These are notes on fellow apprentices' notes.

25/06/2021: No longer kept up to date, but I'm monitoring note changes and updating other parts of my notion notebook.

• Aaryamann
  • Rollups vs state channels
    • see Rollups
    • state channels are related to Bitcoin's Lightning Network
      • bidirectional payment (& more) channel forming a network
  • Stateless Client Concept
    • (also see State Expiry)
    • Let "stateless clients" trust some initial (hash of) state, and then let other people supply them with merkle proofs/commitments of the accessed state, so that they can compute the new state hash forward (that requires all relevant hashes in the computation to be provided, still immensely smaller than the whole tree).
    • Trades off network use for long-term storage.
    • This is for non-miners clients. Miners/stakers still need the full state, if only because state access is dynamic and cannot be determined by statical analysis (without running the contract).
      • Wait, don't they need the state just to perform the computation in the first place?
        • Yes, but the transaction submitter could submit all the state proofs if it can be statically analysed. But then who keeps the state?
      • Idea: bundle a list of account that can be accessed alongside transactions.
      • Stateless miners adds a slew of benefits. It removes a DOS vector, improves parallelization (transactions that touch different states can be run in parallel)
        • Also prevents DOS related to soft forks, as could have happend in the DAO attack
      • In sharding, switching validators between shards increases security, but when you switch shards you need to download the full state of the shard.
      • The issue is: who stores the data?
        • Vitalik proposes to store all recently (past few months) accessed data, then to forget randomly (so that older data would still be remembered with decaying probability depending on last data access).
        • Nodes can query the network for data.
        • Could use state channels to offer payment to ensure further data availability.
  • Network Infrastructure
    • Very good overview, just read it directly
  • Some miners mine empty blocks directly after blocks with transactions, as a way to get an even bigger speed advantage over the competition
• Albus
  • Lots of good links on state expiry & verkle trees / kate proofs
• Jonas Fietz
  • Good project ideas, including info the portal network
  • Some great links
    • Gas limit coordination (relies on a token to reward good behaviour)
    • Avoided attack against Berlin
• Evan
  • The Potential of Cryptographic Code Obsfuscation (not yet available)
    • Indistinguishability obfuscation (might be possible)
      • Breakthrough paper?
    • Black box obfuscation (impossible)
• Francesco
  • Notes on eth1.0-specs, including mention of key reference python implementaitons (py-evm, py-trie & trinity)
  • Other idea: implement EIP-1559 (fee market change) for py-evm
• iauns
  • good collection of links for the C++ and Python ecosystems
• Jared
  • typed transactions
• Milan Cermak
  • more stateless link
  • earlier roadmap to statelessness
  • account abstraction link tree
• odysseas
  • helpful links on L2 and MEV (miner extracted value) (TODO)
• Ognyan
  • More stateless links
  • The Data availability problem under stateless ethereum
  • Beam sync in 80 seconds, using meta witnesses (unread)
  • witness exchange protocol (unread)
  • Witness gas cost reforms (unread)
• owl
  • one idea: modify EVM to make rollups easier to implement
• potuz
  • Project: obtain a real life quantification of the incentives [of validators] to cheat
    • some examples of the incentives to cheat are provided
• th4s
  • some great statelessness links
  • Why it's so important to go stateless
    • Go to full statelessness and not only weak statelessness (with state expiry)
      • Include kate proofs for every touched value in the block
      • It would enable validators to not store the state (proposers still have to)
      • It would make fraud proofs easier because a blocks is its own fraud proof (just run the transactions with the proven values)
  • A Theory of Ethereum State Size Management
  • The state of stateless Ethereum (12/2019)
    • beam sync (not read yet) - progressively fill up the state trie by requesting witnesses for missing states
  • An updated roadmap for stateless ethereum
• Voith
  • is working on eth1.0-specs, an executable specification for the yellow paper, in Python
  • Fe (Language)
    • Solidity alternative with bounded computation with the aim of decidability