4 ms·

> most enthusiasts do not accept that the blockchain will became unmanageable by most just due to it's ever-growing size Some blockchain protocols like Bitcoin

> most enthusiasts do not accept that the blockchain will became unmanageable by most just due to it's ever-growing size Some blockchain protocols like Bitcoin make the silly argument that in order for a chain to be worthwhile, a full history of all transactions needs to be maximally available on the network. Other chains either drop, or plan to drop data that's been on the network for over a year or so (EIP-4444, for Ethereum). This relies on the weak assumption that the consensus algorithm will not finalize invalid data and then continue to build on it for over a year. When you take history expiry alongside state expiry, the technology is absolutely out there to bound blockchain size.

tromp 5y ago
> This relies on the weak assumption that the consensus algorithm will not finalize invalid data and then continue to build on it for over a year. You don't need to rely on such a weak assumption. You can use Incrementally Verifiable Computation to verify the entire blockchain history in constant time. See e.g. https://vitalik.ca/general/2021/11/05/halo.html https://vitalik.ca/general/2021/11/05/halo.html
- dmitriid 5y ago
  So, hackers remove your money or exploit an error in your smart contract, the history is there, okay. Then the data and transaction history is dropped. Yup, the "<magic buzzwords>" tells you that history is correct. So?
  - everfree 5y ago
    Just because you don't understand a data structure doesn't make it a magic buzzword. https://en.wikipedia.org/wiki/Merkle_tree#Uses https://en.wikipedia.org/wiki/Merkle_tree#Uses
    - dmitriid 5y ago
      That is not an answer to my question.
      - everfree 5y ago
        What was your question?
        dmitriid 5y ago
        You're hand waving it away in the parallel discussion.
  - tromp 5y ago
    Yes, losing access to the underlying data is a big downside. I think the ideal blockchain should offer constant time historical verification in addition to rather than as replacement of verifying the tx history. That also makes it robust against possible bugs in the design and implementation of the rather complex IVC technology. Even if IVC is the only way to verify the full history, then it could still be limited to older history, e.g. up to a week or month ago. That would give you some time to investigate recent hiccups.
dmitriid 5y ago
> Other chains either drop, or plan to drop data that's been on the network for over a year or so So when the next $625-million exploit happens [1] the hackers would just need to wait a year before any trace of them disappears? Awesome [1] https://web3isgoinggreat.com/?id=2022-03-29-0 https://web3isgoinggreat.com/?id=2022-03-29-0
- everfree 5y ago
  No, that is incorrect. I'm talking about the data availability requirement that the protocol imposes on nodes if they want to remain connected to the swarm. I'm not sure what you're talking about - data doesn't magically get deleted from the internet with "no trace".
  - dmitriid 5y ago
    > Data doesn't magically get deleted from the internet with "no trace". This is what you said, emphasis mine: "Bitcoin make the silly argument that ... a full history of all transactions needs to be maximally available on the network. Other chains either drop, or plan to drop data that's been on the network for over a year or so". I'm reading exactly what you wrote. I don't know, may be the meaning of words in crypto world is "wrong".
    - everfree 5y ago
      There's a wide spectrum between a piece of data being maintained by every participant of a particular blockchain and maximally available, and that data being completely unavailable to anyone. It's that spectrum between 0-of-N and consensus-enforced N-of-N that you don't seem to grasp. I'll say it again, once a piece of data is no longer maintained by all N-of-N participants in a blockchain as a matter of its consensus protocol, that doesn't mean that the data gets automatically deleted from the internet and from everyone's hard drives. On the contrary, there are many reasons why particular actors would want to retain that data, one of which is the very thing we are discussing - to keep a transaction history in case old transactions need to be audited for potential criminal prosecution of $625-million exploits. > Other chains either drop, or plan to drop data that's been on the network for over a year or so Yes, the chains, i.e. the actors that are keeping the minimal amount of data to continue to participate in the protocol, drop that data. That doesn't mean the data disappears from the internet. Instead, data availability is likely to remain incentivized within dedicated sub-protocols. That's not to mention the numerous private entities that will have their own external incentives to retain that data. Quote from Vitalik: > Older blocks, transactions and receipts/logs would still be accessible through dedicated sub-protocols (eg. the Portal Network) or externally developed protocols (eg. TheGraph), in addition to a much smaller but still sufficient number of volunteer nodes and block explorers. Note that many dapps are already moving their historical data queries to TheGraph and similar protocols for efficiency. > https://www.reddit.com/r/ethereum/comments/qzvsfq/impromptu_technical_ama_on_history_expiry/ https://www.reddit.com/r/ethereum/comments/qzvsfq/impromptu_...

