Quote from: DeathAndTaxes on May 22, 2014, 03:35:08 AM

Quote from: Mtinie on May 22, 2014, 03:25:16 AM

I agree with you that a World without checkpoints would be infinitely better than a World with them, but all things considered, if they are a concern for Peercoin, it only should follow that they are a concern (though maybe slightly less) for Bitcoin, right?

You are right the same centralization risks apply to Bitcoin or any software project. The key difference is the level of control and the value of seizing that control. Developers could be forced to make malicious changes to the client but those changes, have the potential to be detected by other developers. As much as possible client development is an open transparent process (deterministic build process and multiple developer signatures). A single wary developer could alert the entire community. Users can choose to delay upgrading their clients, and the network doesn't consist of a single version of a single client. No major client support auto updating to make malicious code injection more difficult. The window for detection is longer.

If the Peercoin checkpoint keys were compromised an attack would occur instantly, without any window for detection. You would have a transaction with a lot of confirmations and then the network would receive a checkpoint update and your node (along with every other node) would jump to another chain. If you were double spent on that chain, your coins would simply disappear. It doesn't matter if you are online or offline, how closely you monitor the code, or if you prudently wait to upgrade. It would be very similar to a 51% attack except the attacker could perform this attack without a majority of the hashrate and/or stake. The checkpoint authority (in the hands of a single person) has "veto" power over the network.

You need to trust more than just that the developer won't be directly malicious. For the record I don't think he will be, but he may not have a choice. If a criminal organization (or three letter agency) held a gun to the head of the developer (or maybe to the head of a loved one would be more effective) do you still trust him to not checkpoint a chain containing double spends? If someone tortured him how long could he hold out before he gave up control of the keys which sign the checkpoints? If you think these types of scenarios are James Bondish now, well what about when the coin supply is valued in the billions of dollars? Tens of billions? Hundreds of billion? The only confirmed transaction is one that is behind the last checkpoint. Centralized security is easy. It is decentralized security that is hard. A decentralized network which relies on centralized security is an oxymoron.

While Bitcoin and Peercoin both use checkpoints they don't serve the same purpose. Bitcoin doesn't use checkpoints to prevent reorgs and thus the security model doesn't depend on them. I guess you didn't see the humor in the fact that the OP was so worried that a pool (consisting of thousands of miners) had 51% of the hashrate so the obvious solution was to jump to a coin where a single person has control over 100% of the checkpoints.

Very interesting! I am looking for a document(s) that outlines the workings of the bitcoin code & netwerk and possibly the issues that came in its wake. I would like to avoid reading all the code at this point. Does documentation like this exist or do I need to scavenge the internet for all the bits and pieces?

baare

Posts

Merits