[mailhist-discuss] Definition of email

Dave Crocker dcrocker at gmail.com
Mon Jun 11 12:08:00 PDT 2012


Good exchange.  Let's press for resolution...

On 6/7/2012 12:22 PM, Ray Tomlinson wrote:
> On 6/7/2012 11:06 AM, Terry Gray wrote:
>> Electronic mail (Email) is a service provided by computer programs to
>> send  messages from the user of one computer account to the personal
>> mailbox(es) of others for later retrieval.  An electronic mail message
>> consists of packaged data, analogous to a physical letter, specifying
>> one or more authors and one or more recipients.
> Getting close...
>
> Perhaps "computer account" is not relevant.

While I see the careful precision it provides, I don't think it's 
essential and I suspect most readers would find it distracting. 
(Implicit goal:  make the definition reasonably comfortable for use by 
the mass market of readers, not just technicians.)

I think a single reference to computers suffices.  Something like: 
email is a computer service to send messages...


>  A computer with no login required and an
> array of mailboxes named by their respective users would work just fine.

Well, yeah, that's an entirely pragmatic reason for dropping the qualifier.

Separately, the problem with saying "personal" mailboxes is that it 
excludes role and other mailboxes such as mailing list redistributors. 
mailhist-discuss at emailhistory.org is not a personal mailbox.  Does it 
not receive and send email?


> Maybe we can find a better phrase than "packaged data", too.

"content" ?

However the deeper problem is the "analogous to a physical letter". 
This invites debate about a physical letter.  That is, is redirects, 
rather than resolves, basic questions. I suggest that our definition 
stand on its own rather than rely on (mis)understandings of another service.


> Early versions of email (CTSS, SNDMSG) did not identify the recipient;
> if the message was in your mailbox you were a recipient.

If the recipient isn't identified, how can the system know who to 
deliver the message to?

That is, there's a difference between tagging a /display/ list for 
recipients, versus "indicating" recipients of the message.  The text I 
originally suggested is meant to say the latter, without commenting on 
the former.


> I think a slightly reworded version of the first sentence above captures
> the essence of email.
>
> Electronic mail (Email) is a service that allows a computer user to send
> messages to the mailboxes of other computer users for later reading.

It doesn't clarify what "message" means.  While it's been easy to ignore 
that precision over the years, I'd say the recent public claims and 
their constraints have made it worth elaborating.

Also, these latest drafts have dropped the concept of a message's having 
multiple authors.  Again while that is rare, it's legal and is still 
sometimes done.


> Anything beyond that needs a justification. Even "computer user" is open
> to misinterpretation. It implies a person using a computer, but allows
> Amazon to automatically send order status messages if you think the
> daemon that does the sending is a computer user.

Automated sending or receiving -- as well as 'role' addresses - do raise 
the question about "user".  The usual way to dodge it is to work in 
something like "on behalf of".


  The nature of the
> message is a particularly slippery slope; almost any statement here will
> have numerous exceptions. Is it text? No. Are the recipients named?

What does "named" mean here?




On 6/7/2012 3:11 PM, Jack Haverty wrote:
 > Hi Ray!   Been a long time....
 > In writing a definition,  we're describing a line-in-the-sand,
 > separating what's included from what's excluded.

+1


 > I would tend to exclude voice-mail systems of any vintage.

Just to be clear:  why?


 > Similarly,  I'd tend to want to exclude bulletin-board type of
 > systems,   where a message is sent to a group of users where the
 > sender doesn't know exactly who will receive it.

How is that different from what is happening with 
mailhist-discuss at emailhistory.org?  Is this not email?


 > IMHO:
 >
 > Electronic mail (Email) is a service that allows a computer user to
 > send messages to the personal mailboxes of other computer users for
 > later reading, as an electronic alternative to paper messages.

Again:  I believe that language like "as an electronic alternative to 
paper messages" is useful for pedagogy but poor for denotation.;


So, to offer a concrete alternate version:

      Electronic mail (Email) is a computer-based service for sending
messages from one author to the online mailboxes of one or more 
recipients, for later retrieval.  The content of an email may contain 
any sort of information.


Open question:  MCI Mail permitted sending to postal addresses as well 
as online mailboxes.  Weren't those emails?  Does the use of gateways to 
external representations make these no longer emails???

Also, I'm still not clear that this covers role addresses, nevermind 
automated mail creation or mailing list forwarders.

d/
-- 
  Dave Crocker
  bbiw.net


More information about the mailhist-discuss mailing list