tromp 5y ago

> This relies on the weak assumption that the consensus algorithm will not finalize invalid data and then continue to build on it for over a year. You don't need to rely on such a weak assumption. You can use Incrementally Verifiable Computation to verify the entire blockchain history in constant time. See e.g. https://vitalik.ca/general/2021/11/05/halo.html https://vitalik.ca/general/2021/11/05/halo.html

dmitriid 5y ago
So, hackers remove your money or exploit an error in your smart contract, the history is there, okay. Then the data and transaction history is dropped. Yup, the "<magic buzzwords>" tells you that history is correct. So?
- everfree 5y ago
  Just because you don't understand a data structure doesn't make it a magic buzzword. https://en.wikipedia.org/wiki/Merkle_tree#Uses https://en.wikipedia.org/wiki/Merkle_tree#Uses
  - dmitriid 5y ago
    That is not an answer to my question.
    - everfree 5y ago
      What was your question?
      - dmitriid 5y ago
        You're hand waving it away in the parallel discussion.
- tromp 5y ago
  Yes, losing access to the underlying data is a big downside. I think the ideal blockchain should offer constant time historical verification in addition to rather than as replacement of verifying the tx history. That also makes it robust against possible bugs in the design and implementation of the rather complex IVC technology. Even if IVC is the only way to verify the full history, then it could still be limited to older history, e.g. up to a week or month ago. That would give you some time to investigate recent hiccups.

dmitriid 5y ago

> Other chains either drop, or plan to drop data that's been on the network for over a year or so So when the next $625-million exploit happens [1] the hackers would just need to wait a year before any trace of them disappears? Awesome [1] https://web3isgoinggreat.com/?id=2022-03-29-0 https://web3isgoinggreat.com/?id=2022-03-29-0

everfree 5y ago
No, that is incorrect. I'm talking about the data availability requirement that the protocol imposes on nodes if they want to remain connected to the swarm. I'm not sure what you're talking about - data doesn't magically get deleted from the internet with "no trace".
- dmitriid 5y ago
  > Data doesn't magically get deleted from the internet with "no trace". This is what you said, emphasis mine: "Bitcoin make the silly argument that ... a full history of all transactions needs to be maximally available on the network. Other chains either drop, or plan to drop data that's been on the network for over a year or so". I'm reading exactly what you wrote. I don't know, may be the meaning of words in crypto world is "wrong".
  - everfree 5y ago
    There's a wide spectrum between a piece of data being maintained by every participant of a particular blockchain and maximally available, and that data being completely unavailable to anyone. It's that spectrum between 0-of-N and consensus-enforced N-of-N that you don't seem to grasp. I'll say it again, once a piece of data is no longer maintained by all N-of-N participants in a blockchain as a matter of its consensus protocol, that doesn't mean that the data gets automatically deleted from the internet and from everyone's hard drives. On the contrary, there are many reasons why particular actors would want to retain that data, one of which is the very thing we are discussing - to keep a transaction history in case old transactions need to be audited for potential criminal prosecution of $625-million exploits. > Other chains either drop, or plan to drop data that's been on the network for over a year or so Yes, the chains, i.e. the actors that are keeping the minimal amount of data to continue to participate in the protocol, drop that data. That doesn't mean the data disappears from the internet. Instead, data availability is likely to remain incentivized within dedicated sub-protocols. That's not to mention the numerous private entities that will have their own external incentives to retain that data. Quote from Vitalik: > Older blocks, transactions and receipts/logs would still be accessible through dedicated sub-protocols (eg. the Portal Network) or externally developed protocols (eg. TheGraph), in addition to a much smaller but still sufficient number of volunteer nodes and block explorers. Note that many dapps are already moving their historical data queries to TheGraph and similar protocols for efficiency. > https://www.reddit.com/r/ethereum/comments/qzvsfq/impromptu_technical_ama_on_history_expiry/ https://www.reddit.com/r/ethereum/comments/qzvsfq/impromptu_...