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

[P E Smee]

In article <3096 at cirrusl.UUCP> Rahul Dhesi <dhesi at cirrus.COM> writes:
>In <1991Apr30.142053.2313 at sctc.com> stachour at sctc.com (Paul Stachour) writes:
>
>>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?

Mostly because of internal politics at Honeywell.  (Trust me, I worked
there.) Centered around an internal power struggle (after HIS bought
GE's large computer division) between the group that had been in
Honeywell, and the group that had been bought in with the GE purchase.

Honeywell finally took the machine off the market a couple of years ago
because it wasn't selling.  Not surprising that it wasn't selling, as
they never really tried to sell it.  HIS sales reps were, for example,
given quotas of GCOS machines to sell, but no quotas for Multics
sales.  Some sales branches didn't even know about them.  When
University of Southwestern Louisiana decided they wanted to buy a
Multics (because they used some MIT-developed courses) they called up
their local HIS sales office and said they wanted a Multics.  The sales
rep they got replied along lines of 'Multics?  Someone else must make
that.  This is HIS, we make GCOS machines.'  USL actually had to put in
a fair amount of work just to convince the salesman to check and see if
HIS did have such a product.  (In the end, they got one, but my friends
there said it was clear the sales rep didn't like the idea, because it
didn't count towards his sales quota.)

When HIS stopped producing Multics, a number of companies expressed
interest in acquiring the system source code so that they could take
over production.  (Building compatible hardware would have been fairly
easy, as the requirements were well documented.)  These included both a
small (hypothetical) startup company mooted by a friend of mine (who
had solid venture capital backing lined up), CDC I believe, and for
sure (and ironically) HIS' French 'subsidiary', CII/Bull.  For a while
HIS looked like they might go for the idea, but then they backed down
and stated that they would not release or sell rights to the source
code.

Most of the system bowels were developed under government contract and
so were in the public domain, but most of the tools, compilers,
utilities, ... were not.  The would-be takeover groups felt (probably
rightly) that they could not implement a profitable Multics replacement
in reasonable time if they had to reimplement the entire user interface
and support levels from scratch.  This last bit explains why the
concepts still show up everywhere, but the system itself does not.

A lot of us feel that having to move from Multics to anything else 
represents a giant step backwards.  Unix is just about OK.  (There
are, in fact, SOME things that Unix does better.  There are a very
large number where it falls short.  My current annoyance is with the
limitations on Unix access control, but that's another story.)

(Followups directed to alt.folklore.computers.)
-- 
Paul Smee, Computing Service, University of Bristol, Bristol BS8 1UD, UK
 P.Smee at bristol.ac.uk - ..!uunet!ukc!bsmail!p.smee - Tel +44 272 303132