**Action Plan and Open Questions**

- Challenges
- **Option 2: hash of a commitment, not sure what is this exactly**
- Implementation
- Root Circuit
- add checks between chunks
- check fingerprints
- check challenges
- PI circuit
- ???
- RWTable
- ???
- EVM & State circuits (chunk the table, EVM steps need to be consequent, dummy steps)
- state circuit gets info from PI
- BeginChunk / EndChunk logic
- Keccak Circuit
- reduce its size?
- Option 2 for challenges
- bus-mapping
- snark-verifier
- Supercircuit
- ???
- Chunk fingerprint chronologically ordered
- Chunk fingerprint ordered by address
- fingerprint for row
- fingerprint for chunk
- Bind chunk proofs by overlapping them by one row
- add a constraint that last row of previous chunk is equal to 1st row of current chunk
- fingerprint for the whole trace
- removing redundancies from the trace fingerprint (rows taken into account more than once due to chunk linking, see below)
- Constrains for global challenges
- **What are the checks that need to happen on last chunk?**
- Option 2 for challenges
- Poseidon (could we use https://github.com/zcash/halo2/tree/main/halo2_gadgets/src/poseidon ?)
- bus-mapping modifications
- snark-verifier modifications
- Testing
- Benchmarking
- Targets
- 24GB
- 32GB
- 64GB

**PSE meeting 3/10/2023**

Questions

- GG: do we create chunks based on Tchr or Taddr (expressed differently, is the first chunk created by the first N entries of Tchr or Taddr?)
- GG: Is the “local” correctness of Tchr already implemented?
- GG: Is the “local” correctness of Taddr already implemented?
- GG: address sorted trace, do we provide this as a witness and check it?
- GG: don’t understand the 2nd option for challenge generation. Some questions:
- what is proof chunk?
- do we need to have ‘Commit’ in-circuit?
- It is a KZG commitment?
- GG: not yet understood `BeginChunk` (check the previous chunk details are passed correctly) /`EndChunk` (like a write step context into RW table)
- GG: “Change scope summary” paragraph, I don’t understand this yet