[mailhist-discuss] Definition of email

Jack Haverty jack at 3kitty.org
Tue Jun 5 12:01:58 PDT 2012

In the early 70s, there were probably a hundred or so people involved
in the research of "Computer Assisted Human Communications", basically
brainstorming about how interconnected computers might be applied to
human communications.  At the time, there was rudimentary human-human
communications inside single computers.  Various contemporary
timesharing systems had "link" facilities.  Even back in the
prehistoric days of card punches and batch jobs on mainframes in the
60s, it was common to send a message (by using a punchcard!), e.g., to
ask the operator to mount a tape, and sometimes even get a reply (in
the printer output) that he couldn't find it.  (Been there...)

In that 70s context, thoughts coalesced around the concept of taking
an old, well-established human communications service and migrating it
into the then-new interconnected-computer world.    Postal mail had
been around for several centuries and was well-established as a
worldwide service.  Other existing services, e.g., telephone or
telegraph or radio also provided human-human communications.  Wire and
radio telegrams had already implemented an electronic alternative to
postal mail by early in the 20th century.  But postal mail had been
around since at least the early 19th century, and was arguably the
most pervasive form of human communications over distances.  It was
not the first such service - Greek and Roman runners delivered a lot
of scrolls thousands of years earlier.

I recall that the postal mail system became a kind of prototype for
doing something new with computers and networks.  When we talked about
new ideas, we used terminology from the postal world.  The various
discussions congealed on an unwritten goal of using the new technology
to create a service modelled after the postal service, with the main
difference being the electronic transportation of the contents rather
than physical.

There may not be any universally accepted definition of "mail", but
everyone had sent and received letters and knew what the postal
service was.  So, initially there was discussion of "envelopes" and
other related postal terms.  Later, terms such as "headers" took over.

It was natural to call such a new service "electronic mail", simply
reflecting that it was "mail", but done electronically.  I have no
idea who first spoke those words.

So, IMHO, the definition of "electronic mail" should simply be a
derivation from the definition of "mail".   Mail of course has evolved
over the centuries, but the early days of "mail", e.g., the early
1800s, could serve as a good baseline definition.  Letters had to be
light enough to be carried on horseback, they were addressed to a
single person, the address specified a geographic location, etc.

So, IMHO, "electronic mail" is a service which provides at least the
basic services of 1800s-era postal mail, using electronic mechanisms
rather than physical ones.

Of course, this then requires a definition of "mail", which delineates
and defines the characteristics of mail service.  After several
centuries, maybe such a definition of mail service has been
written...?  Elctronic mail is simply providing the analogous service
but through electronic means.

In thinking about this, it occurred to me that there are fundamental
aspects of 1800s Mail Service that don't yet exist in "electronic
mail".   For centuries one can send postal mail to another person and
be very confident that it won't be viewed or altered in transit; not
true in email.  Second, the basic concept of a "stamp" serves in
postal mail to partially fund the system, and limit abuse of the
service by constraining usage; our free email service has permitted
spam to flourish.  Third, mail service usually is provided by
governments, and is surrounded by a mesh of legal and enforcement
mechanisms;  email service is provided by ... well, that's a good
question.  There are more.

Someday, electronic mail might be invented.  It hasn't happened yet.


On Mon, Jun 4, 2012 at 1:18 PM, Thomas Haigh <thaigh at computer.org> wrote:
> As Dave Crocker mentioned, when I was preparing my article on the invention
> of email for the Washington Post’s website it seemed like a good idea to
> attempt a definition of the necessary and sufficient characteristics of an
> electronic mail system. This wouldn’t attempt to capture the features
> present in all or most modern systems, but to pin down the essence of what
> we mean by electronic mail.
> So the tricky thing was to be as broad as possible while excluding things
> like faxes, Twitter, newsgroups, electronic bulletin boards, etc. which
> would not generally be thought of as electronic mail. Most of those were
> never thought of as electronic mail – the exception being fax, as early
> usage (through the 1970s) sometimes did include things like electronically
> routed telexes and faxes.
> People providing comments during the discussion included Dave Walden, Dave
> Croker, Ray Tomlinson, Tom Van Vleck and a couple of historians. However,
> the definition is mine and I do remember that Dave Crocker favored a simpler
> one.
> Definition
> Electronic mail is a service provided by computer programs to send
> unstructured textual messages of about the same length as paper letters from
> the account of one user to recipients' personal electronic mailboxes, where
> they are stored for later retrieval.
> What each phase is doing there and why we need to keep it:
> 1.      “Electronic mail is a service”: puts the stress on the system
> itself, not individual computer programs
> 2.      “provided by computer programs”: May not be necessary, I think was
> there to focus the definition on computer systems rather than something like
> fax.
> 3.      “to send unstructured textual messages”: unstructured separates this
> from earlier network communication where highly structured data was
> exchanged within systems like SAGE. Of course email text can be highly
> structured, but any system than can transmit an unstructured textual message
> can also transmit a structured one. The reverse is not true. Finally,
> “textual” further distinguishes electronic mail from fax, image
> transmission, etc.
> 4.      “of about the same length as paper letters”: we went backward and
> forward on whether to say instead something like “able to handle messages of
> at least 2,000 words” or “of unlimited length.” Realistically, any
> electronic mail system (particularly an early one) has a limit, but the
> point is that it must be large enough to permit transmission of the
> equivalent of a long letter. This distinguishes electronic mail from
> something like SMS.
> 5.      “from the account of one user”: electronic mail always has a single
> sender (whether or not the sender is identified to the recipient) and comes
> from an account rather than a human (though accounts are often personal).
> Also, someone needs an account to send a message, whereas with paper mail
> she only need a stamp.
> 6.      “to recipients' personal electronic mailboxes”: we thought about
> saying private or secure, but in reality they aren’t always. However, the
> message does go to the mailbox of one or more recipients and is shown only
> to them. This distinguishes electronic mail from something like a post to a
> bulletin board which can be seen by all users.
> 7.      “where they are stored for later retrieval.” i.e. asynchronous,
> which separates electronic mail from chat.
> The definition would include private message functions within conferencing
> systems, which I think is perfectly reasonable – they’re basically
> electronic mail capabilities within a larger system.
> Finally, though we did not say it explicitly here, we follow almost everyone
> else in the world by treating email and e-mail as contractions of electronic
> mail with the same meaning.
> Tom
> _______________________________________________
> mailhist-discuss mailing list
> mailhist-discuss at emailhistory.org
> http://emailhistory.org/mailman/listinfo/mailhist-discuss

More information about the mailhist-discuss mailing list