Interview with Sam Morgan.
MSM: You were saying that
there was some effort to...sometime ago to portray UNIX as having
been management inspired. You were about to say but that was
not so.
Morgan: It was not so and
in the way that the question originally came to me. And why don't
we get to that in the proper course of things.
MSM: Okay
Morgan: Let me ask you,
whom have you talked to already about UNIX? Uh, some, some set
of people not all of whom are right around this immediate corridor
that were in on the early part of the work and would have interesting
things to say. Who, who have you seen?
MSM: Well I have talked
to Doug, talked to Ken Thompson and talked to Sandy Fraser, Peter
Weinberger, Brian Kernighan. I am suppose to see Condon and Ritchie
tomorrow.
Morgan: Ritchie, Ritchie
is important as you know.
MSM: Yes. Let me see who
else is in my book. Lorinda Cherry and Al Aho on Friday I am
going to talk to. And then I think we have another interview,
and I have a list of other people that I want to talk to. Basically
one of the lists I have been using is the list of people that
Doug put together in that little piece he did on the early manual.
Identifying where pieces had come from and then I got some other
names. Ted Dolotta gave me a call when he heard about this project
cause I knew Ted from swimming.
Morgan: Okay, a person who
you certainly ought see is Berkeley Tague. The reason I say that
is well Berkeley is an articulate and philosophical person. But
he was involved from very early on in the UNIX support group.
When, when UNIX got to be big enough and interesting enough to
the organizations at Bell Labs outside research Tague put together
a support group that gradually took over the management and extension
of UNIX as a software base for a lot of development efforts.
And I think it was in 1982, um, Tague put together a fairly long
talk on the history of UNIX up to till then as he had seen it.
And I thought it was both entertaining and quite perceptive.
Tague had during, Tague is now a department head in the Murray
Hill comp Center. He can be found in the book and his name would
be suggested to you anyway I think by someone. But he did make
some effort to do a one man's history of UNIX. Which I thought
was quite effective. And so you certainly want to talk to him
as to how he saw it because his viewpoint was a little different
from the viewpoint of the research organization. But not so different
that you couldn't see some similarities and he did considerable
work in 1982 on putting the record together. So he is somebody
you ought to see.
MSM: Good. What was your
position here at the time this UNIX project was done?
Morgan: I was Director of
Computing Science Research. The same position that Al Aho holds
now and that Sandy Fraser held after I did. Up until a couple
of years ago. I was director of the organization which UNIX was
done.
MSM: Were you director
during the period of the MULTICS department.
Morgan: I became director
in 1967 and MULTICS was a project that was then ongoing. MULTICS
had been presented to the Bell Laboratories computer using community
as a kind of computer utility. As an operating system that would
be a major advance over the then existing mostly batch oriented
operating systems and this was to run on GE 645 computers. And
people users had been told don't make further major commitments
to developing software for the then existing IBM700-7000 series.
Uh, wait, uh, MULTICS which is being done as a joint effort
by Bell Labs, General Electric and MIT, uh, this will be the
operating system of the future. At the time MULTICS began to
be presented to computer users in that style I was elsewhere and
was a computer user myself. By the time I became Director of
Computing Science Research there were beginning to be doubts about
MULTICS. It was, the development was clearly moving more slowly
then had been expected. And users were sighing with varying degrees
of pungency, "We can't wait for this." But the MULTICS
development effort as far as Bell Labs was concerned went on for
about two years after I became director. And then Bell Labs declared
that it's commitment to the MULTICS development effort had been
fulfilled. This was made, this decision was made early in 1969
and Bell Labs work on MULTICS was terminated I believe at the
end of March of 1969. So I was in on the last part of the MULTICS
efforts so far as Bell Labs was concerned. And UNIX was born
late in 1969 as you probably have heard and developed in research
for awhile and then eventually spread.
MSM: I'm interested if
you can you tell a little about that of the MULTICS project.
Realizing that you had, that you came into your position while
that was going on. But you did I guess that MULTICS, there were
promises associated with MULTICS.
Morgan: There were promises
associated with MULTICS that were not fulfilled on a scale and
in a time period that would have made MULTICS generally useful
to the Bell Laboratories computing community.
MSM: Was this an assignment that the computing research group got or had it had a role in shaping the project. Making those promises in the first place.
Morgan: I do not know how,
I do not know for sure who was the driving force behind MULTICS.
I associate it with an Executive Director named Ed David. I
was not, I was not closely associated with the computing enterprise
until I became Director of Computing Science Research. I was an
applied mathematician and really had no expectation that I would
be going into a computing organization. And so I simply heard
about MULTICS. The person who did most of the talking in auditorium
and general meetings describing this system that was coming and
encouraging users to not make major commitments to the IBM700-7000
series but to wait for the GE 645 and MULTICS, the person who
did most of this was Ed David. And there was a man named Corbato
at MIT and I don't remember who the principle GE promoter of MULTICS
was, but I simply, I simply was not aware of whose brain MULTICS
sprang, sprang from. The Executive Director of the organization
that Sandy Fraser is now Executive Director of was the principle
person who promoted it and who advertised it. And, oh yes, I
should say, at that time computing service with comp centers such
as they were at the different Bell Laboratories locations were
run by technical organizations. The comp center at Murray Hill
had originally been run by the Mathematics Research Organization.
