On Thu, Aug 18, 2016 at 02:21:27PM +0200, Jari Arkko wrote: > Thanks for your review, Roni! > > Authors, did you observe the editorial suggestions? Did the changes to Editor's Copy. The comment about corner-cases I just deleted[1]. My understanding (I am a newbie to writing IETF RFCs) is that in this stage of process, I should not submit updated version without being told to do that (IIRC, by shepherd, AD or datatracker state). [1] Basically, ECDSA specifies that if the hash is longer than the bitlength of the order, the hash gets truncated. With NIST curves and hashes any such truncation is always to integral number of octets. However, Curve25519 is so small even SHA-256 would truncate, and since its order is ~2^252, the truncation would be to fractional number of octets. -Ilari.