At the risk of restarting a holy war...
I understand that there is apprehension to allowing nested encodings
in the MIME spec. From the text:
"Nested encodings add considerable complexity to user agents: ... they
can obscure the basic structure of a message."
This sounds like a very desirable (perhaps mandatory) trait when
adding privacy-enhancement to messages. The formation of a
"message/pem" content type/subtype sounds like a nice fit for merging
MIME and privacy-enhancement if not for this nesting prohibition.
Therefore, I would like to see this requirement revisited. Otherwise,
we are left defining a application/pem and commiting the nesting sin
under the guise of an specific application.
Does not the allowance of nesting permit greater transportability and
other benefits ?
Is recursion at the UA level that bad ?
Perhaps there is a compromise that can be worked out. For instance, if
nesting were allowed, having a UA automatically un-nest should not be
a conformance requirement (i.e. might require an explicit action by
the user...)
Comments (flame shields up...)
Steve Dusse