An audio cassette of Georgina Perry's book A computer called LEO as book of the week on Radio 4. 

The book describes the beginnings of Lyon's Teashops, the working environment and the challenges they faced. It discusses the exploration of electronic computers, Lyon's visit and report in America and their work with EDSAC, the beginnings of LEO I, and the progression of LEO Computers Ltd. It concludes by discussing the mergers with English Electric, ICL, and Fujitsu. 

Related links:

Lyons Invoices on Microfilm - Document - Computing History

Research Visits - Computing History

LEO Jobs - L1 Payroll - Computing History

LEO Jobs - L2 Teashop Orders - Computing History

LEO Jobs - L3 Bakery Sales and Valuations - Computing History

Transcript :

[Presenter] We have how Lyons tea shops came to be pioneers in the design and manufacture of the computer.
The reader is Maggie Tagney.
[MT] To anyone who has lived in Britain for more than 30 years, the name of Lyons is instantly
recognisable. The fame of J. Lyons and Company rested principally on its chain of high street cafes
known as the Lyons tea shops. At its peak, the Lyons' empire also included grander restaurants and
hotels in London and other big cities, including the legendary Lyons' corner houses in the
Trocadero, a food manufacturing and distribution business, and a catering service for big events.
J. Lyons & Company had its origins in a family tobacco business established by Samuel Gluckstein,
whose father had brought him and his seven siblings from their native Prussia to settle in London.
In 1872, Samuel started a cigar-making business in Whitechapel Road in partnership with his sons
Isidore and Montague and his son-in-law Barnett Salmon. They shook hands on an extraordinary
agreement. They placed the assets of the new company, Salmon and Gluckstein, in a family fund in
which each had an equal share. The family has held its property in common ever since.
sharing all the proceeds of the business equally and owning houses and even cars communally rather
than individually.
By the end of the century, Salmon and Gluckstein was the world's largest chain of tobacconists with
140 shops. Two years later, the family sold the business to Imperial Tobacco,
but by then they were already well established in the catering trade. The way it came about was
this.
Montague Gluckstein was the driving force of the business and spent much of his time on the road,
promoting the company's tobacco products at trade fairs and exhibitions around the country. It was
this experience that first brought home to him the dreary and standstill methods of the catering
establishments he was forced to patronise. The ordinary man visiting a strange town and wanting a
meal had a choice between a public house, where he would get cold meat, pickles and beer, or a
coffee house with its dirty little horse box-like compartments, untidy shirt-sleeved waiters,
grimy tablecloths, bad food and worse smells.
Surely there was money to be made by offering people at least a good cup of tea when they were away
from home. But the rest of the family thought this was hardly the thing for people engaged in the
aristocratic business of cigar manufacturing. Eventually they concurred,
as long as the catering venture was screened behind a different trading name.
The compromise was to find a partner to run the new venture who was almost family but not quite.
Joseph Lyons, an entrepreneur and salesman, was a distant relative of Isidore Gluckstein's wife,
Rose. Born in Southwark in 1847, Joseph Nathaniel Lyons had begun his working life as an optician's
apprentice, but his quick imagination and gift for selling had led him into a colourful assortment
of other occupations. He invented a device called a chromatic stereoscope,
a combination of telescope, microscope, magnifying glass and binoculars, and sold it for one in
sixpence. He wrote detective stories, musical sketches and songs, and painted watercolours.
His marriage certificate gave his occupation as artist.
Joe Lyons went out and won the catering contract for a large exhibition taking place in Newcastle
in 1887 to mark the Golden Jubilee of Queen Victoria, and that was good enough for the Glucksteins.
J. Lyons and company were in business.
There was nothing tentative about their first venture into catering. Customers in the tea pavilion
at the Newcastle exhibition were entertained by a Hungarian string band. They could choose from a
varied menu and enjoy attentive service, and, of course, they could wash down their meals with a
pot of excellent tea for thruppence. Out of that humble but very important trio...
tea, bread and butter of the best kind sold at a reasonable price, said Montague Gluckstein later,
the foundation was laid of what was afterwards to be the largest catering business in the world.
In 1894, the year J. Lyons & Co was formally registered as a limited company,
it moved its headquarters from Whitechapel to Cadby Hall, a former piano showroom and factory in
Hammersmith Road, conveniently near to Olympia.
Following the earlier model of the tobacco business, the company began to manufacture the products
needed to supply its catering enterprises, beginning with bread and rolls from the Cadby Hall
Bakery. Within 25 years, the site held a complex of red brick factory buildings,
erected fortress-like around a central yard, and employed thousands of workers.
In the last years of the 19th century, ever-increasing numbers of clerical workers were commuting
into central London from the suburbs to work, and they needed somewhere to buy their lunch.
The Lyons directors saw a gap in the market and resolved to open a chain of establishments offering
good temperance fare at economic prices in attractive surroundings and with polite and dignified
service.
With the opening of the first Lyons tea shop at 213 Piccadilly, Joseph Lyons and Company set
standards of service to customers and sumptuousness of surroundings that astonished and delighted
their clientele. Between the drab shopfronts of late Victorian London,
the name of J. Lyons & Co. shone out in hand-carved Art Nouveau lettering,
ornamented with floral swags and finished in real gold leaf against a white background.
Inside, there were gas chandeliers, red damask wall coverings, elegant chairs and marble-topped
tables, silver-plated teapots and fine china. Highly trained waitresses,
originally known by the name Gladys, but later chrismed Nippies, a shrewd PR move,
