Email delivery protocols / methods.

[Paul Koning]

> On Jul 8, 2019, at 11:27 PM, Grant Taylor via cctalk <cctalk at classiccmp.org> wrote:
> 
> ...
> On 7/6/19 12:57 PM, Paul Koning via cctalk wrote:
>> There's the MAIL-11 protocol (end to end, no MTAs) and the DECmail protocol which may be some OSI-like thing, I'm not sure anymore.
> 
> I guess I don't know enough about MAIL-11 to understand why you say end-to-end / no MTA.

No mail servers.  You address mail to node::user and it contacts the mail protocol listener at that node, which drops the message into the mailbox of that user on that system.  

> Was DECmail the OSI X.400 email implementation that DEC produced (I think) in the '90s?

Yes, in ALL-IN-ONE and the like.  An interesting point is that it was not really accepted as the internal mail tool (except by some corporate overhead departments); engineering persisted in using MAIL-11 based email on the internal network.

>> For real strangeness there is the PLATO mail protocol, which involves writing the mail into files, which are then extracted from PLATO into the OS file system by a periodic batch job, then sent to another system via file transfer (FTP or a predecessor), then pushed into the PLATO file system, then picked up by a mail agent at that end.  Ugh.
> 
> $ReadingList++

You're unlikely to find documentation for this, unfortunately.  It's part of the "linked systems" capability of PLATO, a loosely connected collection of systems which could exchange email, notes (as in Lotus Notes, which goes back to a very popular PLATO tool) and, in a very limited way, files.  It used very strange custom network hardware originally, and eventually moved to TCP/IP under CDCnet.

	paul