And then it was run by the Computing Science Research Organization
when math and computing science split. And the comp center at
Holmdel was run by a technical organization that kind of ran computing
service with its left hand, and at Indian Hill. Each of the comp
centers was run by a technical organization whose business was
not primarily running a comp center. They were using, the organizations
that ran comp centers had been amongst the early users of electronic
computers, and so they took on as a part of their business computing
service to the whole location at which they were. Alright. Uh.
The Computing Science Research Organization, um, which I believe
was split off from mathematics about 1965 and was under Ed David's
direction until he go promoted to Executive Director, the Computing
Science Research Organization did the software development such
as it was at Murray Hill at that time. The Computing Science
Researchers had put together various operating systems for IBM
701, 704, 709, 7090, 7094 under the names of BESYS 1, BESYS 2,
up through BESYS 7. This work, this operating system design work
was done by computing science researchers. They called it all
research and we ran with the aid of a fairly small number of associate
technical types. We ran the Murray Hill computing service. So
when MULTICS was commitment to by Bell Labs and I am sure that
Ed David was the spear head of this effort, David committed his
people in computing science research to carry the Bell Labs end
of this. I don't know Doug McIlroy was around during that time
and he will probably be able to tell you whether this MULTICS
effort was supported whole heartily by everybody or whether there
were doubters from the beginning. But in any rate, at any rate
in 1967 when I became director of Computing Science Research the
computing science research organization was committed as a major
part of its then ongoing work, to see the MULTICS project through,
and MULTICS was going slowly and was clearly. it was becoming
clear to people that MULTICS was an attempt to climb too many
trees at once. It was a mixture of research and exploratory development
and final development in so far as those terms can be applied
to software and it was simply not going to produce usable amounts
of computing service to customers. This feeling was already pretty
strong in the computing science research center when I took it
over, and users, outside users who were never very patient with
anybody, users had decided already that MULTICS was not going
to fly. Well the MULTICS effort never the less continued for
another two year approximately and was officially terminated,
Bell Labs quit working on MULTICS, as of the end of March in 1969.
MSM: You said you had,
that you came out of applied mathematics and had not been heading
to running a computer, computing environment.
Morgan: I had, I was simply
a user of computers up to that point.
MSM: Did you, when you
took over in '67 what did you see the mission of that group to
be. Did your appointment represent a shift of direction of any
sort?
Morgan: No it didn't represent
a shift in direction it represented the fact that the person who
was then director of computing science research had been laterally
transferred to one of our Whippany organizations. Was doing military
software for the Safeguard Project and I am, this was a lateral
transfer . He became a director at Whippany and I was not aware
of any particular reason why this was done I assume that he, it
was felt he was needed at Whippany.
MSM: Was this Vyssotsky?
Morgan: No this was not
Vyssotsky. Vyssotsky was a department head in computing science
research at the time the man was Tom Crowly. Crowly was shifted
to Whippany, lateraled as a director and I assume that it was
because he was urgently needed at Whippany. And there was I presume
a search for a suitable person to take over the computing science
research center and I was asked to do it but I was not told that
there shall be any change in mission. My understanding of the
mission was that it was a research group. We were-[break]- to
understand the foundations of computer science. Computer science
at that time was not nearly as large and flowering a structure
as it is now. We had work going in formal language theory. We
had work going in numerical analysis. We had work going in operating
systems. We had some work going in switching theory and we had
this exploratory development project MULTICS which had been taken
on under the aegis of Ed David and I believed at that time, and
was not given any other sailing orders, I believed that our business
was to do research to understand the power and limitations of
computers in so far as computers ought to be useful to Bell Laboratories,
and that as a transient part of this we were committed to seeing
the MULTICS operating system through. But I always believed that
this was a transient thing that we had been committed to and that
we were not in the software development business for the long
pull. We were suppose to assist software developers and so that
MULTICS would ultimately be finished up. Well it wasn't finished
up quite in the way that the originators expected it to be, but
it did come to an end.
MSM: Was that a fairly
sudden end?
Morgan: Uh, there had been
an increasing feeling both outside computing science research
and inside computing science research that the project was not
possible to finish in order, in a way that would meet the original
expectations. And our then, Vice President, Bill Baker, decided
that the work was simply going to stop. It was Baker's understanding
and mine too that we had in some sense met the commitments that
we had originally made, that we were not backing out of any contractual
arrangement without, well we were not breaking any contract and
the day on which the work stopped I believe was the last day of
March. There was an announcement, a formal announcement that
the work was stopping, that we were charging no more effort to
the MULTICS charging case. And that we were simply out of it.
I believe it had to be done that way because some folk were,
well unless there is a definite statement that you are not going
to work on a project anymore, why, some people will continue to
work on it. That's the way, that's the way research goes. Folk
from day to day do their own thing. And so there was a clear
announcement that the work was over. And this was due to Bill
Baker. I believe that Ed David actually read the formal statement
to the members of the computing science research department but
they were words that had been, this decision was made by our Vice
President. It is not unprecedented that a project shall stop,
I have known of other much larger efforts that were simply by
management edicts stopped on a certain day. And MULTICS stopped.
It was, I believe the general reaction was that it was understood
why it had to stop. It was simply, it was simply using up effort
and was not, was not advancing and showed no promise of, of turning
into a user useful thing.
MSM: You have talked to
Ken, Doug, and those that have written about it, Dennis when he
writes about it in his retrospective. We talk about the sense
of disappointed they felt. Where as on the one hand they agreed
and they agreed with the technical judgment of MULTICS's inadequacy.
That is that this was a system that was simply wasn't going to
do what it promised to do. For one thing is was not going to
deliver the cycles that sort of...
