We still benefit from a function that recovers the public key from the signature. This EIP unfortunately does not provide that.
I have examined the way this decision was made. Essentially the rollup people rushed out the first spec they thought of and then it became their standard as they rushed to adopt passkeys. Attempts to discuss this were put off. It was rushed. Then this proposal copied the RIP also without scrutiny.
I don’t think the interface has to match. It’s already at a different address.
I might be mistaken, but both signature malleability and recovery ID can be eliminated without increasing the signature size from 512 to 513.