How you can really know electronic voting is a bad idea - all the people who would be the ones to be hired and paid a ton to build and implement it, cryptographers and infosec experts, are the same people who loudly oppose it.
There are so many different problems;
Assuming you manage to build the perfect system, can you actually explain it to people? If you put a prospective voter in front of the real deal, a secure electronic voting machine, and put a fake replica next to it, will ANYBODY EVER succeed in telling them apart? Or will you be forced to continuously audit the hardware from production to shipment to use, and somehow still respect voters’ privacy as they use the machine?
And how can you be confident your implementation is secure? You can to prove the algorithms are correct, that the implementation is correct, that the implementation behaves as end users (voters) expect, and that even the hardware is flawless (both in terms of logic and protection against manipulation).
How do you ensure people only vote once, yet also protect their privacy, but ALSO prove to them their vote was really counted? Individual keypairs? How do you distribute and protect them? Physical ID cards? What if they get stolen (in significant enough numbers)? What if a significant fraction of the population won’t use the new system, do you still run old school voting and combine the results?
How do you give somebody a receipt they can understand without enabling coercion to try to force people to vote a certain way?
How you can really know electronic voting is a bad idea - all the people who would be the ones to be hired and paid a ton to build and implement it, cryptographers and infosec experts, are the same people who loudly oppose it.
There are so many different problems;
Assuming you manage to build the perfect system, can you actually explain it to people? If you put a prospective voter in front of the real deal, a secure electronic voting machine, and put a fake replica next to it, will ANYBODY EVER succeed in telling them apart? Or will you be forced to continuously audit the hardware from production to shipment to use, and somehow still respect voters’ privacy as they use the machine?
And how can you be confident your implementation is secure? You can to prove the algorithms are correct, that the implementation is correct, that the implementation behaves as end users (voters) expect, and that even the hardware is flawless (both in terms of logic and protection against manipulation).
How do you ensure people only vote once, yet also protect their privacy, but ALSO prove to them their vote was really counted? Individual keypairs? How do you distribute and protect them? Physical ID cards? What if they get stolen (in significant enough numbers)? What if a significant fraction of the population won’t use the new system, do you still run old school voting and combine the results?
How do you give somebody a receipt they can understand without enabling coercion to try to force people to vote a certain way?
Yeah, that about sums it up.