Morgan: Well that, well
that was the main, that was the main reason that MULTICS was stopped
here so far as I can see. It, it had, it was originally a promise
to users of very flexible cycles. It was the first real time
sharing system. Well there was IBM's TSS which was running at
Indian Hill. I have never been very familiar with the details
of TSS. The Indian Hill people were a group that decided quite
early on that they were not going to wait for research and for
the GE 645 and for MULTICS they had to have something sooner,
and they wanted something that was supported by a vendor which
TSS was. I don't know the TSS story really and I can't tell it
to you but MULTICS if it had met it's promised goals would have
been a much more flexible, convenient system than TSS, so I am
told and I believe. But MULTICS was not about to meet it's goals
in 1969 and I think this was generally understood. Now the sense
of disappointment that Dennis and Ken and a few other people,
I guess all the people who were working on MULTICS felt, I think
was that they had had a system that was for a few... Well in
the first place it was intellectually elegant. There was a lot
of nice conceptual stuff in MULTICS and it is always good to feel
that you are on the leading edge. But the other thing was that
it was a, it was a very pleasant and convenient system to work
with if, to use, if there were only a few people on the machine,
if you had so to speak, and I am not trying to be pejorative here,
if you had exclusive use of this large expensive toy, great!
It was fun. You could develop software, you could do all sort
of things with it. It just wasn't adaptive to supporting, it
wasn't cost effective. But heck it was something that they liked
and as you have undoubtedly been told repeatedly it was to make
well UNIX turned out to be a much simpler, more cost effective
environment which provided users with the pleasures of a, of the
same kind of sandbox. So MULTICS was fun. I, I never used MULTICS,
but I was told that it was a lot of fun for the people who were
developing it. It's just that it, it wouldn't carry the load
for a, for a big organization.
MSM: Well they've, they
said both in conversation with me, but also in writing it what
they missed was the shared environment.
Morgan: Yes.
MSM: That here was a way
in which you could, you were all working, you could share one
another's files, and delete things for one another.
Morgan: It was a very nice
environment for a small number of users, but the Labs just could
not afford and users generally could not afford the cost of the
GE 645 for the small number of users that it was supporting.
Now I am giving this to you by hearsay, but that was my understanding.
Not that UNIX was unsuccessful for a limited community of computer
science, MULTICS, not that MULTICS was unsuccessful for a limited
number of computer scientist users. But that it was not, it was
not cost effective, it would not have been cost effective to support
a large laboratory.
MSM: Now when the decision was announced that there would be no more charges made to the MULTICS case were there, were there any statements either explicit or implicit. Edicts about lines of research that would or would not be followed. I mean in a sense the MULTICS project didn't stop, stop it just got redirected.
Morgan: It got redirected.
There were no statements about what lines of research were acceptable.
We don't make statements around here about what lines of research
are acceptable in the short run. AT&T is interested in a
very wide variety of subjects. We are not interested in translating
my hieroglyphics, we are not interested in fusion research, we
are not interested in various other things but there are a lot
of things that we are interested in. Centering, mostly centering
around communication and around information handling and this
provides a great many directions that one can go. When a new
researcher is hired here, there will be a point to this in a minute
just hang on. When a new researcher is hired here anywhere in
the research area, he or she is hired because of a doctoral thesis
or some other track record that indicates proficiency, creativity
in some field that we are interested in. When the person comes
in the department head's business is not to tell the individual
what to do but to see that he or she gets introduced to folks
with common interests or perhaps common interests and that various
people come around and ask questions, try and consult with the
person, and we suggest that a new employee spend considerable
amount of time, up even to a few months, thinking about what he
or she would like to do. Very often publishing the Ph.D. thesis
is a good thing to start on. But anyway, people essentially
found their own things to do influenced by the total environment
they are in. And if a person wants to undertake a certain line
of research, a certain kind of investigation, presumably because
of the individual's technical background before coming here this
would something that's in a field that is of interest to Bell
Labs and we allow this to continue. And if the person seems to
be getting somewhere, seems to be interacting with people and
well seems to be making some progress, and particularly seems
to be communicating with his or her colleagues a considerable
length of time can go on before the department head try's to steer
the person in some other direction. The problem that sometimes
arises with people that come to with a bright new Ph.D. and seem
to be very bright is that the person will somehow have been closely
directed during his or her graduate life, worked for a professor
who held the reins very tightly and then in a kind of free environment
like this where nobody tells you what to do from day to day the
person will just not be able to find new and interesting things
to do. Now if a person does find new and interesting things to
do and if they get written up and published and if there seems
to be lot of interaction going on. And particularly, if the person
seems to have people coming in to talk to him or her from other
departments, that this person is interacting and communicating
that is fine. The person will essentially never get directed
by his or her department boss. If with the person doesn't seem
to really to, be producing anything and in particular doesn't
seem to be communicating or interacting with people the boss will
over the next several months spend more and more time encouraging
the person to communicate and to look at specific problems.
And as a year or two goes by there is always performance reviews
every, every year and a matter of communications is stressed.
The boss will become more and more explicit about, "why
don't you spend some of your time looking at this particular problem?"
But a person says, "I want to do such and so." You
are not going to be told not to. Eventually if the direction
seems to be completely at variance with others things that are
going on it may turn out that the person had better pursue this
at a university or some place like that. But in the short run
folks are not told you are not going to do this. Development
organizations. If a development project is canceled usually that
involves a team of people working on a particular direction.
If a development project is canceled, alright you are not going
to be developing XYZ widgets any more, that is clear enough.
