You have interoperability issues when you multiply SHOULDs and MUSTs, not MAYs.
MAYs define a way to use algorithms and allow applications and users to use it.
You never know what future is made of. We already have demands from clients who
wish these algorithms to be supported in S/MIME, TLS, ASx, and so on, FOR
INTEROPERABILITY with their current or potential partners. So I think it is
better to define one single way to use algorithms as soon as possible rather
than let everybody define its own way, and merge methods later on.
Regards.
Antoine Alberti
XPP project manager
+33 (0) 1 47 17 24 37
aalberti(_at_)axway(_dot_)com <mailto:aalberti(_at_)axway(_dot_)com>
Axway. software a Sopra Group company
www.axway.com <http://www.axway.com>
-----Message d'origine-----
De : owner-ietf-smime(_at_)mail(_dot_)imc(_dot_)org
[mailto:owner-ietf-smime(_at_)mail(_dot_)imc(_dot_)org]De la part de Paul
Hoffman
Envoyé : mardi 28 juin 2005 02:45
À : Blake Ramsdell; Russ Housley
Cc : ietf-smime(_at_)imc(_dot_)org
Objet : Re: Support for hash algorithms other than SHA-1
At 12:58 PM -0700 6/27/05, Blake Ramsdell wrote:
Are there any reasons besides implementation ease to promote or
discourage the SHA-512-derived algorithms?
Flinging unnecessary options at implementers and users does not help
get use or interoperability. If someone believes that there is a real
class of users for whom a collision attack that requires *more than*
2^128 work exists, I'd like to hear it.
--Paul Hoffman, Director
--Internet Mail Consortium