A decentralized distributed consensus currency, so to speak. At least aiming to be decentralized.
I don't really get the why. Why not some variation of proof of stake? For example a version where one stake equals one virtual node. No need for trust when there's financial interest.
Yes. We actually have two layers of proof-of-stake. There is the blockchain consensus, then within the blockchain we have coinhours, which are burned as fees. That is a form of proof-of-stake. A forked blockchain with transactions reverted will have a lower total coinhour burn.
Why will it have? What happens when attacker has an older and bigger stake and transacts it in a fork?
It only adds age as a security parameter.
Open issues, with this are whether to have the votes on the blockchain or off-blockchain and whether quorum decisions should be on blockchain.
If they aren't, how does new node, with old blockchain, decide which blocks to accept? It doesn't have the necessary information.
You can combine, web-of-trust with the proof of stake/voting mechanism. You could require that changes to the Skycoin source code (for the blockchain module) require a majority of coin holders to vote to ratify the changes. You hash the files in some way and get a root hash and people vote on the hash. Then if someone gives you source files, you can hash them and get a hash and compare it to what it should be.
What happens when majority of stake tells something other than majority (direct or by some algorithm) of trusted nodes/accounts? The way I see it, if trust wins, you may as well remove PoS, and the opposite if PoS wins.
We think we can go beyond proof-of-stake. We believe it is possible to be 51% attack proof, even if the attacker controls the majority of the coins and the majority of hashing power. We want an absolute, mathematical guarantee and that is what we are working towards.
How? A node with genesis only connects. Two blockchains exist. True - the one that was the first - blockchain has 49% of stake behind it, false (constructed with fake time to force a doublespend from true) 51%. Now what?