Re: [Cfrg] Call for adoption: draft-hdevalence-cfrg-ristretto-01

["Filippo Valsorda" <[e-mail]>]

2019-09-19 00:36 GMT+02:00 Christopher Wood <[e-mail]>:
> I'm in favor of adoption as well, though am curious to know where folks 
> stand on the issues discussed in [1] and earlier messages. What is the 
> plan for reconciling this with the hash-to-curve draft? Of the three 
> options Filippo outlines in [2], the third seems the most appealing and 
> probably the most workable, though does muddy the "curve" focus of 
> hash-to-curve.
> 
> Best,
> Chris
> 
> [1] https://mailarchive.ietf.org/arch/msg/cfrg/w91AdGVdLWccafiPub862MObk5c
> [2] https://mailarchive.ietf.org/arch/msg/cfrg/p_r53FtGcaNSB_HwIV0J9FSgYbg

We plan to prepare -02 to address all the unclear parts that emerged
from the discussion on this list. The reason it's taking longer
than we'd like is that we want to simultaneously port one of the
implementations to a different curve, so we can have discussions more in
concrete.

As for hash-to-curve, we agree that invoking FROM_UNIFORM_BYTES
from hash-to-curve is the best way forward. It preserves the API
abstraction and avoids leaking Ristretto internal representatives into
the hash-to-curve spec and its implementations, while still leaving
hash-to-curve in charge of domain separation and everything else.

Thank you,
Filippo