Some things just work, like paper ballots. No reason to re-invent the wheel or to "verschlimmbessern" what works.
We vote a lot in Switzerland on a lot of issues but we do so on paper ballots which we can either drop directly in the box or send in the post. When there is a close vote the maximum wait for a result is usually around 4-5 hours so that isn't really an issue either. Counting is a highly distributed effort and IMO that also reduces the risk for large scale fraud.
It absolutely doesn't work. All paper elections have some (acceptable and accepted) level of fraud. We should move to mathematical system, that still uses paper but let's the voter confirm that thier vote was properly counted. There was a TED presentation about this many years ago.
Ok. If "works" means, "is good eonugh to be used for the purpose", I guess it works. But shamanistic medicine wokrs by the same measure so it's really not a high bar to clear.
It sounds like "3 out of 3" is too risky, as you're basically tripling the risk of losing a key (but you're reducing the risk of compromise). Something like "3 out of 4" would have been a better balance, in my opinion, but I think there were technical issues in requiring such a quorum (I think I read that the encryption scheme didn't support it, but don't quote me).
Previously: A cryptography research body held an election and they can't decrypt the results https://news.ycombinator.com/item?id=46020596
Ah, sorry. I'd only searched for cryptology and should've been more thorough.
It's the same story.
Previously:
A cryptography research body held an election and they can't decrypt the results - https://news.ycombinator.com/item?id=46020596 - Nov 2025 (38 comments)
Some things just work, like paper ballots. No reason to re-invent the wheel or to "verschlimmbessern" what works.
We vote a lot in Switzerland on a lot of issues but we do so on paper ballots which we can either drop directly in the box or send in the post. When there is a close vote the maximum wait for a result is usually around 4-5 hours so that isn't really an issue either. Counting is a highly distributed effort and IMO that also reduces the risk for large scale fraud.
It absolutely doesn't work. All paper elections have some (acceptable and accepted) level of fraud. We should move to mathematical system, that still uses paper but let's the voter confirm that thier vote was properly counted. There was a TED presentation about this many years ago.
Evidence says it works. And evidence beats ted talks any second, to the constant surprise of the tech (or influencer) community.
Ok. If "works" means, "is good eonugh to be used for the purpose", I guess it works. But shamanistic medicine wokrs by the same measure so it's really not a high bar to clear.
Why does the IACR use the term "cryptology" rather than "cryptography"?
Cryptology is the science, cryptography the practice.
It sounds like "3 out of 3" is too risky, as you're basically tripling the risk of losing a key (but you're reducing the risk of compromise). Something like "3 out of 4" would have been a better balance, in my opinion, but I think there were technical issues in requiring such a quorum (I think I read that the encryption scheme didn't support it, but don't quote me).
This headline is incorrect, elections were rescheduled, not canceled.
I guess it's my turn to post it -- https://m.xkcd.com/2030/
Like fine wine
Better than losing the key and finding a "workaround" I guess.
https://www.iacr.org/news/item/27138
[dead]