[ih] SMTP History

[Jack Haverty]

The RFCs, IMHO, are useful but only as an imperfect historical record.   
At the time, RFCs, despite their name as "request for comments",  were 
often released well after implementation of whatever they documented, 
when someone (often Jon Postel) took the initiative to write down what 
had happened.

So, with something like FTP, despite what the "official spec" might have 
said in some RFC at the time, it was easy for an actual developer to add 
new functionality and try it out, possibly collaborating with others.   
Adding a MAIL command would have been easy for any FTP developer, and 
invisible to other FTPs who hadn't ever heard of it.   Such things were 
expected as a vital aspect of research.   Ray's introduction of @ on 
TENEX propagated quickly since there were numerous TENEX machines on the 
net at the time.

I don't remember who first added the MAIL command to an FTP 
implementation (might have been Abhay...?), or how its use propagated 
throughout the ARPANET community as other developers added it to their 
FTP software.  It was easy, and common, for developers to see someone 
else's idea and simply adopt it, well before it appeared in any RFC.

I do remember that I first used, and insisted on including, the 
"Message-ID" header field as a way of making it possible for a computer 
program to distinguish specific messages, so that, for example, they 
could be linked together into conversations.  But most details of the 
SMTP and headers just sort of appeared over time and if people found 
them useful, they got more widely implemented.   I have no recollection 
of where they started.

Eventually an innovation could become popular enough that it became 
effectively "standard".   It might only be after that occurred that it 
actually appeared captured in an RFC.   Rough consensus and running code 
came first.   Documentation later.

I wrote RFC722 in an attempt to document some of the issues and 
principles involved in that kind of evolutionary development of network 
mechanisms.  Never got many comments thought to that RFC.

IMHO, the email records such as Noel captured are likely a more 
accurate, but still incomplete, historical record.   But there were lots 
of such email interactions not using email lists, or even, gasp, by 
telephone or in-person discussions, probably now lost, that reflect the 
actual history of what happened when and who did it.

That's why I try to write down just what I personally remember.   My own 
"mail archives" were lost long ago, when my Dectapes turned into 
magnetic dust.

Jack

On 3/28/22 14:24, Dave Crocker wrote:
> On 3/28/2022 2:07 PM, Jack Haverty via Internet-history wrote:
>> There are artifacts in the RFCs capturing some of the early work. FTP 
>> began circa 1971 with RFC172.  At the same time, there was discussion 
>> of a "Mail Box Protocol" intended to enable functions like remote 
>> printing as a way of sending something to someone else over the 
>> ARPANET.   You just send it to their printer.   See RFCs 196, 221.
>>
>> At first, FTP added a "MAIL <user>" command, which each machine 
>> receiving such MAIL could process as it saw fit.  Print it out.
>
>
> RFC 354 (July 1972 and edited by Abhay Bhushan) does not contain the 
> string 'mail'.
>
> RFC 475 (March, 1973 and edited by Abhay Bhushan) discusses FTP's MAIL 
> and MLFL commands. It is a meeting report discussing agreement to 
> create those commands.
>
> RFC 542 (August 1973 and edited by Nancy Neigus) does not contain the 
> string 'mail'.
>
> RFC 765 (Aug, 1973 and edit by Jon Postel) does. But while is cites a 
> mail command, it does not specify it.
>
>
> d/
>