Edsger Wybe Dijkstra 1930-2002
Edsger Wybe Dijkstra is one of the computer pioneers and has
developed the framework for proper programming.
His early years
Edsger Wybe Dijkstra was born on May, 11, 1930 at Rotterdam. His
father, Douwe Wybe Dijkstra was a chemist and his mother, Brechtje Cornelia
Kruyper, was a mathematician. This scientific background had
a strong influence on his career, and on his life. He attended the
Gymnasium Erasmianum (High School) in Rotterdam. In 1948, he entered Leyden
University. Instead of the law career at the United Nations he had fancied in his
Erasmianum period, he now chose mathematics and physics.
First programming experiences
Within three years he got his bachelors degree and his father
was so pleased that in September 1951, he gave him permission to go to England for a summer
course. It was a Cambridge University course in Programming
for electronic computing devices, given by the famous M. V. Wilkes. His
professor made him write a letter to the head of the Computing Department of the
Amsterdam Mathematical Center, Aad van Wijngaarden, to check if his basic
subject knowledge was sufficient.
The previous year, Van Wijngaarden had attended the first
Cambridge course. Van Wijngaarden immediately wrote back two lines: first to confirm that
Dijkstra's knowledge was sufficient, and second to ask to come to Amsterdam and work for him
as a programmer at the MC.
Which Dijkstra did.
For Dijkstra, still a student at the time, that 1951 summer
course, given by Wilkes, became the basis for his future career.
Freedom at the Mathematisch Centrum
1953, MC Amsterdam, Edsger Dijkstra, B.J. Loopstra and Dijkstra's future wife Ria Debets (Picture taken by Gerrit
Blaauw).
Without any knowledge of Zuse?s Plannkalk?ll (1946) or the
work of Hopper, Dijkstra started his programming career by rewriting the
programming mutations and input programs Van Wijngaarden had written. Van
Wijngaarden allowed him to do it. These programs had been developed for MC?s
first computer, the
ARRA I
, an in house development project by
C.S.
Scholten
and J.
Loopstra.
Zeeland Delta Project
The Computing Department of the MC worked day and night to solve
the numerous problems connected with the large scale projects that were
developed in The Netherlands, like the
DELTA project
for the safety of the
Zeeland province, containing the truly gigantic Haringvlietsluizen-complex.
Delta Project, Haringvliet sluices
Another major project was the development of the Fokker
Friendship aircraft , with its
wing flutter calculations, that demanded the maximum of MC 's capacity.
In
1953,
Gerrit Blaauw joined the MC team. First the
ARRA II
was constructed and because of the reliability of this machine Fokker Aircraft
Corporation ordered a similar computer, called the
FERTA
. This
FERTA
was twice as fast as the ARRA II and had a different type of coding. For both these
machines Dijkstra developed the software, as well as for its 1956 successor: the
ARMAC,
the last computer developed for the MC.
ARMAC,
1956
After the completion of the
FERTA
Gerrit Blaauw went to work for IBM, Poughkeepsie, USA, where he would
work on the development of the IBM 7030, the "
Stretch
" and eventually
design and build the IBM System 360.
A new challenge: Electrologica
Because developments of large scale calculating machines
was an ongoing process the preparations for a further computer had already been
finished by Dijkstra, Scholten and Loopstra when the
MC management
and
Life
Insurance Company Nillmij
decided to establish an independent company for the
production of commercial computers:
Electrologica
, in 1956.
Electrologica
X1, 1957
Therefore, Electrologica could immediately start with a brand new
machine: the
Electrologica X1.
A brand new computer language...ALGOL
In these years between
1952
and
1956,
programming
saw an
evolution, partly because the ever growing complexity of the systems ordered a
more structured operating system, and partly because the scientific, mathematical
approach of programming produced an ever clearer insight in how the job could be
done efficiently. Remarkable in this process was Dijkstra?s discovery of the
Shortest Path Algorithm. Because this evolution was a worldwide happening,
with worldwide contributors, slowly the basis for a scientific computer language
was laid:
ALGOL
was soon to be born.
