[SIGCIS-Members] Fw: Bringing up Baby - Oxford DNB Life of the Day

Roger Neil Barton neil.barton at uclmail.net
Tue Aug 11 04:21:24 PDT 2009


FYI
kind regards
neil



Dr Roger Neil Barton
Visiting Research Fellow, Institute of Historical Research
http://www.uclmail.net/~neil.barton/
----- Original Message ----- 
From: <oxforddnb-lotd at oup.com>
To: <ODNBLIFEOFTHEDAY-L at WEBBER.UK.HUB.OUP.COM>
Sent: Tuesday, August 11, 2009 12:00 PM
Subject: Bringing up Baby - Oxford DNB Life of the Day


To read this Life of the Day complete with a picture of the subject,
visit http://www.oxforddnb.com/view/lotw/2009-08-11



Kilburn,  Tom  (1921-2001), computer scientist, was born on 11 August 1921 
at 111 Town Street, Soothill Nether, near Dewsbury in Yorkshire, the only 
son of John William Kilburn, a clerk in a local woollen mill and later 
company secretary, and his wife, Ivy, nee Mortimer. From 1932 to 1940 he was 
educated at Wheelwright Grammar School, Dewsbury, where he did well, 
especially in chemistry and mathematics. He also enjoyed sport, particularly 
running. His headmaster persuaded him to concentrate on mathematics and, 
having earned a state scholarship and a county major scholarship, as well as 
a minor open scholarship, he went to Sidney Sussex College, Cambridge, in 
1940. He followed the mathematical tripos, which was restricted under 
wartime regulations to two years, and was awarded a first class in both part 
one and part two. In later life he said that the examinations at times 
seemed very simple because he had a photographic memory, and could readily 
recall the detail in notes he had taken in lectures.

Pioneering work at Malvern and Manchester

On completion of his studies at Cambridge, Kilburn was invited to move 
directly into a post where his skills in mathematics and related disciplines 
could be of service in the war effort. He was therefore sent on a short City 
and Guilds course in electronics in London, before being posted to the 
Telecommunications Research Establishment (TRE), then billeted at Malvern. 
There he joined F. C. (Freddie) Williams, whose group (working in Malvern 
School's cricket pavilion) solved hard problems in electronic circuitry in 
radar and other important areas. Kilburn later enjoyed telling the story of 
his arrival at TRE, where Williams's disappointment at being given a 
mathematician rather than a 'proper' electronics expert was barely 
disguised. However, Kilburn soon established himself in the group and became 
a key member of the team, making steady progress up through the grades of 
the scientific civil service. His personal life settled down too, and on 14 
August 1943, in a Congregationalist ceremony, he married Irene Marsden 
(1921-1981), a stores assistant, also from Dewsbury. She was the daughter of 
Joseph Marsden, a marine dealer. They had a son, John, and a daughter, Anne.

During the war and immediately afterwards it was well understood that 
scientists and mathematicians would benefit greatly from access to some form 
of general electronic computer. However, no such computer existed. The 
biggest obstacle to creating a general computer was the lack of a way of 
storing instructions and data so that access to them could keep up with the 
speed of the computing circuitry. Williams and Kilburn began to investigate 
the possibility of using cathode ray tubes (CRTs) to store the information. 
Early experiments at TRE demonstrated the successful operation of a 
single-bit memory using the anticipation pulse method, on which a 
provisional patent was lodged in 1946. The bit was stored in the form of a 
charge on the phosphor of the CRT screen, and an electron beam could control 
whether a zero or a one was written. The charge leaked away, however, so it 
had to be rewritten continuously in order that information could be stored 
for long periods. This principle of regeneration was still in use more than 
half a century later.

In December 1946 Williams took up the chair of electrotechnics at the 
University of Manchester and persuaded Kilburn to move with him on 
secondment from TRE, in order to continue their experimental work on the CRT 
storage. At this stage, the major challenge was to find a technique to allow 
large numbers of bits to be held on the CRT in such a way that each bit 
could be written or read at high speed and would keep its value indefinitely 
between rewrites. The continuing relationship with TRE was very useful to 
the project because it meant that Williams and Kilburn had access to a 
steady supply of parts for use in the circuits they were designing. After 
carrying out various experiments, including storing 2048 bits, Kilburn in 
December 1947 completed a report to TRE on his progress. In this he 
presented a definitive explanation of how a CRT could be used as an 
electronic storage device.

