[mailhist-discuss] First email attachment?

Dave Crocker dhc at dcrocker.net
Sun Mar 12 15:48:18 PDT 2017

On 3/12/2017 3:42 PM, Jack Haverty wrote:
> Dave,
> I'm not sure if it fits your definition of "attachment", but the COMSYS
> mail system at MIT in 1976 (and probably a bit earlier) had a notion of
> "Enclosures" which enabled large or non-textual content to be included
> in a message outside of the body of the text itself.  The message would
> include a pointer to a file containing the content to be enclosed, which
> the receiving system could retrieve by the local file system, or
> mechanism such as FTP, if and when the recipient wanted it.

ewww.  Forgot the scenario of external references (though MIME included 
the construct from the start.)

Mumble.  If a convention for citing/including external references does 
not count as an 'attachment' then we'd have to define yet-another 
category of construct and I think it's not necessary.  So, yeah, that 
probably counts.

1976?  Cool!

> So the main text message would be delivered by COMSYS, and the enclosure
> would be retrieved later by the recipient.
> Since there wasn't any protocol defined yet for attachments, messages
> sent to recipients outside of the COMSYS world got the enclosure as part
> of the message body.

Was there any distinctive/standard packaging and/or encoding?

> That behavior allowed us to conveniently send large documents in
> internal email.  It also led to a few complaints.  One episode was
> captured in the HEADER-PEOPLE archives, when I had just sent something
> large (a ~65,000 character document, Oh My!) that caused other people's
> mail systems to choke.  See below.

Yeah, when author's can't feel the effect of a large message body and 
it's convenient to create them... they do.

Thanks for the reference and example!


Dave Crocker
Brandenburg InternetWorking

More information about the mailhist-discuss mailing list