Re: MIME is a complex object description, not a transport

• Messages sorted by: [ date ][ thread ][ subject ][ author ]
• Previous message: Ned Freed: "Re: the 8-bit dilemma"

At 01:23 PM 2/29/96 EST, Jueneman@gte.com wrote:
>
>My suggestion, therefore, is that we view MIME and similar protocols as being 
>descriptions of a class of generalized, labelled, encrypted/signed, 
>complex/compound, local/remote/distributed/aggregated objects that have to be 
>securely stored and occasionally transmitted, whether the transmission takes 
>place via SMTP, HTTP, FTP, SSL, MSP, or sneaker net.

Agreed.  The more I look at this, security needs an envelope.  MIME might
work, might as well be made to work.  Transmission is a separate matter.

>In particular, I do NOT want to keep perpetuating the serious flaw, IMHO,
which 
>cripples X.400 and to a certain extent MSP, i.e., that of securely delivering 
>the bits to the destination in a very nice, neat package, and then letting
them 
>spill all over the floor in a disorganized jumble of clear text once the 
>package arrives. 

Right.  The transport mechinism can leave the secure envelope intact and
open it at the user's request, or deliver it as a secure envelope outside of
the transport program, for another program to manipulate it.  Afterall, we
do have MPACK from CMU as a standalone MIME processor already.

>Once you have a reasonably compact representation of such an object, you
should 
>be able to bit-stuff it, forward-error-correct it, disperse it over various 
>kinds of media, and put it all back together again in a completely transparent 
>manner.  But it seems very difficult to do this if the message transport
system 
>is supposed to peer down into the object in a non-transparent manner, and try 
>to second-guess what the object consists of. So I think the anwer is clear -- 
>7-bit bit stuffing and other transmission system kludges should be a function 
>of the transmission system to solve, not a requirement laid on the object 
>representation. (I have no objection if you want to use MIME recursively to 
>accomplish those functions, however.)

A complex 8-bit MIME should be able to be put in a 7bit base64 MIME to
transport it if need be, returning it at the end to 8-bit MIME.

Robert Moskowitz
Chrysler Corporation
(810) 758-8212

• Previous message: Ned Freed: "Re: the 8-bit dilemma"