4 ms·

Scammers isn't the least of it. The system they have designed and executed is ludicrously insecure and falls to even the most basic of flaws. It's unlikely that

by 0x8D3A 11y ago

Scammers isn't the least of it. The system they have designed and executed is ludicrously insecure and falls to even the most basic of flaws. It's unlikely that it will ever be able to achieve consensus given the endless stream of crashes and logic errors, like being able to send yourself a negative balance (thus, infinite money) or crashing every node in the network by sending a block header with a nonce of zero. The sheer quantity of security bugs and undefined behavior in the two clients attempting to act in bit-for-bit identical ways in two different languages is quite alarming. https://github.com/ethereum/go-ethereum/issues?utf8=%E2%9C%93&q=label%3Avuln+ https://github.com/ethereum/go-ethereum/issues?utf8=%E2%9C%9... I'd bet against it ever getting to try it's consensus given that their launch is now going to be completely centralized (ie, it can't stand up on it's own let alone with malicious attackers, so it's going to be remotely controlled by the developers for now). Will they ever remove the crutch and risk everything crumbling? Seems very doubtful.

drcode 11y ago
Though I disagree with you, I upvoted your comment because its content is all factually correct.
bachback 11y ago
Right. It's not really a block-chain, but a block-tree (whatever that is supposed to be). I suspect to make smart contracts work, one needs a different time-stamping/incentive system altogether (non-blockchain), but haven't seen any progress in that direction.
- DennisP 11y ago
  For more information on the block tree see the GHOST protocol paper, which predates Ethereum. It does collapse into a linear chain for all but the most recent blocks. http://eprint.iacr.org/2013/881.pdf http://eprint.iacr.org/2013/881.pdf
  - bachback 11y ago
    I know that paper. The authors have no background in distributed systems and use math to prove absurd things. If you look at the core proposed algorithm, it's not even a P2P algorithm.
    - DennisP 11y ago
      Eh? It's P2P just like Bitcoin. The only difference is how forks are resolved. Bitcoin nodes choose the longest chain they see. Another way to put that is: at any forking point, Bitcoin nodes choose the block with the longest chain after it. GHOST nodes choose the block with the most total computation after it (which may include additional branches).

by 0x8D3A 11y ago

drcode 11y ago

Though I disagree with you, I upvoted your comment because its content is all factually correct.

bachback 11y ago

Right. It's not really a block-chain, but a block-tree (whatever that is supposed to be). I suspect to make smart contracts work, one needs a different time-stamping/incentive system altogether (non-blockchain), but haven't seen any progress in that direction.

DennisP 11y ago
For more information on the block tree see the GHOST protocol paper, which predates Ethereum. It does collapse into a linear chain for all but the most recent blocks. http://eprint.iacr.org/2013/881.pdf http://eprint.iacr.org/2013/881.pdf
- bachback 11y ago
  I know that paper. The authors have no background in distributed systems and use math to prove absurd things. If you look at the core proposed algorithm, it's not even a P2P algorithm.
  - DennisP 11y ago
    Eh? It's P2P just like Bitcoin. The only difference is how forks are resolved. Bitcoin nodes choose the longest chain they see. Another way to put that is: at any forking point, Bitcoin nodes choose the block with the longest chain after it. GHOST nodes choose the block with the most total computation after it (which may include additional branches).