However, Kilburn had not yet demonstrated in practice that a CRT could be 
used as the storage mechanism for an electronic computer. To achieve that, a 
test device needed to be built-and the obvious choice was to build a 
computer. So Kilburn, with the assistance of Geoff Tootill, also on 
secondment from TRE, designed and built the simplest general purpose 
computer he could. It had a store size of just 32-bit words, and a very 
limited instruction set (including subtraction but not addition!), but this 
sufficed for carrying out any general calculation. The machine, known as the 
small scale experimental machine (or more affectionately as 'the Baby'), 
worked successfully for the first time on Monday 21 June 1948. The Baby was 
the world's first stored-program computer. It was 16 feet long, contained 
some 650 valves, and weighed half a ton. On that day in 1948, Kilburn not 
only demonstrated a working stored-program computer for the first time, but 
also ran the first proper computer program. This program determined the 
highest factor of a number. The number chosen for the first test was quite 
small, but within days the machine was able to run the program on 218; it 
found the correct answer (217) in 52 minutes, having obeyed about 2.1 
million instructions with about 3.5 million store accesses. Kilburn was 
awarded a PhD for his work on the Baby by the University of Manchester in 
1948.

Further work at Manchester

Having completed the project for which he had been seconded from TRE, 
Kilburn expected to return to Malvern. However, he was persuaded by Williams 
that he should stay on in Manchester. A major reason for this was that the 
government had placed a contract with a Manchester-based firm, Ferranti Ltd, 
to build a full-scale, commercially viable computer to Williams's 
specification, and Kilburn was the key person needed to lead the design and 
construction of the prototype. Kilburn therefore accepted a lecturing post 
at the University of Manchester in 1949 (he was promoted senior lecturer in 
1951 and reader in 1955) and increasingly took over the leadership of 
computer research and development.

Kilburn's experience at TRE and when working on the Baby ensured that he 
approached projects in a very practical way; he understood the value of the 
engineering prototype and the working model. The work on the new computer 
was therefore tackled in stages, starting from the original Baby and 
enhancing and expanding it through a series of working computers. In the 
process, a number of inventions were incorporated, including the 
introduction of index (or address modification) registers, and a two-level 
store based on a number of CRTs and a magnetic drum store. In the late 
summer of 1949 the ideas demonstrated by the enhanced computer, known as the 
Manchester mark 1, were transferred to Ferranti as the basis for the 
computer they produced and marketed in early 1951 as the Ferranti mark 1. Of 
course, the various stages of the Manchester mark 1 were more than an 
experiment in computer design; the machine was also available for others to 
use. One of the earliest users of the evolving Manchester mark 1 was Alan 
Turing, who came to the mathematics department at the University of 
Manchester towards the end of 1948.

Kilburn continued to lead the computer developments at Manchester. His 
approach was always one of problem-solving and experimentation, but at the 
same time he fully recognized the need for practical feedback from real 
users of the systems that he and his team designed. In the 1950s and 1960s 
he led the design of further highly innovative computers, each of which, 
like the mark 1, was exploited commercially. In 1951 he started work on two 
new computers at the same time. One, called Meg, was really an enhanced mark 
1 with floating point arithmetic included, and was probably the world's 
first floating point computer. This design, with some modification, became 
the Ferranti Mercury computer, nineteen of which were sold. The other 
machine, known as the Transistor Machine, was an experiment in the use of 
transistors for computer design, and was probably the world's first 
transistor computer. It too was built commercially, this time by 
Metropolitan Vickers.

These successes were followed by the design of possibly the most influential 
computer produced in Manchester under Kilburn, the Atlas computer. Work on 
Atlas started in 1956, initially without any government or commercial 
backing, and progress was slow. However, in 1959 Ferranti agreed to support 
the project and to provide additional manpower. The goal was to produce a 
really fast computer which could provide a great deal of computing power for 
users. The design used the idea of a controlling program (called the 
supervisor) which would allow the computer to switch between different 
programs. This meant that when one program was held up (perhaps waiting for 
input data from a slow peripheral device), the supervisor would allow 
another program to take over the processing power. A host of other new 
ideas-some, like virtual memory and the 'compiler compiler', invented in 
Manchester-were included in the design of the hardware and software of 
Atlas. A small number of Atlas computers (and its derivatives) were 
produced, with machines at Manchester, Harwell, London University, and 
Cambridge doing sterling service supporting the academic and scientific 
community.

In 1966 Kilburn started his final large computer project, the design of a 
machine called MU5. It introduced new ideas intended to help support the 
efficient running of system software, such as compilers and operating 
systems. The machine was a major influence on the 2900 series of computers 
sold by ICL.

