From: rh...@ciao.cc.columbia.edu (Ronda Hauben)
Subject: John Lions: Interview from 1994
Date: 1998/12/09
Message-ID: <74kp23$2fq$4@apakabar.cc.columbia.edu>
X-Deja-AN: 420285061
Organization: Columbia University
Reply-To: rh...@columbia.edu
Newsgroups: comp.unix.misc,comp.unix.admin,comp.org.ieee,alt.culture.internet,alt.amateur-comp,alt.folklore.computers

In comp.org.usenix, Peter Salus posted that John Lions died on
December 5, 1998. That is a great loss to the computer science
and UNIX community. I am posting this Interview that I did with
John in 1994 as a tribute to his important contribution to 
computer science.

This interview that appeared in the special issue
of The Amateur Computerist in celebration of the 25th anniversay
of UNIX. It appeared in the summer 1994 issue.

-----------------------------------------------------------------

                Spreading UNIX Around the World:
                  An Interview with John Lions

   [Editor's Note: Looking through some magazines in a local
university library, I came upon back issues of UNIX Review from
the mid 1980's. In these issues were articles by or interviews
with several of the pioneers who developed UNIX. As part of my
research for a paper about the history and development of the
early days of UNIX, I felt it would be helpful to be able to ask
some of these pioneers additional questions based on the events
and developments described in the UNIX Review Interviews.
   Following is an interview conducted via E-mail with John
Lions, who wrote A Commentary on the UNIX Operating System
describing Version 6 UNIX to accompany the "UNIX Operating System
Source Code Level 6" for the students in his operating systems
class at the University of New South Wales in Australia. Lions'
important book provided some of the earliest printed commentary
and documentation of the UNIX kernel. John Lions is a Professor
of Computer Science in the School of Computer Science and
Engineering, at the University of New South Wales.]

Q: John, I have been reading with joy the interview with you that
was published in UNIX Review in October, 1985. I found it
inspiring because it showed the hard fight you and your
colleagues and students took up to be able to adopt UNIX at your
University and to help to spread it in Australia and around the
world. In the UNIX Review article you describe the arrival of
UNIX saying "UNIX was a revolutionary force on our campus." You
tell how the University of New South Wales decided to purchase a
Cyber 72 computer in 1974. But, since the Cyber could only
recognize User200 terminals which were by that time obsolete, the
University bought some PDP-11/40's to emulate User200s. You
describe how you wrote for information about UNIX after reading
an article by Ritchie and Thompson published in the "Communica-
tions of the ACM," and explained how a copy of Edition 5 tape and
manuals arrived in late December, 1974. A little later in the
Interview you relate how Ian Johnstone with assistance from
others wrote a new User200 emulator "that ran under UNIX. That,"
you point out, "became the first application of UNIX to be
written in Australia .... This exercise proved to be extremely
important. With a PDP-11," you explain, " completely to our-
selves, we most likely would have run vanilla UNIX on it and been
happy. But because we had to provide the User200 emulator, we had
to learn a lot about the system and pay a lot of attention to
performance issues. We needed help, but we couldn't get any from
outside sources. So we ended up generating our own expertise."

Lions: Undoubtedly true ....

Q: What was it about UNIX that led you to do the hard work that
you did? Were you aware of the power that it promised? Was that
some of the consideration or was it more practical -- that you had
certain things you wanted to be able to do and could hack to get
the UNIX system to do it?

Lions: UNIX was wonderfully plastic. We changed things to adapt
them to our situation ... because it was a challenge, and we were
having fun!

Q: You then say that through your work on UNIX you started to
make a few friends elsewhere on campus. Were they from any other
particular department? How did you begin to build a user group?
Did you start having formal meetings?

Lions: Other people at the other batch stations were interested
in solving the same problems as we were, so we found a common
cause. This included the Library which in those days had
passwords for the ordinary user accounts, but not for the super-
user ... very convenient!

Q: Can you say what kinds of similar problems people in other
universities were encountering at the time that led you to be
able to work together?

Lions: In a word ... isolation.