But researchers you don't tell them what to do. What we did we
tell people that had been working on MULTICS was you cannot have
a large computer a the [inaudible] Tand was proposed as a substitute
for the GE645 during the summer of 1969. They didn't have a GE645
anymore and so there was some effort to propose the purchase of
a large computer on which to do time shared systems research and
proposals that one should by a substantial chunk of, of plant
equipment, naturally have to be approved upstairs. The MULTICS
people were told, "we are not going to provide you with a
large computer for a small number of people to do time shared
system research on." We didn't tell them what they should
do. We told them various things that we were not going to buy
equipment for. And so, that's standard operating practice for
researchers. They are suppose to be able to find things to do
and so eventually they did.
MSM: When did you become
aware of this new project, this new file system and attempts to
implement it?
Morgan: Oh, must have been late in 1969. We have a performance review every year. Starts about November and people write up a page of what they have done, it is commonly called the, I Am Great Report, although that is not an official name for it. And Ken Thompson had on his I Am Great Report in 1969 that he was working on a, a file system for a, a small computer and by 1970 why the system had been named UNIX and several people were working on UNIX. Another theme that merged with the time shared development was the text processing, text editing and formatting theme. There was, I believe at MIT or from someplace a type script formatting scheme called runoff which Doug McIlroy worked on quite a bit and turned into a thing called roff and this, I got a head of myself a little bit. Now this when we got a cheap photo typesetter, got approved to a formatting scheme called T-ROLF which is still around. People were interested in text editing and formatting and you could do this on a small computer and if you had a, if you had a operating system that involved easy handling of files and presently time sharing so that the people could access and use each other's files. The UNIX file system and the text editing and formatting work kind of came together. I understand that Ken Thompson had a space war game that he played for a while when he was looking for something to replace MULTICS. That was fun but the, the first real application of the file system that became UNIX was a text processing system. And this was kind of the merging of file system work and work that folks had been interested in for a long time, text processing.
MSM: Well this is the system that the Labs was willing to buy a machine for.
Morgan: This was a system
the Labs was willing to buy a machine for. When this, the proposal
was first made that we should buy a machine for text processing
it was presented to me because I had to sign the purchase order.
And I didn't understand at the time the innovative nature of
the UNIX file system and we had done text processing work in the
past and I didn't see that we were, that any great research advance
was being made. It sounded as if people wanted to provide a service
or something with a typing pool. So the first proposal to buy
a I guess it was a PDP11/20 I turned down, this was in 1970 and
another director Max Matthews managed to gin up a machine. He
was also interested in text processing and by the time people
had worked on the file system and on text processing for another
year by 1971 it was quite clear to everybody locally that there
was something rather special going on. And from then on we bought
a sequence of PDP-11's, and research, and by and by the patent
division was using the text processing scheme, and by and by it
was used in the word processing typing organizations, and eventually
it spread all over.
MSM: What is it that persuaded
you that there was a research component?
Morgan: Well, I do not remember at the moment any particular day on which I decided there was a research component there. I guess it was being, being shown some of the things that the text processing system could do. I guess I felt that. Let's back up a little bit. The computing science organization here like all of the research area, but I think maybe they are a little stronger than some, has always been very outspoken and very sharp tongued in promoting their viewpoints and there was a good deal of noise when MULTICS came to an end. And it seemed to me, that well I guess I had some difficulty in sorting out the signal from the noise. I was quite well aware that my bosses wouldn't approve the purchase of a really large computer to support any surreptitious continuation of the MULTICS effort. And I was, I think willing to wait for the initial shouting to die down and I figured that if there was a research component involved in the text processing work that it would appear in due course. And indeed it did. I have told this tale more than once to people who, this is were we came in, who would have liked to demonstrate that it was management perspicacity that caused UNIX to be, to be born. Some parts of management did not understand UNIX as rapidly as other parts did. I think that we did understand management principles. The management principles here are that you hire bright people and you introduce them to the environment, and you give them general directions as to what sort of thing is wanted, and you give them lots of freedom. Doesn't mean you always necessarily give them all the money that they want. And then you exercise selective enthusiasm that is one of Bill Baker's favorite phrases. You exercise selective enthusiasm over what they do. And if you mistakenly discourage or fail to respond to something that later on turns out to be good. If it is really a strong idea it will come back. So anyhow that is...
MSM: We had a dean at Princeton
who was famous. No matter what the proposal the first answer
was no. Then you came back.
Morgan: Alright, well it
is a little bit like patent examiners. I was told in my early
career when I got a few patents that there is always an interplay
between the local patent attorney and the patent examiner at the
bureau of patents in Washington. And you had chit chat for a
while and then you get from the patent examiner a final rejection.
This patent is now finally rejected. All claims are rejected
and that is an invitation to put forth your strongest argument.
You now give me your best argument and I will listen to it.
But I was told that the final rejection is the last, last stage
before you finally get some things admitted. Well, we didn't
quite play it like that around here. But that. I have never
been in an organization that had enough money, or enough hiring
slots, or enough office or land space to do everything that we
would like to do. So one provides some back pressure. And in
the case of the transition from MULTICS to UNIX the MULTICS faucet
had to be turn off reasonably hard. It was a part of turning
off, I mean it was a management decision that this was going to
be turned off. And part of turning it off was not immediately
buying hardware on which MULTICS could be continued. In retrospect,
Thompson and Ritchie and other people did find partly through
their own efforts and partly by looking for a director that was
willing to buy a small amount of hardware, did find machines on
which they could work. And in due course when it was clear to
everybody around the research area that UNIX was going to go somewhere
and needed to be supported they have had the machines that they
wanted.
