Closed chris-wood closed 4 years ago
Originally posted by @chris-wood in https://github.com/cfrg/draft-irtf-cfrg-voprf/pull/78
Unrelated to this change, but what is
u
here? Is DST guaranteed to be of sizeu
? (2HashDH doesn't specify the hash this way. Maybe we ought to go back to what's in the original paper?)
This was done #97
Finalize uses
H_2
like so:The reference implementation basically uses
DST
as a keyed hash, yet we define it as a hash function. Maybe we should just say this is a PRF, and then concretely say it's HMAC-H (with H depending on the suite) keyed byDST
? (That works fine for HMAC as it's a dualPRF, but perhaps not other PRFs, so I might suggest swapping the input orders.)Also, for what it's worth, 2HashDH doesn't require H_2 to be a PRF, so maybe this is a bit much. Maybe we should just use a normal hash (SHA2) and define an encoding of the inputs?