Q: Do you have any idea why UNIX was so widely adopted at other
Australian Universities?

Lions: We spread the news evangelically ... We were very anxious
to share our accumulated knowledge and to experiment ... and we
wanted to share it with others. We were having fun!

Q: You say that UNIX has possibly made a deeper penetration in
Australia than in any other country.

Lions: That comment has to be understood in its proper context. I
would not make it today. UNIX penetration is now 100% by
university, though not by department within universities. The
Internet is heavily UNIX-dependent, so I believe.

Q: In the UNIX Review interview, you describe how in 1975, Ian
Johnstone who was acting as a tutor for the operating systems
course you taught, asked, "Why don't we run off a few of the
source code files for the kernel and ask the students to take a
look at them? Then we can ask them some questions; maybe it will
be interesting." What kind of questions did you folks intend to
ask?

Lions: The same kind that the Commentary answers ....

Q: After you took his suggestion and you both selected what
seemed like a reasonable subset of the kernel and handed it out
to students, you report that you asked them questions, but that
they didn't have enough information to answer them so "they came
back to us with questions of their own many of which we couldn't
answer." Can you say any more about how the students suggested
that you offer the complete kernel for study?

Lions: They suggested that it should be all or nothing. The
selection of code I finally printed (on a DECwriter) is only
complete in a limited sense. Section Five that deals with device
drivers could have been much longer.

Q: Was there any special reason that you took their suggestions?
What led to the preparation in 1976 of the booklet containing the
source files for a version of Edition 6 UNIX that could run on a
PDP-11/40 system?

Lions: Seemed reasonable at the time ... what other options
reasonably existed?

Q: You say "Writing these was a real learning exercise for me. By
slowly and methodically surveying the whole kernel, I came to
understand things that others had overlooked." Can you give any
examples of what you came to understand that others had over-
looked?

Lions: No. I guess what I meant to say was that I obtained an
integrated view that allowed me to see more connections in the
code than others did. I used to test the students' knowledge and
understanding by weekly tests. Most years there would be two
tests on each of the first four sections of the code.
   Students could sit for both tests in each section, but they
were discouraged from submitting more than one answer for
marking. If they chose to submit two answers, their mark was the
better of the two, less 10%. This allowed students to recover
from a bad first result, while discouraging them from trying
again if their first attempt was reasonable. Marking was always a
problem as overnight turnaround was needed.
   The sophistication of the questions increased over the years
and towards the end, some new questions were quite devious.
(Don't ask me for examples!)

Q: In the Commentary you say: "A decision had to be made quite
early regarding the order of presentation of the source code. The
intention was to provide a reasonably logical sequence for the
student who wanted to learn the whole system. With the benefit of
hindsight, a great many improvements in detail are still
possible, and it is intended that these changes will be made in
some future edition." Did you ever write the future edition
making the changes?

Lions: No. There had been a three year gap between [UNIX -ed]
Editions Six and Seven. This created a window of opportunity for
us that never really occurred again.

Q: In your Commentary you say "You will find that most of the
code in UNIX is of a very high standard. Many sections which
initially seem complex and obscure, appear in the light of
further investigation and reflection, to be perfectly obvious and
'the only way to fly.' For this reason, the occasional comments
in the notes on programming style, almost invariably refer to
apparent lapses from the usual standard of near perfection ....
But on the whole you will find that the authors of UNIX, Ken
Thompson and Dennis Ritchie, have created a program of great
strength, integrity and effectiveness, which you should admire
and seek to emulate."

Lions: That is what I believed then ... and still do.

Q: Can you say any more about the conclusion you drew of the high
standard of code in the UNIX kernel? Do you feel that students
and others who studied your book and the code did emulate it? Did
that help improve the level of code of those who had access to
your book and the source code?

Lions: In a general sense, I believe the answer is 'yes':
students did learn better coding practices.

Q: In the UNIX Review article, you relate that in 1977 at the
University of New South Wales you were developing your own PDP
version of UNIX to handle heavy student loads and that Ian
Johnstone, Peter Ivanov and Greg Rose developed a "sanitized
extended version of UNIX." And you made some changes to the
kernel. Can you say what the most important ones were?

