Ned Freed <NED(_at_)innosoft(_dot_)com> writes:
Another way to say it would be "the body prior to the application of the
content transfer encoding".
Being a formalist, I would say that prior to the application of the
content transfer encoding, it's not the body. It's like talking about
"the message prior to the addition of the headers".
RFC 1521 uses the term "object" consistently to refer to the thing
that the MD5 digest is applied to.
Its unfortunate that content-md5 has to be separate from the primary MIME
documents -- it would be both simplest and best to tie it into the canonical
encoding model at the proper place.
If the timing permits, I'd like to see content-md5 merged into the the
primary MIME documents by the time they go to Standard. Content-MD5
is a great example of why one has to get the canonical encoding model
_.John G. Myers Internet: jgm+(_at_)CMU(_dot_)EDU