Hacker News Reader

Comment by DennisP | Hacker News Reader

That is not what they're doing. When you have a known set of entities and trust that at least 1/3 of them will be honest, you can use other consensus systems like Byzantine Paxos. There's at least one project combining that consensus algorithm with the Ethereum virtual machine.

Just passing signed messages around wouldn't be sufficient, you need a way of ensuring overall consensus.

View on HN

I agree with your post, but Byzantine Paxos require at least 2/3rd to be honest not 1/3rd

Ah that's right, thanks.

That's not a blockchain, paxos has been around since the 90s and don't offer objectively true information (only subjectively true) . they haven't been used by banks prior because they're insecure and not nearly as efficient as centralized message passing protocols.

You're using an arbitrary, uncommon definition of blockchain. A blockchain is a Merkle tree of validated transactions with a consensus algorithm.

That is the definition of blockchain. It's what was invented in Bitcoin. All other definitions would suggest blockchains were invented in the 90s or 80s

thats the best consolidated broad definition of a blockchain I've heard. Do you mind if I use it?

Should we also include elliptic curve signatures?

Some kind of public key signature seems essential, doesn't necessarily have to be elliptic curves. There's already talk about using post-quantum algorithms.

But verifying the signature is part of the validation algorithm, so "validated transactions" already covers it.

It's a terrible definition. There's no mention of pow, which was the innovation of Bitcoin, and that which secures the network

Reply on news.ycombinator.com