MSM: One of the things
that impressed me, that has impressed me about UNIX is the way
in which it's a system with a lot of really effective coding,
programming. On the other hand it is a system that if it's not
theory driven at least has roots in theory. There is a, if one
works ones way through the various books by Aho, Hopcroft, Ullman
or the various combinations of three things taking two at a time,
that series represents. And one looks at UNIX, one sees their
other piece, and that is the product by in large of the computing
research group. How did that unity come about. Was that something
that you were looking to create or is it pure coincidence, or
it is...
Morgan: I...
MSM: It is a view of computing
really as an enterprise...
Morgan: I, I don't, there
was a conscious decision to do this unified sort of way. As I,
yes I was saying earlier we have had at Bell Labs for a long time,
certainly since before I came and I have been here forty two years,
we have the philosophy that you hire bright people, you expose
them to interesting problem areas, and you keep an eye on what
they are doing and in particular on their interactions with other
people. You attempt to give them guidance only in a very general
sort of way. Often times this guidance is simply a lot of enthusiasm
for something that they are working on. An environment like that
is self perpetuating, so long as you keep your hiring standards
up and your management keeps its eyes open. Remember that the
management around, around Bell Labs all are, at least around the
research area, all came up through the technical route. And people
don't get promoted to management in the research area here unless
they have a good track record. You may find a few folk who disagree
with that. But my view is that our department heads and directors
and executive directors were once technical hot shots. Sometimes
that back fires, because you get a technical hot shot who has
no people skills. But that is a different story. Your, uh,
your technical people, your managers were once technical hot shots
and they were imbued with this general philosophy of how you conduct,
how you manage research at this kind of place. And good things
come out of this. Well, good things come out of this philosophy
of research management. And as I said it is self perpetuating
in that that's the way that I grew up and that is the way Al Aho
grew up and it is the way Sandy Fraser grew up and it is the way
our local department, department heads have seen it work for a
long while. What we did in the computing science research organization.
I didn't create this, to some extent I keep it going. But was
to have bright, interactive people who were interested in theoretical
computer science, that is Al Aho and now Ravi Sethi and lots of
other people, but Aho was one of the early ones. And folk who
were interested in programming techniques. People like Doug McIlroy
who was a mathematician to start with, but was in on computer
from very early on. And people like Peter Weinberger, who I
guess was an algebraist to begin with but he is interested in
practically everything. And people like Brian Kernighan who are,
is a superb expositor. Kernighan writes beautifully, which a
great many computer scientists don't and is very creative in seeing
something that needs to be, needs to be done. And people like
Mike Lesk perhaps not the world's greatest programmer but Mike
Lesk who is now at Bell Core did the EQN patient setting system.
No I guess that was Brian, that was Brian. Alright Mike Lesk
did a lot of the page layout stuff that got put on top of, of
T-ROLF. Brian, Mike, Lorinda Cherry, could see things that, by
golly, I know how to do this and I bet somebody else would be
interested in it too. The EQN equation layout package for instance
after it has been done. Sure other people do that sort of thing,
there's --
[end of side one]
-- had the inventiveness
to see (A) that I can do it and (B) that it would be nice if somebody
did it. Now maybe it ought to come the other way around, perceiving
the, perceiving the niche and observing the technique, that the
technology exists for doing it. These are two parts of the same
thing. Alright we had people like Brian, we had people like Mike
Lesk, we had people like Doug, we had people like Ken and Dennis.
We had people like Alfie and if I, well it was a small group,
it was about two dozen people for a number of years after, after
1969 when we quit running a comp center service. I will come
back to that in a moment. And when the MULTICS effort and some
people who were associated with MULTICS in a more supportive sort
of way went elsewhere. From about 1970 to about 1976 or 1977
we had about two dozen people, twenty four people in the whole
group. We were essentially not hiring anybody, we were in one
of our chronic hiring freezes and we just had a small group of
good people who generally ate lunch together, and who were quite
willing to argue with each other and to discuss and to use the
techniques they knew about to put together things that they thought
were interesting. It was a lot of work in text processing at
that time. There was a lot of work in practical operating system
development. There was a lot of work in theoretical computer
science, compiler theory and algorithms. It was done essentially
by a hand full of people. But they were people who did a lot
of talking to each other and a lot of shouting and who essentially
collaborated. And that is the way that research is suppose to
be done.
MSM: What I am getting
here is a picture of a group of people hired within a fairly short
period of time, several years. At a time when things are the
field was still in flux.
Morgan: It was still in
flux.
MSM: What computer science
was suppose to be in the 60's was an open question.
Morgan: That's, that's right.
MSM: Brought together they
tend to form a stable community and in essence what I see is a
usual match between practical applications and theoretical foundations.
Certainly the out, the product of a group that had come together
for different reasons but then had found a common, they had established
a community.
Morgan: They established
a community. They were in touch, of course, with the university
community with computer science being established, being defined
as a field in the 1960's. Some of the old timers, myself and
Doug McIlroy, and a few others came from the math research center.
We had a numerical analyst named Dick Hamming, early on who was
kind of the father of service computing at Bell Labs.
MSM: What had been your
background?
