[ih] Fwd: Design choices in SMTP

[Dave Crocker]

On 2/7/2023 6:06 PM, John Day via Internet-history wrote:
> mail was originally part of FTP

Just realized that this needs a bit of clarification.  Thanks to Ray 
Tomlinson, networked email originally used the Tenex CPYNET file 
transfer capability.  He linked it to the Tenex SNDMSG command.

Why this fact is more than a reference to the creation of networked mail 
and actually covers "original use" is due to the popularity of Tenex 
around the Arpanet, and the delay until email commands were added to FTP.

  * Ray did his thing at the end of 1971. I don't know the propagation
    rate it had to the rest of the community, but it was quick. (My
    involvement in the Arpanet didn't start until Spring of 1972.  I
    can't claim memories about this topic that early.)
  * RFC 458 (2/73) set the foundation for MAIL and MLFL. It appears to
    be one of the outputs from a meeting that month about FTP and
    included discussion of adding email capabilities.  But it was only a
    discussion piece.
  * RFC 475 also discusses the topic and introduces the MAIL and MLFL
    constructs.
  * Yet the Aug, 1973 FTP revision (RFC 542) still does not include MAIL
    or MLFL. In fact, it has text that says mail should be a separate
    protocol, even as it defines a mail Reply code...
  * The mail commands did not show up in the FTP specification document
    until 1980 (RFC 765)!  Although they had, of course, been in
    widespread use long before that.

While I have no memory or documentation of when these commands were 
deployed in FTP, it's clear that it was not immediately after Ray 
created networked mail, and yet email was quickly widely used. Offhand, 
I'd guess the delay for the FTP commands was in the 1-2 year range.  
Possibly longer?

Ergo, an original /use/ claim needs to cite CPYNET, not FTP.


d/



-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net
mast:@dcrocker at mastodon.social