lwj(_at_)cs(_dot_)kun(_dot_)nl writes:
Greg Vaudreuil writes:
Quoted-binary is a new feature which has yet to be implemented
and if included in MIME will need to be implemented and tested by
at least two independent implementations before MIME can be advanced.
Given the intended similarity to quoted-printable, it should be
trivial (ie, a very small amount of work) to add quoted-8bit to an
existing implementation supporting quoted-printable. The problem
is then to find an 8-bit transport channel to test
interoperability over.
This strikes me as a little strange. If you have an 8-bit
transport layer, then there is no reason to use 8-bit quoting.
(since you can move all 8-bits across your transport layer.)
If your transport layer is 7-bits, then you would need to quote
8-bit characters, but that is already covered with the
quoted-printable features in MIME.
It seems to me, however, that you are mixing up the layers of the
mail model. My understanding of the MIME RFC is only concerned with
how to represent the multiple-functions of mail (the MUA), as is not
(and should not be) concerned with the lower levers such as how the
mail is ultimatly delivered (the MTA).
Greg Vaudreuil continues:
I do not believe that the increase in complexity to optimize the
sending of Binary data over 8 bit paths is worth the saved
bandwidth. I'd rather see efforts go into defining a newMTP that
will expressly deal with Binary Data.
right. When a binary transfer method becomes common many of the
transfer encodings will become obsolete. So why clutter up the MIME
specs with unnecessary complication.
For example, fidonet already provides a binary transport layer,
and it is possible for any gateway to a 7bit world can translate
the body-parts into base64 or quoted-printable if necessary.
--
Erik Seielstad | Bitnet: erik(_at_)brock1p
Systems Programmer/Analyst | Internet:
erik(_at_)acspr1(_dot_)acs(_dot_)brockport(_dot_)edu
SUNY College at Brockport | or:
erik(_at_)rochgte(_dot_)fidonet(_dot_)org