate.net/publication/327977453_How_Much_...
[2] https://www.wired.com/2008/04/the-cost-of-e-v/ https://www.wired.com/2008/04/the-cost-of-e-v/
- tmm84 7y agoThe problem, at least what I remember, was that each state had a different voting style. Some had punches, some had squares you filled in by hand. The chance to make mistakes and the like was there (crossing out a square or double punching) and some voting booths/voting locations only allow straight votes (all x or y party only). Voting machines were designed to make those mistakes/situations less likely.