Morgan: I was mathematical
physics. And as I said I was an applied mathematician. I did
boundary value problems, radar antennas, microwave, wave guides,
up until I was put in charge of computing science research. And
I suspect that I was put in charge of computing science research
because just at the time that the previous director was transferred
to Whippany there wasn't a really good professional computer
scientist who was ready for the job. Sometimes you know there
is, you have to fill a slot, and there isn't a perfectly qualified
person of the right amount of seniority. So I didn't look for
the job in computer science. It wasn't compatible with my background
really, but I had been steeped in the general way in which research
is managed here. You get good people, you make sure they are
aware of a wide class of problems and you let them do their own
thing. But you give them feedback.
MSM: Did you have a sense
yourself of what computer science would look like at the time?
Morgan: Not really. I was
more interested in the mathematical parts of it. Of course I
was interested in seeing that people got, so long as we ran a
service organization I was interested in seeing that service was
supplied. In the middle of 1969, this is a new topic but it
is relevant I think...
MSM: Right. It was about
to ask you what happened when you ceased to be a service organization.
Morgan: In the middle of
1969 service computing at all Bell Labs locations was taken away
from technical organizations, that is organization whose primary
business was something else. And merged into a single division
under a man named Phil Thayer, who is now retired. All the comp
centers were put under unified management and for a year and a
half I was both director of service computing at Murray Hill and
Whippany, and director of computing science research. It was
expected that this would be a temporary arrangement and so at
the end of 1970 I chose to give up the service computing aspect
and from then on it was under its own separate management. Well
anyhow. The move to a separate computing service organization
I think was long over due. It just wasn't reasonable for a single,
for a research organization also to try to manage a stable computing
center. Researchers you know always like to fiddle and, "we
are going improve it!" And users, "my program that
ran yesterday won't run today what the hell have you done with
the operating system now? And besides I want this thing fixed!"
and the researcher may feel, "I am not really interested
in fixing that. That is mundane..." So anyhow, having a
separate computing service organization was long over due. It
was started in the middle of 1969. And this was another impetus
I think toward the development of an operating system for small
machines namely UNIX, that went on in computing science research.
Because once the comp center machines were moved out from under
the research aegis, we had in research had our own machines and
management would not buy big DEC PDP 10's and so we had to do
something on minicomputers. And that was another impetus toward
Thompson and Ritchie in the UNIX direction. Anyway, that's kind
of the history of things. I was interested in both in seeing
that users got good service and got new tools as rapidly as they
could be delivered. And also in seeing that the theoretical understanding
of computer science got advanced. And if I had any personal interests
here I found the text processing work particularly entertaining.
But I did not do any of the, I did not do any of the research
myself. I was, I was the research manager.
MSM: Was there any point
at which you were running interference for these people?
Morgan: Well you always
run interference in that part of the business of management is
to explain to higher management and to other parts of the laboratories
what is going on. So that was just part of the job. I never
felt that, I never felt that I had to go out and fight for them
in the sense that I was contending with anybody. I was, I was
if anything a publicist and explainer rather than a person who
was going out and contending with other people for money. Our
research area here has always been I think quite well supported.
It doesn't mean that we have everything we wanted. But if you
can make a good case for something, and it is the director's business
to make a good case, if you could make a good case for something
well you could get it.
MSM: At a certain point
I gathered between 73-74 I am not sure what the date is. UNIX
in a sense out grew the research group or...
Morgan: Well basically it
out grew
MSM: Did it outgrow it
and therefore did you feel that it had to someone else had to
take over...
Morgan: Well
MSM: Take over or...?
Morgan: Remember in 73 and
74 we had a couple of dozen people who included some numerical
analysts, and some people like Al Aho, Jeff Ullman was around
here for a while. He is now in Stanford he has become a big man
in academic computer science. We had a small number of people
and we had essentially no support staff and there were things
that folks wanted to do with UNIX that our people were not terribly
interested in putting the time on. My perception of the way things
grew is roughly as follows and you will want to talk to Berkeley
Tague on this because he was heavily involved in this next stage.
My perception is that various software development organizations
around Bell Labs picked up UNIX and its researchy state and began
adding things to it, to meet their needs. The first application
of UNIX outside research where it was a test bed or an environment
for software researchers to use, the first thing that was done
outside, was early 1971 where the patent organization took UNIX
on as a word processing system. And there was another system
whose name I forget now I could look it up, it was a commercial
system that was a competitor. Well, the patent organization was
persuaded by some of our folks including particular Joe Osanna
who was the father of troff. Osanna died I think in 1977. But
anyway Osanna persuaded the patent people that they ought to use
UNIX as a text processing system. And so computing science research
organization provided what extra bells and whistles were needed
and hand holding for the patent typists and so forth. But by
1972 there were a number of outside organizations, outside research
at Bell Labs that were beginning to use UNIX for software development.
And each one of them had special needs and so they came and they
talked to our people and then they added their own bells and whistles.
It was about this point that Berkeley Tague who'd been in and
out of research and was at the time associated with the comp centers,
decided that, that was a game that he wanted to play. His story
now is that he was tired of the work that he was doing at the
comp centers. He wanted to get into the UNIX action and this
was a very interesting area and it was clear to Tague, who has
always had his head screwed on right, it was clear to Tague that
somebody was going to need to support UNIX. And so he persuaded
his management as part of the comp center that there needs to
be a UNIX support group. His version of this is that he got together
a few people and the way that he became a UNIX support group was
that when anybody would ask a question, "We had this problem
with UNIX," Tague would say, "well I will take it under
consideration." Then he would get on the phone and fly around
like mad to every UNIX user he could think of and say, "have
you every encountered this problem? What did you do about it?"
