It isn't allowed to have this. I hope this is clearer now.
[sigh] I wish you hadn't said that.
Actually, it's probably better that the spec be "too tight"
than that it be too broad. I was about to say, "white space processing
in boundary scanning doesn't have to be such a nightmare", but then I
brought more neurons to bear on what's happening. We're trying to
force binary through that-which-was a text mode service. Right?
Not really. I have not heard anyone express any intention of tranferring binary
using the existing transfer mechanisms (e.g. the DATA command in SMTP). A new
transfer mechanism is needed for binary. Currently it is described as an SMTP
enhancement, although it could have been a completely new protocol.
If that's correct, then use Base64. Did the light go on?
(am I even on the right list? Marilyn told me this morning not to
talk before having my coffee) I think the case was brought forth
where Base64 isn't an option. Fine. Then just don't call it MIME
in that case.
MIME is intended to be transport neutral. This means that we intend it to be
usable in a pure binary environment where there's no need to ever use
quoted-printable or base64. The fact that Internet email doesn't support such
an environment at the present time doesn't mean we are excused from getting
this right -- we're trying to look ahead a bit to a time when such environments
do exist. If we don't then we end up building automatic obsolescence into the
protoocol, which I really don't think is a very good idea.
Ned