Teaching and university administration

Kilburn's research focus was always the design of innovative computer 
systems (as indicated by the fact that he was named as inventor or 
co-inventor on over seventy-five computer and electronics patents), but he 
was able to find time to make important contributions too in the educational 
sphere. In 1960 he became the professor of computer engineering in the 
electrical engineering department. However, he was aware that the study in 
depth of the design and application of computer hardware and software was 
broader than was appropriate for students studying electrical engineering. 
So in 1964 he established a separate department of computer science. This 
was the first department in the UK to offer undergraduate degrees in the 
subject. Kilburn also became increasingly involved in the wider affairs of 
the university; he served as dean of the faculty of science from 1970 to 
1972, and as pro-vice-chancellor from 1976 to 1979. He took early retirement 
in 1981, with the intention of spending more time with his wife, Irene, who 
was by that time seriously ill with chronic bronchitis. Tragically, however, 
she died on 3 August 1981, just before his official retirement.

Throughout his career Kilburn received numerous invitations to visit 
research institutes and research groups in industry in Europe, the USA, and 
further afield. In practice, however, he accepted very few of these 
invitations because he disliked travelling and being away from home. He did, 
however, visit Australia, the USA, and the Russian Academy of Science on two 
occasions.

Wider interests and final years

Kilburn was in many ways a typical Yorkshireman. He had clear vision, 
enormous powers of concentration, and unshakeable determination when he set 
his mind on something. On the other hand, he was also very quiet and 
unassuming, and unfailingly polite. He enjoyed pursuits that may be regarded 
as very ordinary. When his children were young, he and Irene took their 
holidays in Blackpool, because it provided entertainment for the children 
and was close enough to Manchester to allow him to return with ease in time 
for the first football match of the season. He was passionately fond of 
football and described himself as an avid supporter of Manchester United. He 
claimed that the day in 1968 when he went to Wembley to watch the team win 
the European cup was one of the best of his life. He was equally thrilled 
when they won the treble in 1999 (league championship, FA cup, and European 
champions cup). He also enjoyed music, mainly jazz, and when he could he 
would play the piano for his own amusement.

In the late 1940s, and for several years thereafter, the University of 
Manchester led the world in computer design. Kilburn's work was central to 
this remarkable achievement. His contributions to computer science were 
recognized nationally and internationally, and he received many honours in 
his lifetime, including election to the fellowship of the Royal Society in 
1965 and appointment as CBE in 1973. He received the Royal Society's royal 
medal in 1978. One of the most unusual honours, however, was the 
organization in 1998 by the city of Manchester and the University of 
Manchester of a major celebration to commemorate the 50th anniversary of the 
first time the Baby computer worked. Kilburn played a very active part in 
the events that were arranged, including formally switching on a full-scale 
working replica of the original computer. He died at Trafford General 
Hospital, Davyhulme, Manchester, on 17 January 2001, of bronchopneumonia 
following abdominal surgery. An inquest recorded a verdict of accidental 
death. He was survived by his son and daughter.

Hilary J. Kahn

Sources  R. B. E. Napper, 'Tom Kilburn (1921-2001)', 
www.computer50.org/mark1/kilburn.html, accessed 11 Aug 2004 + The Scotsman 
(22 Jan 2001) + The Times (23 Jan 2001) + The Guardian (23 Jan 2001) + The 
Independent (23 Jan 2001) + Daily Telegraph (27 Jan 2001) + M. Wilkes and H. 
J. Kahn, Memoirs FRS, 49 (2003), 283-97 + WW (2001) + personal knowledge 
(2005) + private information (2005) + b. cert. + m. cert. + d. cert.
Archives  SOUND BL NSA, North-Western Sound Archive, 1CD0011074 BD24
Likenesses  C. Djanogly, bromide print, 1998, NPG · S. Samuels, bromide 
print, 1999, NPG [see illus.] · J. Morris, two photographs, PA Photos, 
London · photograph, 1948 (with Freddie Williams), PA Photos, London · 
photograph, repro. in The Independent · photograph, repro. in The Times · 
photograph, repro. in The Guardian · photographs, repro. in 
www.computer50.org/mark1/kilburn.html
Wealth at death  under £210,000: probate, 1 May 2001, CGPLA Eng. & Wales



========================================================================
©    Oxford     University    Press,    2004.    See     legal    notice:
http://www.oup.com/oxforddnb/legal/

We hope you have enjoyed this Life of The Day, but if you do wish to stop
receiving   these   messages,   please   EITHER   send   a   message   to
LISTSERV at WEBBER.UK.HUB.OUP.COM with

signoff ODNBLIFEOFTHEDAY-L

in the body (not the subject line) of the message

OR

send an  email to  epm-oxforddnb at oup.com, asking us  to stop  sending you
these messages.




More information about the Members mailing list