When he found out he would trot this back, and "you know
this is the way that you can fix it!" And after six months
he had some local expertise. Well you should hear Tague tell
this story. Anyway. Starting in 1972 UNIX began to be adopted
here and there and then it was a kind of an avalanche of organizations
that needed it for software development purposes or for running
their own particular version of the operating system on. And
they got to the UNIX organization supported out of the comp centers
and Thompson, Ritchie and Kernighan and Mike Lesk and Doug McIlroy
and the folk who had contributed to the early stages of UNIX went
on adding things as they wanted to, and answering calls from Tague
occasionally, and generally consulting, but they were not in the
support business any longer. They hadn't been in the support
business very much although at the beginning, whoever invents
something has got to give it some support. But a UNIX support
organization was formed I think in 1973 under Tague, and from
there on it...
MSM: Well the system was
first announced if I am not mistaken in 1974 at the IBM conference.
Morgan: I believe that..
MSM: Was there a definite
decision about when to go public with this and... Cause there's
a, one sees a flurry of papers around the mid '70s, is that a
sense of back-up?
Morgan: Well we. Computer
scientists most of them we don't like to write papers. And so
it was a while after the system existed before it was, before
it was written up. I forget, I could look up when we first began
licensing UNIX to universities. There is another interesting
aspect of the UNIX story which might not happen the same way again.
AT&T management could not understand the importance of what
it had in UNIX. And so we began licensing UNIX to universities
for a very small amount of money. At most a few hundred dollars.
And UNIX was licensed all over, it spread over the academic world,
Berkeley but many other places, it just went all over the academic
world in the early 1970's. Doug McIlroy can tell you when or
I could dig up out of my records the first dozen or twenty or
fifty universities that we licensed. AT&T had simply let
universities have UNIX for a song on the basis that they would
not do certain things with it that universities didn't very often
do anyway. And as a result of that graduate students got on UNIX
and decided that, "I love this, where can I get more of it?"
And this piece of serendipity was I think very important in spreading
UNIX over the world. The fact that it was essentially given free
of charge to universities. And I suspect that we would not do
that again. But it was a case of AT&T's top management not
realizing that software could be financially, extremely valuable.
MSM: You also weren't suppose
to be in the computing...
Morgan: We were not suppose
to be in the computing business. So we were not making computers
and people did not realize that the computer is more than just
hardware. It's also software. So giving UNIX away essentially
free, talking about it freely and having open lines of communication
with our university colleagues especially those at Berkeley, but
also elsewhere, this was a major thing in the spread of UNIX.
And to some extent it was an accident.
MSM: So they, that set
of papers that appear in the mid '70s is that just a..
Morgan: Well now.
MSM: The timing there accidental?
Morgan: Well, the first
big set of papers was the 1978, BSTJ. I have got that around
someplace.
MSM: Yeah, I have got the
reprint of them.
Morgan: Alright. It was,
I was involved in that rather heavily and in putting it together.
The BSTJ decided that computer software was a good thing and
that we ought to have, we ought to have an issue on it. And there
was a general consensus that the UNIX operating system by 1978
needed a special issue. And so we kind of made that a, well we
made that a project. I was the, yes guest editor who put the
issue together and we sold the issue at two dollars and a half
or three dollars and a half a copy, and it has out sold every
other issue of the BSTJ that we ever published by a factor of
many. It over took the transistor issue almost at once, and it
was a tremendously heavy seller, and they sold it for the standard
rate on single issues of BSTJ which I forget how much it was then,
I believe it was two dollars and a half. And that was another
thing that AT&T management didn't recognize that they could
have done much better by charging five dollars an issue. The
big spate the first big spate of papers on UNIX was the 1978 BSTJ
issue and I don't believe except for some conference papers and
the original Ritchie-Thompson paper, I don't believe there had
been much before then. I know I beat quite hard on folk to get
papers written for the 1978 BSTJ issue. And we kind of regarded
that as the journal unveiling of the subject. There should have
been more publications before then but it was a case of lots of
word of mouth and lots of electronic mail. And the fact that
most computer scientists with the except of Kernighan and Aho
don't much like to write. There was no conscious deci..., before
the BSTJ special issue, there was no conscious decision, "we
will now publish on this." Our folk were encouraged to publish
anything that is worth publishing. But some people you have to
push pretty hard to get them to publish and we don't push perhaps
as much as we should.
MSM: So there was no sense
of sort of a formal release to the world?
Morgan: UNIX didn't get
any formal treatment in it's early stages it was a thing that
Ken and Dennis made and other people as they heard about it by
word of mouth, got on to it, "oh this is great!" And
university people heard about it and, "Oh yeah you can have
it. We will set up a mechanism, you pay us three hundred dollars
or whatever and sign this agreement that you won't set up your
own company to market this stuff." And it was all a very
small scale and low key in the beginning.
MSM: If you look back on
it from a point of view of manager, is UNIX the sort of thing
that management can make happen?
Morgan: No. You can't make
it happen. You hire bright people, provide them with a stimulating
environment. And you do selective pruning and encouragement of
what they undertake. But you don't do this with too short a time
constant. As I said the first time that Thompson and company
asked for a computer they asked for a DEC PDP10 and they were
told no on that. It was simply to big and you are not going to
do operating system research for a big computer just after MULTICS
has been turned off. The second time they asked, they wanted
a PDP11/20 and I said, "I am not convinced yet, I want to
see more of what you say you are going to do with your text processing
system." So I didn't stomp on them, but neither did I sign
their order and they found somebody else, another director to
sign the order, and the third time they came around they wanted
an 1145 and by that time they had a perfectly plausible and defensible
story. So they got selective enthusiasm used on them but not
too violently or with too short of a time scale. If it is going
to be good it will prove itself.
