Re: smtp charter (revised)

[David Herron <david@xxxxxxx>]

> For the record, I support Stef's suggestion for a new port number for 8-bit
> SMTP.

Technically I agree.  I wonder (aloud) the difference between accepting
SMTP++ on a new port number, versus the acceptance of X.400.  Both
are a "new" protocol, admittedly SMTP++ has the advantage of similarity
to existing practice and easy implementation as opposed to X.400
which is very different & much harder to implement. 




> I would support any SMTP extension or modification proposal that involves a
> new port number.  I recommend (at least) the following changes in new SMTP:
>  1) 8 bit instead of 7 bit character set
>  2) no "lines" in message.  Everything between "DATA<cr><lf>" and
>     "<cr><lf>.<cr><lf>" is a byte stream that is transferred in image form.
>     This removes both line length and non-binary restrictions.

My first thought was:  But Wait!  What if those 5 characters are in
the data?!?  Almost immediate answer:  stuff in some DLE type characters?

>  3) Explicit batching support; definition of server states when a command
>     fails.

What do you mean by "batching"?  There are at least two things you
can mean:

1) BSMTP.  That is, bundle up the sender half of the conversation
   into a file and push this through a file-transfer-protocol as
   one bundle to the recieving process.  Useful for networks where
   file-transfer is the unit of interaction (e.g. BITNET and UUCP).

2) Accumulate the "MAIL FROM:<>" and "RCPT TO:<>" commands into
   one group to be sent across the wire at one time.  Useful
   in cutting down on the number of round trips required to send
   a message with a lot of recipients.



	David