[ih] internet-history Digest, Vol 14, Issue 5

[Bob Braden]

  *> ------------------------------
  *> 
  *> Message: 2
  *> Date: Fri, 14 Apr 2006 09:01:49 -0400
  *> From: Craig Partridge <craig at aland.bbn.com>
  *> Subject: [ih] very early email question
  *> To: internet-history at postel.org
  *> Message-ID: <20060414130149.65F2867 at aland.bbn.com>
  *> 
  *> 
  *> Hi folks:
  *> 
  *> Strange question.  I've been delving into the history of email and discovered
  *> that the original SNDMSG did not spool messages or seek to retry if a remote
  *> host was down.  It simply returned an error.

Even better, for many years the UCLA-CCN site (IBM 360/91) was a
receive-only site for email.  Incoming email (originating from FTP, at
first) was simply spooled to an IBM 1403 printer with a TN train, to
get upper-lower case that was hard to find in EBCDICland.  We used our
Tenex account at ISI for sending email.

Bob Thomas would probably recall more about Tenex sndmsg, but I believe
that the mail subprotocol within FTP was supposed to provide reliable
delivery to a user.

Bob Braden


  *> 
  *> PS: While we're on this general topic.  Another unanswered question.
  *> RFC 354 (first FTP spec that looks like FTP we know) was the product of
  *> an April 1972 meeting in which one of the issues was getting email support
  *> into FTP.  RFC 354 makes clear the idea was to use the APPEND to file
  *> command (which was what SNDMSG used).  Then five weeks later RFC 385
  *> modifies 354 to define MAIL and MLFL in place of APPEND for mail support.
  *> What happened in the five weeks between the two RFCs to cause the sudden
  *> change of approach?
  *> 
  *> 

I think Jon changed his mind ;-))

Bob