MSM: Well, you know, you.
The selective enthusiasm with a counterpart. You were displaying
selective enthusiasm. They were displaying entrepreneurship.
You said no, so they went and found someone that said yes.
Morgan: That is right.
MSM: Is that also, is that
conscious management? That is, encouraging people to do sort
of in-house entrepreneurship.
Morgan: Not in the sense
that people are told that if you want this funding you have got
to find somebody to fund it. We don't ride our folk with a short
rein. And we don't tell them who they can talk to or who they
can't talk to. And I didn't tell the guys when they came around
and asked for the 11/20 and I said, "Look I am not convinced
by this story." I didn't say, "If you can find somebody
else to do it, feel free." But I didn't tell them "Look
you must not talk to anybody else," and when I discovered
there were no secrets about this I discovered, I was told "Look,
Max Matthews can support this." He was the other director.
And why could he support it? He could support it because he
was interested in text processing, he was doing, he was in behavioral
sciences and psychology and he had people who were working on
text processing, and in fact one of the folk who was, came in
with the proposition for the 11/20 that I turned down. One of
the folk was Lee McMahon who was one of Max's department heads.
Lee has recently and very sadly died. But anyway, Lee McMahon
was the department head of Max Matthew's so the pitch was not
really that two of Morgan's MTS, having been thrown out of Morgan's
office, I exaggerate here, but having been thrown out of Morgan's
office, trot down the hall to Matthew's and, "Morgan doesn't
love us what can you do for us?" They came in through McMahon's
director and in fact I am, I hadn't really reviewed this but as
I think about it comes back to mind. The people who originally
made the proposition to me were Osanna, Thompson and McMahon who
were working together. They were in different organizations,
they were working together. McMahon was particularly interested
in English grammar and in text analysis at the time. Osanna was
interested in editing and formatting systems, things like the
runoff system. Thompson was interested in file systems and ways
of handling information conveniently in a computer. So these
birds were talking to each other and they had discovered they
were all interested in small computers. They had discovered the
PDP-11/20, and they felt that they could make a good text processing
system out of an 11/20 which would exercise Thompson's file system.
I don't know whether it was called UNIX then or not. It would
exercise Thompson's file system, it would give Osanna a chance
to write text processing programs. It would give Lee McMahon
a chance to process text in his analysis of grammatical structure.
Which is what he was doing as a psychologist in Max Matthew's
area. So they came to see me. And I obviously didn't fully understand
where their proposition was going and so I said I couldn't do
it. Not then. So they go around and they give this pitch to
Matthews who is Lee McMahon's director. And so it wasn't a case
of guys from my organization going outside my organization to
find somebody. They simply went to the director of the other
person in the trio. And I didn't think there was anything wrong
with this or unreasonable about it. Max Matthew's was a person
who collected little computers anyway and it may be that he had
more money at the time than I did. We have certain plant budgets.
Anyway, I didn't think that there was anything out of the way
about this and I didn't feel that somebody was criticizing my
judgment. Max was perhaps in a better position to support the
machine or I don't know what. But anyway it didn't seem unusual
to me and now that I recall McMahon reported to me for a long
time. He transferred from Max's organization into mine not too
long after this particular incident. But at the time he was a
department head for Max Matthew's. So he added his voice to the
desires that the other fellows had and so Max bought the first
computer. But there was nothing, there was nothing particularly
unusual about this. And since it would have been a little more
unusual if people entirely from one center had gone to a different
director and said can you support this. In which case, I am sure
that Max would have come to me and said, "Look, two of your
guys have come to me and have said will I buy them a computer.
If a computer is going be bought you ought to buy it. Let's
discuss whether it should be bought or not." The reason
it didn't happen that way was that one of Max's department heads
was in this story.
MSM: It sounds less entrepreneur...
Morgan: Well, now I should
have told you, told that to you that way in the beginning. I
remember now who the people were. It was Thompson, Osanna and
McMahon, and at that time McMahon was a department head for Max
Matthew's, and Thompson and Osanna were MTS in my shop, and each
one of them had a separate interest in the particular project,
and Max was willing to buy it. If he hadn't had one of his department
heads pitching it to him he wouldn't have bought it. I would
have done it or nobody. And if I hadn't, if Max had come to me
and said, "Look, I have listened to these guys and I think
they have got a good story, and why don't you listen to them again,"
I probably would have. Anyway there was nothing unusual about
this and it wasn't a case of people either being encouraged or
choosing to find a back door. We do encourage people to be enterprising,
that if they want something done, or if they want somebody to
cooperate with them. There is another kind of question that comes
up occasionally with a new MTS. You will occasionally get someone
with feels that he would like to have somebody working with him
and this won't happen until somebody's boss says, "You two
guys collaborate." One does not tell researchers to collaborate
with each other. You find, you find common interests in somebody
and then the collaboration occurs. So if somebody came to me
and said, "You tell somebody to, you tell such and so to
work with me." You make your own contacts. So folks, folk
are encouraged to be entrepreneurial in the sense that they make
contacts and they get collaborations going. They are not encouraged
to go and ask some other member of management for money that they
can't get from their own management. But in the case of the Thompson,
Osanna, McMahon thing the pitch to Matthews came from one of his
own department heads.
MSM: Well, we have gone
on for a while and I think that I have run out of questions.
Morgan: Alright.
MSM: But I also probably..
END OF TAPE