Lions: We fixed bugs that we found ... or had introduced
ourselves. I cannot recall what they were ... and of course the
Seventh Edition changed everything anyway. We only did it once:
that was enough.

Q: Was your examination of the kernel for the Commentary helpful
in determining what changes to the kernel were needed. For
example, in the Commentary on pg. 82 under "Some Comments" you
say " 'namei' is a key procedure which would seem to have been
written very early, to have been thoroughly debugged and then to
have been left essentially unchanged. The interface between
'namei' and the rest of the system is rather complex, and for
that reason alone, it would not win the prize for 'Procedure of
the Year.' Earlier in the Commentary in chapter 19 (pg. 82) you
say "Copy the eight words of the directory entry into the array
'u.u_dent'." Then you comment, "The reason for copying before
comparing is obscure! Can this actually be more efficient? (The
reason for copying the whole directory at all is rather
perplexing to the author of these notes.);" Were these problems
clarified upon further examination or if not, did you make any
effort to solve them when you folks made changes to the kernel?

Lions: No comment now. My understanding changed over the years,
and some questions that may been obscure once were no longer so.

Q: In the UNIX Review Interview you explain that you were the
first person from the UNIX community in Australia to spend a
sabbatical at Bell Labs. Who invited you to the Labs? When? Why?
What did you do once there?

Lions: After I started distributing copies of my notes on UNIX
(Source Code and Commentary), I sent more than two hundred copies
to BTL [Bell Telephone Laboratories -ed]. One night (sometime in
1978?), I had a phone call from Doug McIlroy saying BTL would
like to assume responsibility for distributing those documents,
and would I agree? I did. It saved me much work.
   At the beginning of 1978, when I was starting to wonder what
to do for my first sabbatical leave, I had another late night
call, this time from Berkley Tague enquiring whether I might be
willing to visit BTL, another easy decision.
   In the middle of 1978, my family (us and two daughters) set
off for the USA, Madison, NJ in particular, where Berkley had
arranged for us to rent the house of an academic from Drew
University. (They were going to the south of France for his
sabbatical!)
   I can still remember arriving at 26 Morris Place, tired but
pleased to be there (I think we must have rented a car from
Newark airport). Shortly afterwards Berk arrived and introduced
himself. We have been firm friends since then, with both him and
his wife, Anne-Marie. He is undoubtedly one of nature's
gentlemen.
   Madison, N.J. is only a few miles (less than 10 -- I forget!)
from BTL. Incidentally, Berkley used to collect me each morning
and drive me from Madison to the Labs, so my wife could have our
car!

Q: You say in the UNIX Review interview that you worked in the
UNIX Support Group while at Bell Labs during that first
sabbatical and were able to introduce a number of utilities,
including pack, etc. Can you say more about what your work was
during that first sabbatical?

Lions: There were no expectations and I was given a free hand to
follow my own interests. Fortunately for BTL I had lots of ideas,
so there was never a problem.

Q: Do you know how your book was used as part of the work that
USG [UNIX Support Group -ed] was doing? Do you know how it was
used both elsewhere in Bell Labs and outside?

Lions: I had sent them the original copies of my notes. They
reproduced them and provided one copy to each new licensee (so I
believe). Each new licensee was allowed to make additional copies
under specified conditions.

Q: At the end of the UNIX Review Interview you say that it was
not so much that the UNIX system is friendly but that the people
who use it are. Do you have any sense of what about UNIX makes
this true?

Lions: Not really. That was just my experience. I think you ought
to remember that BTL is a very special place, and its Research
Department is also very special.

Q: What would you see as an appropriate way to commemorate the
25th anniversary of the creation of UNIX in 1994?

Lions: I gather Usenix is attempting to organize a meeting.

-----------------------------------------------------------------
Reprinted from the Amateur Computerist vol 6 no. 1
The issue is online at http://www.ais.org/~jrh/acn

Ronda
ro...@panix.com