Willem L. van der Poel

1984 Computer Pioneer Award
________________________________________________

 

 

"For the serial computer ZEBRA"

 

I was born on the 2nd of december,  1926 in The  Hague,  The Netherlands. During the German occupation I took my final examination at the Secondary school (this is now called an Atheneum) in 1944. Universities and other higher school were all closed during
that  very difficult period of the war.  There was no  light,  no electricity,  no  gas,  no public transport and hardly any  food.

Formally  I  was  registered at an intermediate  level  Technical College, but in reality one could not venture out in the streets, risking  to be caught for forced labour in Germany.  During  this period I designed my first computer.  Computers always had had my interest  and  I  had at that time combed every  library  in  the vicinity for literature on computers,  i.e.  mechanical desk calculators,  integrators, slide rules, mechanical carry propagation systems.  I  was aware of the work of Torres Quevedo and  of  the binary  system.  Otherwise  nothing was known to me of  American, British or even German work.  Drawings  of this first computer I have provided  separately to this curriculum. This machine used telephone stepping switches which  could  step both ways.  That would have been  possible  in principle but they were not available in practice. Let it suffice to  say,  that in this machine 3 levels of "subroutines" could be distinguished.  For a sine function the "main" program was on one rotary switch,  the operations such as multiplication were "micro programmed"  on  a  second level and the additions  were  on  the bottom level.

During  the  winter of 1944/45 there was plenty of  time  to study  mathematics  originally meant of some  intermediate  level teaching  degree,  but  used  to advantage when I  entered  Delft University of Technology (Technische Hogeschool,  Delft),  so  to say  the MIT of the The Netherlands.  I enrolled in the study  of physics.  In  1947  D.R.  Hartree  gave  a few  lectures  on  his experience  with the ENIAC and I grabbed the opportunity  to  get acquinted with Van Wijngaarden and Van der Corput who were at the just founded Mathematical Centre in Amsterdam.

N.G.  de  Bruyn offered me a position of research assistant.  At that time I had already developed the plans for a relay computer. With this design I entered a prize contest of the University and  got a honourable mention for it.  This  design was  reworked and a realistic plan was completed in 1949.  The control part was working when I took my degree of Physics Engineering in February, 1950 under A.C.S.  van Heel. The machine was completed by several other students after this design and started work in 1952.

Its  main  purpose was ray tracing for the  design  of  lens systems.  Few  working registers were needed and most of the programming was on stepping switches.  Nevertheless,  it was a  true Von Neumann machine. Instructions could be executed from the live registers  (the  RAM we would say now) as well as from  the  dead
registers (plugboard, the ROM) and the stepping switches. It kept working from 1952 until 1964.  Although it was slow,  it was very reliable  and could beat a human calculator,  working 8 hours  of day time in 16 hours overnight, unattended. Representative speeds were:  Addition 30 sec,  multiplication 45 sec,  division, square
root  and square root of (1 - x square) in 45 sec.  This built-in operation (!) was handy for converting sine in cosines. Its very slowness gave it the name:  Testudo,  the latin for tortoise.

Influential   in   my  further  education   has   been   the participation  in one of the first extra mural courses on  computers in Cambridge,  England under the direction of M.V. Wilkes in 1950, with EDSAC 1 as the "hands-on" machine.

After  taking my engineering degree I was asked to come  and work  with the Central Laboratory of the Dutch Postal &  Telecommunications  services.  Kosten was there and under his  direction the development of the first Dutch electronic computer, the PTERA was started.  This was a vacuum tube machine with a magnetic drum as  main  memory.  The input and output was the usual teletype  5 hole  paper tape.  This machine was put to work in 1953  and  has worked until 1958 when the next machine came in.

During  the construction of the PTERA the ideas for the next machine were brooded upon and a prototype machine was temporarily built from the components of the PTERA. The registers were there, the  magnetic  drum worked and these parts were put  together  to form a machine,  afterwards called ZERO.  It has been extensively described in.  This machine had the basic ideas of the  later ZEBRA,  i.e. the functional bit coding schema, something we would now  call horizontal micro coding.  I still regard the design  of ZERO  the  most elegant one.  Nowadays it could be called a  RISC design.

In  1956  I  got the degree of  Doctor  of  Mathematics  and Physics on a thesis called "The Logical Principles of Some Simple Computers" from the University of Amsterdam.  Promotor was A. van  Wijngaarden,  at  that time the only professor  of  computer science in the Netherlands.

