Re: Nothing-at-Stake & Long Range Attack on Proof-of-Stake (Consensus Research)
Yes I see. Thanks. In effect, the 'bootstrap number' I've been talking about, is now 'PART OF THE PROTOCOL'.
Welcome. Indeed, it's been part of it from the beginning.
Although - if the initial stake holders ever sold their private keys once those accounts were empty, or lost / hacked / QC cracked, then maybe you could cause a little heart ache..
This is not possible because of checkpoints in the client. A newbie with a clean NRS downloading the blockchain from scratch would still reject the fake blockchain.
Quote from nxtforum.org:
Kushti has proved (and no one has produced any counter evidence) that the long range attack (Kushti's definition: Long-range attack - attacker can start fork hundreds or thousands blocks behind current chain) isn't possible.
As you say, you put the checkpoint in after 720 blocks had passed so it was behind the decentralised rolling checkpoints.
So why put the second checkpoint in at all? 'Just in case', maybe similar to having two deadbolts on your door?
It is just in case indeed (and I have 4 locks on my door, not two). It is to prevent history rewriting attack where somebody buys no longer used early stakeholder accounts to build a complete alternative blockchain, regardless of what the theory says if such attack is possible or not.As you say, you put the checkpoint in after 720 blocks had passed so it was behind the decentralised rolling checkpoints.
So why put the second checkpoint in at all? 'Just in case', maybe similar to having two deadbolts on your door?
A practical use of such checkpoint would be also if people provide for download a full copy of the blockchain, ending just before the checkpoint, then a node starting with such a copy and passing the checkpoint can be sure that blockchain is the same that other nodes running 1.4.8 or later use, up to the checkpoint at least. When downloading a compressed archive of the full database, a rescan is also needed if one wants to be sure that all tables containing account balances, assets and so on are populated correctly, but the checkpoint at least guarantees that the transactions up to it are the same.
it is refreshing to have such civil discussions here on btt for a change
