🚀 go-pugleaf

I haven't heard from SEF.  But it seems that Jim De Arras bounces
MIME.  It seems that there are some implementation deficiencies in the
Jim De Arras bouncer though; it seems to take the entire text and
transmit it as a Next attachment.  More specifically, it takes
perfectly good email-safe encoded text, and performs the following
operations:
	1) prepends a short RTF header at the beginning.
	2) prepends a "> " to the front of every line
	3) appends the text "Pretty cute, an almost 1 meg post.
	   What's your point?" in RTF along with a signature.
	4) UNIX-compresses the entire mess.
	5) uuencodes the entire mess.

This seems to be a lot of extra overhead.  It added about 100
kilobytes plus adding several extra steps required to decode and
extract individual pieces of the message.  The "> " characters
inserted in every line are a particularly large part of overhead, as
is the encoding of data which has already been encoded.  I would think
that any machine running the Jim De Arras bouncer would end up
spending a lot of CPU overhead in these extra steps.

The equivalent could be done in MIME, including signatures and making
all parts of the innermost encapsulated message available, with a
couple of hundred bytes.  Refer to the MIME document, which you can
find on thumper.bellcore.com on pub/nsb/BodyFormats.ps (or .txt if you
can't read PostScript).  Specifically, it would add something like:
	MIME-Version: 1.0
	Content-Type: MULTIPART/MIXED;boundary="105foobar"

to the header.  Then it would start the body with:

	--105foobar
	Content-Type: MESSAGE/RFC822

then incorporate the text of the message to be bounced.  Then append:

	--105foobar
	Content-Type: TEXT/PLAIN;charset=US-ASCII

	Pretty cute, al almost 1 meg post.  What's your point?
	
followed by the signature and

	--105foobar--

afterwards.  I used plaintext here since I didn't see any particular
usage of the RTF encoding.  I'm sure this can be implemented much more
efficiently than the Jim De Arras bouncer.  Also, the receiving agent
can conveniently extract the bounced message (or its contents) in this
format for processing in some other fashion.

If/when I hear from the SEF bouncer, I will report the results along
with a comparative analysis of its overhead and ease of parsing by the
recipient.

With tongue firmly in cheek...