In 1958, Edsger Dijkstra represented the Dutch MC at the November
Conference in Mainz, a preparatory conference for defining the specifications of
ALGOL In December 1959 ALGOL 60 was defined. In the words of Dijkstra: "?A
miracle was performed by simple mortals?". Finally, the APRIL 1962
Rome convention agreed upon most of the specifications, and in August, 1962 IFIP,
the International Federation on Programming Languages reviewed the report, and
approved of it.
As early as January 1960, after
ALGOL 60
had been defined, the
Mathematical Center started courses in Programming in ALGOL 60, at first in
Holland, and, in 1961, also at Brighton, England.
This
really was the start of a
new discipline of the MC: programming education.
TH Eindhoven: chances and deceptions
In 1962, Edsger Dijkstra was offered a full professorship at the
TH Eindhoven
(Eindhoven Polytechnic). Although chairs abroad were
already named Chairs in Computer Science, Dijkstra strongly opposed to this
approach, mainly because of a lack of scientific substance of the profession.
His chair was that of Professor of Mathematics. His students received at least 3
years of thorough mathematical education, after which period they could specialize
in Informatics. The mathematical training was based upon the
principles of applied mathematics. In the long run Informatics developed the
forms of mathematics it needed by itself.
In 1967, Dijkstra went through a deep emotional crisis. The
thesis of his first Ph.D. student had been rejected by his mathematical colleagues at Eindhoven, who were still contemptuous of computer science. For him
and his wife this period of deep depression was the hardest time of their life,
but he regained his original strength and started to write:
Notes on structured
Programming
.
His colleagues at Eindhoven stayed silent or reacted in an
utterly negative way, but Dijkstra chose the right way of attack: he xeroxed some 20
copies to colleagues in Europe and America. This was the start of a collection of
scientific notes and papers that outnumbers the Bach Werke Verzeichnis, or BWV,
with ease: the
EWD?s
.
Burroughs and total freedom
In
1973
Dijkstra became a
Research Fellow
for Burroughs, and he
reduced his work for the Eindhoven TH to an extraordinary chair. These decisions
enabled him to produce scientific reports, of which he wrote over 500 for
Burroughs, and to travel abroad at will. He was a free man and had the smallest
laboratory of the company: his study.
Austin, start of a new era
On several occasions during his trips he visited the University
of Texas, Austin, where he also gave lectures.
Texas University, Austin.
When, in 1984, he was offered a
full professorship
(Professor
and Schlumberger Centennial Chair in Computer Sciences), he immediately knew
he
would feel at home, so he accepted and he and his wife came to live in America. It was the start
of a 15 year period of teaching, writing and discussing in the center of
evolving programming techniques. Besides, both he and his wife were impressed by
the overall American hospitality.
Cray supercomputer at Texas University, Austin
In 1999, at the age of 69, there came an end to his career as a
professor. Looking back at those 47 years of hard work, of a continuous fight
for a better, simpler and more accurate way of programming, a concern for a
"clarity of notation and exposition", Edsger Dijkstra values most his teaching
of students, and showing people that the work could be done more
elegantly than they knew or imagined. Fascinating others has been his most
rewarding activity.
"
For me, the first challenge for Computer Science is
to discover how to maintain order in a finite, but very large, discrete universe
that is intricately intertwined. And a second, but not less important challenge
is how to mould what you have achieved in solving the first problem, into a
teachable discipline: it does not suffice to hone your own intellect (that will
join you in your grave), you must teach others how to hone theirs. The more you
concentrate on those two challenges, the more you will see that they are only
two sides of the same coin: teaching yourself is discovering what is
teachable".
("My hopes of Computing Science")(EWD 709).
Edsger Wybe Dijkstra died of cancer on August 6 2002 in Nuenen, The
Netherlands.
On the next pages you can find
an impression of his written work and a possible list of his publications,
plus a list of his Honors and Awards.