[ih] first use of "email" (redux)

Wed Jun 5 12:11:20 PDT 2013

I can offer a little perspective on this...it's been a long time!

"Electronic mail" came in a lot of forms in the 1970s -- even telegrams, 
telex, fax, etc., but if we limit the context to "network mail" in the 
ARPANET, I think the history of FTP provides a good timeline.

I was at MIT in the "Dynamic Modelling" or DM group, from 1970 through 
1977 as grad student and then staff.  Licklider ran this group, so it 
was focussed on his interests, i.e., how to use computers to assist 
human activities.  Abhay Bhushan was also in that group.  With our shiny 
new ARPANET IMP connection and NCP, we experimented with lots of things, 
including how to use the ARPANET to do things that we had been doing 
within our own isolated machine environment previously.  One of those 
was "mail", which was simply the ability to drop a small "note" that 
another user would see when he or she next logged in.   That was a 
feature common to many operating systems of the day.  Everyone on the 
ARPANET was doing similar things.

Abhay was involved in the FTP genesis, and authored the first spec (RFC 
114) in April 1971, describing what was being done at MIT on at least 2 
machines - MIT-Multics and MIT-DM - to create a file transfer 
capability.  That capability was then used to implement crude 
"electronic mail", which worked simply by connecting to the "other" 
machine with FTP, and either STORing, or more likely APPEnding your 
"mail" to the existing mail file in the recipient's directory.  People 
quickly wrote little wrapper programs so you didn't have to interact 
directly with FTP commands, and little programs that would run when you 
logged in and see if the file had been changed since last logout.  ("You 
have new mail")  It was crude but it worked.

There was no structure to such messages -- no headers at all.  The 
message was just whatever the human sender wrote.  But people quickly 
started using standard office conventions for their mail, e.g., to show 
who it was from, what it was about, etc.  Of course there were many such 
formats, since everyone could make their own just as they did on 
typewriters.

The FTP spec was revised by Abhay in RFC 265, 17 November 1971, where it 
mentions that the minimal implementation provided the primitives 
necessary for the most common uses, i.e., it provided the "basic control 
functions necessary for basic file transfer and "mail" operations".   As 
far as I have found, this is the earliest ARPANET community reference to 
"mail", almost certainly referring to the shenanigans going on at MIT 
and elsewhere to use the new ARPANET.  "Mail" was accomplished by 
opening an FTP connection to the recipient's machine and using APPEnd to 
add your message to the recipient's mailbox file.  The FTP spec was 
further expanded and revised in RFC 354 in July 1972.

About that time, I was doing the mail implementation on MIT-DM (Lick 
*really* wanted to send mail to people at ARPA), and Ken Pogran was 
doing the Multics implementation.  MIT-DM was an ITS operating system, 
which meant it was wide open, no security at all.  Multics, OTOH, had 
very strict ideas about who was allowed to do what, and this was causing 
all sorts of problems with using those crude "mail" procedures involving 
who was allowed to write into files in someone else's space, and how to 
have a "server" process that acted on behalf of many users.

As I recall, it was easy to collar Abhay -- his office was just down the 
hall.  After sufficient badgering about the problems with using FTP to 
do these mail activities, he issued RFC 385 (August 18, 1972) which 
extended the FTP spec by the addition of two new commands "MAIL" and 
MLFL", which separated electronic mail functions from the rest of file 
transfer and manipulation.   By distinguishing those as separate 
functions from regular operations like APPEnd, it became possible to 
write programs that had different capabilities for the different 
functions -- in particular a "mail server" that could have permission to 
manipulate certain files like "mailboxes".   I view that RFC as the 
point where "mail" on the ARPANET began as a distinct service, and set 
the stage for further development of headers, mail protocols, etc., 
which occurred later through the 70s - as captured in RFCs 524, 561, 
680, 724, and finally 733.

All during the 70s, there was a *lot* of discussion about the "network 
mail" technology, but I can believe that the term "email" didn't arise 
until later.  The term "electronic mail" would of course have to arise 
before the contraction to "email".   But there were legal and political 
issues that, IMHO, forced the use of "ARPANET mail", or "network mail" 
during that period.

"Electronic mail" was a forbidden term, in the ARPANET community. Here's 
why...