The  ZEBRA (Zeer Eenvoudige Binaire Reken Automaat,  or Very Simple Binary Reckoning Automat) described in the thesis was made under contract by Standard Telephones and Cables Ltd. in Britain. About 55 copies of the machine were made,  which were sold around the world. 9 of them were in the Netherlands. The universities of Groningen, Utrecht and Delft had the ZEBRA as their first machine of the newly erected university computation centres. The last one left  service  about 4 years ago in England.  A  rather  complete description of its code can be found in.  I succeeded Kosten as Chief Engineer and head of the  Mathematical  Department  of the Dr.  Neher Laboratory  (formerly  the Central Laboratory) after his nomination in Delft.

Mention  should be made of my close collaborator Dr.  Gerrit van der Mey.  Van der Mey was a deaf and blind mathematician. The work  of  programming  for automatic  computers  was  excellently suited for him. There was very little literature at that time and all  software had to be developed from scratch.  Van der Mey  has worked with me from 1951 to 1979 when he was pensioned and he has done wonderful things.  Great parts of the "simple code" and  the operating  system for PTERA and ZEBRA were made by him.  Later he made the design of an ALGOL compiler and a LISP system for ZEBRA. For later machines he worked with me on IPL V,  LISP, SNOBOL3 and ALGOL 68 compilers.

For communicating with Van der Mey several special  machines were  built.  Among them was a mechanical typewriter keyboard  to braille translator, a braille telephone and a braille telex. With this  last  machine  we kept contact over  an  ordinary  teletype connection.  For the construction and design of these machine Dr.
H. Mol and I received the Visser-Neerlandia Prize in 1960.

In  1962  I was nominated on an extra-ordinary chair in  the theory and construction of computers in the University of Technology,  my  Alma  Mater in Delft.  In 1967 I left  the  Dr.  Neher Laboratory for an ordinary chair in Delft.  Of course I took  Van der Mey with me to Delft.

The  year  1962 was also the beginning of the Working  Group 2.1  on ALGOL of IFIP.  From 1962 to 1969 I served  this  Working Group as chairman and from then on as a member.

In 1959 the Dutch Computer Society (Nederlands Reken Machine Genootschap,  NRMG) was founded and I served a term of 4 years as treasurer  and a term of 4 years as chairman on the  board.  From 1971  to  1977  I represented the NRMG as member of  the  General Assembly of IFIP.

The last "machine" I designed was a hypothetical one. In the course  of  giving lectures on programming  and  architecture  of machines we needed an ideal machine,  more systematic and regular than any machine on the market at that time.  Furthermore, if one takes  a  particular  machine as the basis for lecturing  on  the  structure  of  instruction codes one would be  biased  towards  a particular manufacturer.  As a result of this the machine SERA 69 was  described  in  a defining report.  Its whole  operation  was formally  given in the form of a program written in its  own  assembler code.  The machine was designed in a committee. I chaired the committee and edited the final report.

In  June  1971  I was appointed member  of  the  Koninklijke Nederlandse   Akademie   van  Wetenschappen  (Royal  Academy   of Sciences).  In  July 1971 I was conferred the honorary degree  of Doctor of Technology from the University of Bradford, England.

In the winter of 1964-1965 I  spent  a  short  sabbatical at Stanford University. In 1975 I have been working several month at the Watson  Laboratory  of  IBM,  Yorktown Heights, and in 1977 I again spent a prolonged period at the University of Tokyo, Japan, where I have returned every three years ever since.

In  the  last  10  years of teaching at  the  University  of Technology in Delft my field of interest has shifted considerably from  hardware to software.  The theory of programming  languages and  their  implementation,   especially  systems  implementation languages  and  portability are my main  interest.  I  have  been  instrumental   in  the  foundation  of  IFIP  WG2.4  on   Systems Implementation  Languages  and  I was the  editor  of  the  first working  conference.  Sidelines of my interests  are  number theory and theory of computation.  The most important publication of this period is on combinators and lambda calculus.

In  the  personal field I have two  hobbies.  The  first  is collecting and designing mechanical puzzles,  interlocking wooden bars etc. The other hobby is music. I have played the piano since I  was  5 years old and I still do.  Later in life I took up  the flute and the bassoon. For more than 18 years I have been playing the  bassoon  in  the  students  symphony  orchestra.   Since  my retirement I have changed to another smaller ensemble, still with the bassoon.

I  retired from a full professorship somewhat  earlier  than usual in May 1988 and I gave my valediction in October, 1988. Fortunately  I  was able to serve another 2 years on a  part-time chair on the Theoretical Aspects of Programming Languages.  Since May  1991 I am fully retired.  At the occasion of my retirement I
was  promoted  to  the  order  of  "Ridder  in  de  orde  van  de Nederlandse Leeuw" (Knight in the order of the Netherlands Lion).