[andrewla]

First of all, what you're talking about more resembles a Merkle Tree rather than a blockchain, because the "chaining" property is really useless in this scenario. Each election can publish the Merkle Tree of its results and you can be sure that your vote was properly registered. Or frankly, just publish the list of one-way hashes and their vote, and you can dispense with all the Merkle-ing.

But what about a Sybil attack? How do you ensure "one person == one or zero votes"? I could submit a jillion votes for Donald Duck and how would you ever know that those votes were all cast by the same person? Any sort of election scheme has to deal with messy real-world identity, and there's no cryptographic solution to that, only various weak social network approaches that are pretty much the norm.

[creeble]

A Sybil attack in this case is just a reveal (once again) of the oracle problem - a blockchain doesn’t provide proof that you are you. Therefore, it cannot provide proof that you cast only one (or no) vote.

Verifying your identity is outside the blockchain. Thus it can provide no value for voting.

[maneesh]

As to your first point, I'm not familiar with Merkle Tree's, so I'll learn about that before I respond. Thank you for the insight.

To the second point--I would imagine you would vote in the same way we do today for MVP, in-person / mailing, etc. So the main function would be to verify that your vote was properly recorded and counted.