Alex Xiong

> thought a bit further... I kind of like that point being associated with SWParameter as before, rather than a given field. This makes more logical sense to me --...

## Problem I figured out why, because we have both [`From for Point`](https://github.com/EspressoSystems/jellyfish/blob/main/plonk/src/circuit/customized/ecc/mod.rs#L55-L68) and [`From for Point`](https://github.com/EspressoSystems/jellyfish/blob/main/plonk/src/circuit/customized/ecc/conversion.rs#L22-L58), where the latter was reserved for `F: SWToTEParam` or for BLS12-377 in our...

okay, after some attempt, I can't avoid using `` in the generic param, so I will close this issue now. that said: > we should remove the option to convert...

now arkworks already have HashToCurve implementation, correct? should we close this? @zhenfeizhang

Sounds good, the only concern would be, we can improve the documentation more strategically given that our constraint system API (i.e. `jf_plonk::circuit::*`) might be updated significantly. (contingent on our investigation...

sorry to chime in here, > The reason to call it Builder is because you use it incrementally build up the R1CS matrices and variable assignments. imo, "builder" has the...

I'll take a look at CHANGELOG today and post my opinion afterwards.

@adklempner `RefundApp` is not meant to be a `CounterfactualApp`, it's used during deposit protocol, to ensure newly deposited eth/token into the multisig can be withdrawn. The documentation is still under...

Ah, I see. I do agree with converting `ETHRefundApp` to be `CounterfactualApp`, but prefer the current "abstract + implementation" over "default implementation + overload function". To convert, 1. `AppState` should...

because to comply with `pure` accessibility, reading the balance needs to have a workaround, so I assume that just specifying the exact withdrawable amount would also work.