Lucas Xia

A lot of them are not const, but they should be.

compute_logderivative_inverse seemingly creates the lookup_inverses polynomial and doesn't rely on shifts so it could be possible to use a provingkey instead.

Prover_instance and verifier_instance shouldn't be in sumcheck/

For native file structure, we are differentiating between plonk and honk by putting plonk stuff in a plonk/ directory. However, the reverse is true for stdlib - we have a...

This could also include the recursive verifiers.

Currently, the flavor-specific Transcripts, which define the structure of a transcript and provide serialization/deserialization functions, are parameterized by Commitment to generalize to the recursive case, which is wrong. They should...

We compute the power polynomial with an outer loop that iterates over 2^d (d being the log size and also the number of betas), and use a O(d) inner loop...

We currently inefficiently compute G^(0) in nlogn time where n is the length of the polynomial. We can reduce this to to O(n) time by precomputing coefficients, which should be...

Currently, the UltraCircuitBuilder doesn't add any gates for hashing. One fix would be to implement the stdlib Poseidon2 permutation using standard arithmetic gates.

Brought up because of usage in bigfield constructor and an old usage in convert_to_grumpkin_fr: https://github.com/AztecProtocol/aztec-packages/pull/4161/files#r1480434996. This function may be unnecessary but should be double checked. Also should audit convert_to_grumpkin_fr which...