At 15:53 11/07/95, John Gardiner Myers wrote:
RFC 1521 uses the term "object" consistently to refer to the thing
that the MD5 digest is applied to.
is a great example of why one has to get the canonical encoding model
It would have been possible to refine the set of definitions in 1521
to include term to designate, without ambiguity, the canonical content
of the body and the body as actually sent, with encoding applied. But
since I do not agree at all with the way the definitions are modified
in the latest drafts, I'm not going to make a proposal.