Currently, chilldkg_recover really just replays the transcript, i.e., it calls _round1, _round2, etc. This is simple, but it does many redundant computations.
We could instead write a specialized recover that computes only what's necessary. In particular, PoKs don't need to be checked (or even included in the transcript).
Before we go that route, we'd probably need to factor out some more routines out of the _round1, _round2, etc. to avoid code duplication.
Currently,
chilldkg_recoverreally just replays the transcript, i.e., it calls_round1,_round2, etc. This is simple, but it does many redundant computations.We could instead write a specialized recover that computes only what's necessary. In particular, PoKs don't need to be checked (or even included in the transcript).
Before we go that route, we'd probably need to factor out some more routines out of the
_round1,_round2, etc. to avoid code duplication.