On Tue, 13 Dec 2016 14:32:04 +0000
Eran Messeri <[email protected]> wrote:

> * For Hash Algorithm repository, specify "Expert Review" guiding the
> Expert to make sure the proposed hash algorithm has public
> specification and does not suffer from known preimage attacks.

The hash algorithm also needs collision resistance.

How precise does the guidance need to be?  "Not suffer from known
preimage attacks" wouldn't be construed as forbidding SHA-2 because
there are some very expensive preimage attacks on some rounds, would it?

Arguably, "suitable for use as a cryptographic hash function" covers
both preimage and collision resistance, and leaves it to the expert's
discretion what attacks are worth worrying about.  Would leaving it at
this provide sufficient guidance?

Regards,
Andrew

_______________________________________________
Trans mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/trans

Reply via email to