In December 1980, AFIPS hosted an invitational workshop on "technical 
and policy issues in electronic mail and message systems" in Washington 
DC.   I still have a copy of the Proceedings (I was listed as an author, 
which is probably why I still have it).  There's unfortunately no 
publication number or other identification - it's a paperback 207-page 
book titled "Electronic Mail and Message Systems, Technical and Policy 
Perspectives", editted by Robert E. Kahn, Albert Vezza, and Alexander D. 
Roth, copyright 1981 by AFIPS.  It's a good resource for anyone curious 
about the non-technical perspective on electronic mail at that time, and 
also the view from people outside the ARPANET community (i.e., most of 
the world).

For example, one of the papers (by Steve Lukasik of the FCC) traces the 
origin of "electronic mail", or "electrical mail" to "the 
Post-Office-operated telegraph line between Baltimore and Washington 
starting in 1845".   Yes, email began in 1845.  Bet you didn't know 
that.   I didn't.

The paper from the White House notes: "Electronic Mail as 
'Communication' -- there is a question about whether or not this new 
service is 'communications' or is something else.   The FCC has not 
really attempted to answer that question."

So, the Rest of World knew about "Electronic Mail" but wasn't quite sure 
what it was or how it fit in to the legal and bureaucratic framework.  
Whatever was going on within the ARPANET community was just an 
experiment, not to be worried about.

Getting to the original question about "email"....

As I recall, during the early 70s, we techies naturally used the postal 
system as a model for a starting point for our own work in "electronic 
mail".  But we were also strongly encouraged to *avoid* saying anything 
that could be interpreted as working on a project to compete with or 
replace "snail mail" or any other form of established communications 
service.  Such a project would have strong domestic and international 
ramifications, with the regulations, treaties, labor relations, tariffs, 
etc., associated with interstate and international commerce, PTTs, etc., 
etc., etc.  So whatever we were doing, it wasn't that.

"Electronic mail", in those days, meant things like Western Union, 
Telex, TWX, Fax, etc., all of which were well established, heavily 
regulated, defended by stacks of laws and treaties, and entrenched.   So 
whatever we were doing couldn't possibly be "electronic mail".  It was 
far better to call it "network mail". And just wait.

IMHO, the ARPANET created "network mail".  Some time later, people 
recognized it as yet another form of "electronic mail".  Some time after 
that, they forgot about all the earlier forms like TWX etc., and "email" 
became synonymous with and subsumed "network mail".

Enjoy!

/Jack Haverty

On 06/05/2013 05:33 AM, Alex McKenzie wrote:
> On January 4, 1978 ARPA published a "Completion Report" for the 
> ARPAnet project 
> (http://www.cs.utexas.edu/users/chris/DIGITAL_ARCHIVE/ARPANET/DARPA4799.pdf). 
> In that report there are discussions of what we call email on pages 
> III-113 to III-115 and III-131.  In both instances the term "network 
> mail" was used.  This leads me to believe that the word "email" was 
> not in common use in the ARPAnet community in the second half of 1977 
> when the report was written.  So it is easy for me to believe that the 
> 1979 date given by the OED is correct.
>
> Cheers,
> Alex
>
> ------------------------------------------------------------------------
> *From:* Dave Crocker <dhc2 at dcrocker.net>
> *To:* Miles Fidelman <mfidelman at meetinghouse.net>
> *Cc:* "internet-history at postel.org" <internet-history at postel.org>
> *Sent:* Tuesday, June 4, 2013 11:23 AM
> *Subject:* Re: [ih] first use of "email" (redux)
>
> On 6/4/2013 3:44 PM, Miles Fidelman wrote:
> > Apropos recent discussions of the "inventor of email"....
> >
> > Just came across this:
> > http://public.oed.com/appeals/email/
> >
> > An appeal from the folks at the Oxford English Dictionary, seeking any
> > use of the term earlier than 1979.
>
>
> I've sent them a query, to find out their basis for citing 1979.
>
> I also just search the msggroup and header-people archives for the 1970s
> and could not find an occurrence of "email" or "e-mail". This 
> surprised me.
>
> d/
>
>
> -- 
> Dave Crocker
> Brandenburg InternetWorking
> bbiw.net
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20130605/b9095f21/attachment.htm>