Why not Multics? (was Re: BSD tty security, part 3: How to Fix It)

[Eric Lee Green]

>From article <3096 at cirrusl.UUCP>, by dhesi%cirrusl at oliveb.ATC.olivetti.com (Rahul Dhesi):
>>Gee, with that kind of understandings, its no wonder that those of
>>us who have used Multics are kind of upset when we are forced to migrate
>>to systems where [non-Multics things happen].
>
> So why are we all using UNIX and its derivatives?  Why isn't Multics a
> commercial success even though it seems to have a unique place in
> history?
>
> More specifically, where can we buy Multics to run on our favorite
> hardware?  Why can't we buy it?

Multics suffered from Honeywell Brain Damage. First of all, by the time the
commercial implementation was released, the hardware was totally obsolete
(the hardware being a late 60's GE mainframe with a bunch of bags hung on
the side to implement the segmentation and ring security). Worse
than obsolete, it was also very expensive. Second, Honeywell didn't know
what to do with the Multics OS once it was completed. Most of their
customers were already using GECOS on the un-bagged version of the
mainframe, had all their aps already running under that environment, and
saw no reason to change. And, finally, Honeywell was in business to sell
hardware, not operating systems. Unlike Bell Labs, which really wasn't in
business at all, much less the hardware business. Thus Multics was
intimately tied to Honeywell's hardware, to the point where many portions
of the system would munge on pieces of 80-bit pointers or, for that matter,
were written in ALM (Multics assembly language, a truly horrendous beast...
I seem to recall that it had more than 500 instructions, dealing with all
sorts of bags on the side ranging from the address unit to the decimal
unit). Multics wouldn't run on non-Honeywell hardware, unlike Unix, which
was a multi-machine OS almost from the beginning. When Honeywell did the
crash-and-burn with their Multics marketing, there wasn't anybody around to
take up the slack. No Sun Microsystems equivalent to turn Multics into
something ubiquitous in its market segment (like Sun did for Unix and
workstations).

There did exist some other problems, of course. For one thing, some aspects
of the Multics design were inherently less efficient than "normal" design
practices (things such as the dynamic linking, where the first run of a
program would produce a whole lot of traps to pull in routines from other
segments). The OS was big, for another thing (one reason why it was so late
to be released), and somewhat resource-hungry for its day (though "X" puts
it to shame any time of the day). The PL/1 compiler did a decent job of
optimizing for its day, but the underlying architecture was the pits to
begin with (a single accumulator? sheesh!). And of course the whole OS was
designed in the late 60's and early 70's, and it showed... there were some
half-hearted attempts at MIT and elsewhere to bag on support for graphics
terminals and such, but a real professional job of it was never
accomplished. About the only really user-friendly software that ran on the
machine was an excellent version of Emacs written at MIT, as far as I know
the first version of Emacs that integrated Lisp into the editor (actually,
it was a compiled Lisp and PL/1 program that dynamically linked to the
MacLisp interpreter). Unfortunately, the antiquated Multics front-end
hardware was so inefficient at handling user interrupts that running Emacs
slowed terminal I/O to a crawl... the Multics front-end hardware really was
designed for an era in which complete lines were typed in on a teletype and
then processed upon hitting RETURN.

Datedness and poor marketing were what eventually did Multics in. Both are
probably attributable to Honeywell's befuddlement at what to do with an OS
that didn't "fit in" with their existing lineup. One can assume that they
would have eventually updated both the hardware and the OS if Multics had
turned out to be a "best seller", or, for that matter, if they'd ever come
up with a marketing slant on what to do with the silly machine. As it
turned out, they sold a few to the DoD (it was the most secure system in
the world for quite some time, perhaps still is), a few to universities
here and there, and that was it.

Multics-inspired derivatives live on here and there. A friend describes
Primos as "what Multics would have looked like if designed in the USSR".
The folks at Stratus (fault-tolerant computing) made their system look a
lot like Multics. I seem to recall that Apollo's Domain OS stole a few
things here and there from Multics, also. But Multics in its original
conception seems to be dead, period. I recall hearing rumors that someone
was going to try to port it to a non-Honeywell machine, several years ago.
As far as I know, those rumors have gone nowhere. There's few current
architectures which could adequately handle Multics in its full segmented
and ring'ed glory, anyhow... the '386/486 family come to mind as the most
obvious.

--
Eric Lee Green   (318) 984-1820  P.O. Box 92191  Lafayette, LA 70509
elg at elgamy.RAIDERNET.COM               uunet!mjbtn!raider!elgamy!elg
 Looking for a job... tips, leads appreciated... inquire within...