[mailhist-discuss] Origin of SMTP ?

Jack Haverty jack at 3kitty.org
Thu Apr 23 17:18:45 PDT 2015


Hi Dave et al,

IMHO, it's important to remember the culture of that era.  In
particular, RFCs often lagged well behind whatever work they were
documenting.   So the notion that what is documented in RFCxxx was
derived from RFCyyy may be misleading.

If the goal is to reconstruct the history that occurred between MTP and
SMTP, I think the relevant mailing lists would be much closer to showing
the true picture.  In particular, MSGGRP and HEADER-PEOPLE were very
active throughout that era.

My personal recollection is that there were multiple efforts to define a
next generation mail "protocol", although "architecture" might more
accurately describe the thinking.   MTP was one of these.   Somewhere
back in those days, I defined one called MSDTP (see the RFC) which was
very powerful but very different.   IIRC, the conclusion was that all of
those very advanced designs were too much of a change.  They would
require a lot of code.  They would consume a lot of the scarce memory
and processing power.  

So a "Simple" next step was needed that made just a small change to
solve the most immediate problems at that time (which of course I can't
recall).   I'd be more inclined to think that SMTP "derived from" all of
the prior work, both what was implemented broadly and what was just in
proposal form, and that knowledge was used to define SMTP as a small
incremental step in the right direction.  Saying that SMTP did or did
not "derive from" MTP oversimplifies the community interactions that
were being conducted outside of the RFC process per se.

I still can't remember where the "Simple" adjective was to be applied in
going from MTP to SMTP.   Perhaps to handle "simple" mail - as opposed
to compound documents, multimedia, verified-delivery,
deliver-only-to-addressee, etc?   Or "simple" protocol (limited set of
functions supported)?   Or Simple transmission (direct communication not
involving intermediate nodes, translation services, etc.)?  Maybe all of
them.

IMHO, the same thing occurred with network management.   SNMP is Simple
NMP.   At the time, we had grandiose ideas, but the realities of limited
resources (people, memory, processor, line capacity, etc.) motivated
small steps and simplification.

The answers are probably in whatever archives still may exist of all the
interactions on the relevant mailing lists.

/Jack Haverty


On 04/23/2015 01:26 PM, Dave Crocker wrote:
> Greetings, oh quiescent email history list.
>
> Just saw something that prompts a question:
>
>    Would you say that SMTP derived from RFC 780 (MTP)?
>
> There were a number of failed efforts that preceded SMTP and while I
> understanding thinking that things derived from RFC 780, it does not
> match my recollection, other than Jon Postel's involvement in both.  I
> remember the specification meetings, but don't remember being
> particularly influenced by 780.
>
> I'm asking after technical relationship, not about the involvement of
> individuals or even the development process.
>
> Thoughts?
>
> d/



More information about the mailhist-discuss mailing list