It was observed that a MIME-PEM implementation, as currently
defined, requires the user to have a MIME compliant mailer plus
PEM-specific features. Thus the current proposal, if adopted as
a replacement for RFC 1421, would remove support for 822-PEM.
This proposal was not uniformly well received.
Although it may have appeared this way, this was not the intent, which
was unfortunate. Rather, the proposal was redefining the PEM-822
specification. In particular, if the following was true:
1. PEM adopted the MIME conventions for BEGIN/END delimiters.
2. PEM dropped the canonicalization step in its processing and instead
required its input stream to be in a canonical form.
3. PEM adopted the MIME conventions for nesting, instead of the RFC 934
conventions.
it would be entirely possible for there to exist 822-PEM only user
agents. However, it also turns out that MIME compliant user agents that
support PEM would also be able to understand the 822-PEM only generated
messages. This, of course, is a big win.
There's a couple of other minor details, but I think the objective right
now should be to revise the current proposal to make it clear what is
intended and to cover all the details. Then we can all discuss it with
a common frame of reference.
Jim