Re: [MMUSIC] 4572 update: forbid weak hashes?

[[e-mail] (Dale R. Worley)]

Martin Thomson <[e-mail]> writes:
> Maybe three pieces of advice are right.
>
> 1. Implement the strongest hash you can
> 2. Check all the hashes you can
> 3. Don't accept a session if you can only check weak hashes

I'm just now looking at this, but it seems to me that it is sufficient
for the receiver of the hashes to check any one hash that it considers
to have a sufficiently secure algorithm.  If the hash was sent to it,
that shows that the sender considers it to be secure enough.  (Of
course, if it does not receive any hash that it considers secure, the
receiver should reject the negotiation.)

There doesn't seem to be any reason for the receiver to check more than
one sufficiently secure hash.  Because it received them, the receiver
knows that the sender thinks both hashes are secure, and by hypothesis
the receiver thinks they are both secure, so every party to the
transaction expects that if one hash is valid, the other one will be.

Of course, we could allow the receiver to check multiple hashes if it
wishes...

Dale