I have a couple of questions about the algorithms specified in the next draft:
Elliptic Curve: I've added identifiers for EC-encryption and ECDSA. Is this
enough?
X9.42: Should we add it?
ROT-N: The reason this is here is to give an algorithm that all the proper
transforms can be done on -- IV construction, encoding engines, etc. -- but
won't be export controlled in a reference implementation. This would allow
an international team of people to build and test the reference
implementation. However, it has to be defined in more detail than that to
really be interoperable. I think it's both useful and amusing, but I'm not
so committed to it that I've actually done the work to spec it out. Someone
who believes in having it needs to write up the details. If no one does so,
I'm removing it.
Haval: Similar to the above, someone who wants Haval has to decide how many
rounds and the length. In Washington, we agreed that it should be at least
as strong as SHA-1 or RIPEMD160. Someone with an opinion should state it,
or it goes.
Blowfish: Ditto. We need the variant specified.
Jon