Post
Topic
Board Announcements (Altcoins)
Re: A new mining scheme
by
bitfreak!
on 18/11/2014, 06:33:33 UTC
No, proof of stake is a 3rd-party confirmation scheme too. What I'm arguing is that there is no need for 3rd-party confirmation on new transactions because the blockchain enables all nodes to confirm new transactions for themselves and extend the blockchain. As long as all nodes begin with the same genesis block and follow the same rules for confirming new transactions, accidental forking shouldn't happen. Even the number and size of blocks between nodes can differ but they remain effectively synced as long as they share the same transaction history.
I think your understanding of how Bitcoin works and how Cryptonite works is somewhat faulty or incomplete. Blocks are solved periodically by miners for a reason, and I wouldn't exactly call it a 3rd party confirmation scheme, it's a decentralized confirmation scheme that anyone can participate in. Nodes can't just accept any transaction they get without worrying about whether or not other nodes got the same transaction, for all they know they could be the only node which got that transaction. Blocks are periodically solved via a PoW mechanism because that's the only real way that everyone can agree on which transactions should be accepted and in which order they were received. When a miner creates a block they are deciding on which transactions to include in their block, and if they manage to solve the block which they created then they send it out to the network and everyone agrees to place that newly solved blocked onto the end of their chain.

A fork can occur when two miners solve a block at roughly the same time, nodes just accept which ever block they received first and start mining a new block linked to that block. So there will be two different groups of miners which are mining two different chains for a small period of time until someone else solves the next block, which should resolve the fork because everyone will start mining on the longest chain. I have thought about ways that such forking could be eliminated entirely, such as a deterministic rule for choosing one block over another when two blocks are solved at the same time, but if you think about it carefully there is a critical flaw in that idea which could be exploited. As far as I can see there is no viable way to totally eliminate accidental forks.