[ih] Email and address books

[paul vixie]

On 5/24/2012 8:29 AM, Nigel Roberts wrote:
> I worked for the ALL-IN-1 development team in the mid-80s.
> ...
> And sending email to and from the wider Internet was actually easier
> than Paul remembers.
> ...
> There were various gateways between Message Router (the email backend
> of ALL-IN-1) ...
>
> But what I uses was am ALL-IN-1 feature that sent VAXmail from inside
> ALL-IN-1 (my memory is failing as to exactly how, but I remember
> hacking something in the DCL code that actually implemented it. I seem
> to recall that it may have been prefixing the address with '_'
>
> So to email the Internet from ALL-IN-1 was as "simple" as sending it to
>
> _RHEA::DECWRL::"user at host.tld"
>
> You can see what the return address would have looked like to the
> Internet user at
>
> http://catless.ncl.ac.uk/Risks/8.4.html#subj1.1

that's:

Nigel <roberts%untadh.DEC at decwrl.dec.com>

/Wed, 11 Jan 89 03:02:40 PST

/
> This shows a reply address that would translated to an internal
> Easynet address in the form UNTADH::ROBERTS which typically would be
> automatically forwarded into Message Router/ALL-IN-1 by the ALL-IN-1
> user having done a $MAIL SET FORWARD.
>
> :-)

so, from 1988 to 1993 i ran DECWRL:: (as the rest of the company knew
our DECnet Phase IV node name) and i was using a public-domain version
of the mail11 gateway code (originally written by keith moore, then at
UTK) which i'd hacked to do "block mode" since the decnet-savvy people
in the company were really pissy about "line mode" which apparently did
one "line" per round trip (no pipelining). i'd by this time thrown away
the Ultrix version of sendmail, restarted from berkeley sendmail, which
i ended up publishing as "King James Sendmail" because eric was at that
time still at britton-lee software, and sendmail was in its walkabout
phase (before Sendmail, Inc. was started). KJS included lennart
lovstrand's most wonderous "IDA Sendmail" hacks, which allowed for
arbitrary "db" lookups from within rulesets, so that i could do UUCP
routing based on a pathalias database rather than having "rmail" do it.

i also ended up throwing away all the m4-based sendmail.cf generation
logic from both berkeley and ultrix, and starting from my own private
.cf file.

all of this got written up by fred avolio and i in the immortal classic,
"Sendmail: Theory and Practice", amazingly still in print at:

http://www.amazon.com/Sendmail-Second-Edition-Theory-Practice/dp/155558229X

in 1992 or so i convinced my various bosses (mostly this was brian reid,
since i think dave crocker was gone by then) to let me buy a VAX 5400
running vax/vms so that i could run a proper "DIGITAL ALL-IN-1" mail
gateway instead of relying on the underbar hack you're describing here.
that was node name WRLMTS:: and you knew us as @WRL. this effort failed
miserably because i wasn't as able to hack the software on VMS as i had
been on Ultrix (which was really just a sad old cut of BSD).

so, you're welcome, DECWRL was happy as heck to carry all your e-mail to
and from the internet. but from talking to ALL-IN-1 customers outside
the company, who had to use DEC's own products to do this kind of thing,
they couldn't make it work any better than i could. thus my comment,
"weeping and wailing".

:-)

paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20120524/0eb837a0/attachment.htm>