ietf-nntp Re: draft-codogno-mime-nntp8bit-00.txt

[Brad Templeton]

On Tue, Aug 11, 1998 at 12:14:44PM +0200, Maurizio Codogno wrote:
> Dear nntpext and usefor people,
> 
> In June I suggested in usefor to define a more efficient way to send
> binary files via Usenet News - after all, NNTP is defined on a
> 8bit channel, and the only characters not allowed in the body of 
> a message are NUL and bare CR and LF. 
> 
> Eventually I prepared a draft, which has been issued with the name
> draft-codogno-mime-nntp8bit-00.txt, which defines a new MIME type,
> application/nntp8bit.  I am already told that I chose the
> wrong path: such an extension cannot be a new type, but rather a 
> Content-Transfer-Encoding. This means that after Chicago I'll have
> to rewrite the draft...
> 
> May I however ask any people interested in the idea to read the current 
> draft anyway and to send me suggestions? I'll be in Chicago, by the way.
> > ciao, .mau.

Well, on one hand there is a strong push for no more content transfer
encodings.    But there are two obvious missing slots, the one you defined
there, and of course raw binary itself.

Expect a lot of resistance.   Defining a new content-transfer-encoding means
that a vast variety of different software packages suddenly have to
understand it.  If, for example, we end up signing the canonical form
of an article rather than encoded form, every signature checker, which means
all news database systems (or all readers in some scenaria) have to
understand it.