In article <
9d4f092b-b466-4ed1...@googlegroups.com>,
Robert Carnegie <
rja.ca...@excite.com> wrote:
>On Friday, 15 January 2016 17:34:49 UTC, Cryptoengineer wrote:
>> Back in the early days of the internet, before the web, usenet and
>> mailing lists reigned supreme. One of the popular mailing list
>> management programs was 'listserv'. SF-Lovers (the ancestor of
>> this group) ran on listserv.
>I don't know that particular history, but I have used
>e-mail groups where the control messages (subscribe,
>unsubscribe, digest) were to be sent not to the group
>address, but IIRC a "majordomo" address - but getting
>/that/ wrong was quite common, and looked about the same.
LISTSERV predates Majordomo. LISTSERV was an IBM mainframe program,
and it originally was used primarily on BITNET, an early
store-and-forward network that was built on top of leased lines
sending virtual punched-card decks over leased telephone lines
connecting universities. (Among other things, it was infamous for
including in its responses a summary of the CPU and I/O resources used
in responding to the request.) (If I remember correctly, the guy who
wrote it originally was a systems programmer for FRECP1 at the Ecole
Centrale de Paris.) The BITNET email system was simply an IBM email
product that already existed and was used internally to IBM and at
the few commercial IBM customers that were large enough to have more
than one computer.
The ARPANET convention, which predated mailing-list manager software,
was to send requests to <Listname-Request@List-Host>. (Like BITNET,
pre-DNS ARPANET had non-hierarchical host names and a globally
distributed host address file, so MIT-MC was a machine owned by the
Macsyma Consortium at MIT[1], SU-AI was the machine at the Stanford
University Artificial Intelligence lab, ISI-B was the second computer
at Information Sciences Institute, and WSMR-EMH1 was Electronic Mail
Host number 1 at White Sands Missile Range.) People were expected to
know this convention (after all, they must have figured it out to get
on the list in the first place) and were sometimes roundly flamed for
sending "administrivia" to the list address rather than the -request
address.
There are people around who were actually there for this history; I
started just about the time the ARPANET was finally being shut off,
having been obsoleted by NSFNET, although MILNET continued to exist,
with many old machines continuing at their same names (and
26.host.0.imp IP addresses) until they were finally decommissioned,
like the famous WSMR-SIMTEL20 and WRAIR-EMH1, just to name a few.
-GAWollman
[1] The local names at MIT were mostly two letters: at various times
there were machines called AI, DM, MC, ML, MX, OZ, and XX, all PDP-10s
except for one PDP-6. The people who weren't in 36-bit land tended to
prefer longer names, so MIT-EDDIE was a Unix machine in the EECS
department (which still exists, several hardware generations removed),
and MIT-MULTICS was the production Multics machine. Lisp machines had
longer names still, but of course their users almost never had to type
them because the lispm operating systems had (have!) autocomplete.
--
Garrett A. Wollman | What intellectual phenomenon can be older, or more oft
wol...@bimajority.org| repeated, than the story of a large research program
Opinions not shared by| that impaled itself upon a false central assumption
my employers. | accepted by all practitioners? - S.J. Gould, 1993