for their speedy efficiency, eagerly waited to take the orders in made-to-measure uniforms with
starched white aprons. The tea, of course, was only tuppence a cup.
It was an instant success, with queues of customers patiently waiting outside on benches
thoughtfully provided by the management. Within a year, the capacity of the tea shop had to be
increased to cater for 400 rather than 200 customers at a time.
The model was repeated over and over again. There were 37 Lyons tea shops by the end of the century
and 200 by 1925 on prime sites in London alone. Meanwhile,
cafes in Bradford, Manchester, Sheffield, Leeds and Liverpool had been converted into Lyons tea
shops. Whether or not they drove the dramatic social changes that followed the First World War,
they certainly reflected them. Writing in the Daily Mail in 1921,
Lady Angela Forbes observed, For the business girl, not only in the city but in every part of
London, the nearest tea shop is not far away. They share a table with men as naturally as they take
a seat or a strap in tram and tube. From every point of view, and most emphatically from a woman's,
London has changed for the better during the past 25 years. In that metamorphosis,
the tea shops have played a meritorious part.
Working on an even grander scale, the company launched a number of larger and more upmarket
establishments, notably the Trocadero at Piccadilly Circus, a palatial restaurant in the heart of
London's theatre districts, and the Lions Corner Houses. The first corner house opened in Coventry
Street in London's West End and was capable of serving 5,000 people at a time.
There were restaurants catering to different tastes and budgets on the four upper floors, each with
its own live band. There was a food hall on the ground floor, selling tea,
coffee and high-quality cakes and biscuits. You could even get your hair done, book theatre
tickets or avail yourself of that novel instrument, the telephone.
By the end of the 1930s, Lyons had a total workforce of well over 30,000
making it one of the largest businesses in the country.
As the number of outlets to be supplied grew, so did the Cadby Hall site and the range of products
that Lyons made. After bread came tea, cakes,
ice cream, confectionery and eventually ready frozen meals. The food production areas were highly
mechanised. The Lyons' continuous Swiss roll plant, for example, took in raw ingredients at one end
and delivered filled, rolled, wrapped and packaged cakes at the other.
The company supplied almost every grocers in the land with red and green labelled tea,
packed by the quarter pound, foil-wrapped cupcakes and Lyons made ice cream. It developed its own
printing and packaging, laundry and dressmaking and transport and vehicle maintenance operations,
and bought a tea plantation in Nyasaland.
Despite its size, Lyons remained very much a family business. The exception was the company
secretary, George William Booth, who joined the company in 1891.
It was Booth who recognised that concern for quality and value and a fine sense of what the
customer wanted were not always enough to ensure profitability. The problem Lyons faced was simple
to express. Much harder to solve. A typical tea shop customer bought no more than a bun and a cup
of tea, costing a few pence. The profit to the company on that transaction might be as little as a
farthing, barely a tenth of a penny in today's currency. The modest profit margin on each purchase
could easily be wiped out if the clerical work involved in recording and analysing all those
transactions was inefficient. And since almost everything Lyons sold had a limited shelf life.
It was essential to have an ordering and distribution system that accurately matched supply to
demand. The same applied to the retail business. Lyons supplied goods such as tea and cakes
directly to small shops, dealing with as many as 40,000 orders in a week.
In the checking department at Cadby Hall, the clatter of machinery was relentless, light fell
through the high windows on row after row of workers, bent to their identical tasks.
Seated at their desks in ruler-straight rows, the clerks tapped away at their borough's mechanical
calculators, separated from one another by partitions erected to reduce distraction.
The adding machines, solid constructions of steel and varnished wood, had up to a dozen columns of
numbered keys to input the figures and a crank on the side to sum the totals. Like a cash
register, they printed out a record of the calculations on a roll of paper.
The 300 clerks in the department, most of them girls not long out of school, had but a single job
to do, to add up the totals on the waitress's bills from the 250-odd high street tea shops and to
check them against the cash takings banked by the shops. A squad of office boys kept them supplied
with sets of bills received from the tea shops that morning in locked leather bags and sorted into
numerical and alphabetical order by the office juniors. The senior clerks and managers,
invariably male, stalked the aisles between the desks in their sombre suits,
ensuring that every fashionably waved head was bent to its task. It would be their duty to follow
up any discrepancies revealed as the streams of numbers gradually unrolled.
The checking department was one of three central offices at Lyons, supervised in those pre-war
days by a young manager called John Simmons. Yet as Simmons surveyed the room full of clerks and
their clacking machines, all he saw was a waste of human intelligence. Punching a borough's
calculator could be worse drudgery even than unskilled factory work. At least on an assembly line,
he mused, you could chat to the next worker or let your mind wander while you carried out a
repetitive task. Mechanised clerical work demanded total attention, but granted no intellectual
satisfaction in return.
Simmons began to dream of the day when machines would be invented which would be capable of doing
all this work automatically.
Such machines would free managers such as himself from marshalling their armies of clerks and allow
them to examine the figures, to digest them and to learn from them what they had to tell us of
better ways to conduct the company's business.
[Presenter] Maggie Tagney was reading A Computer Called Leo by Georgina Ferry, which is abridged by Peter
Everett. Tomorrow, Lion staff take a boat for America in their quest to find whether a machine can
do clerical work. The producer in Bristol is Francis Burns.
[MT] The term scientific management was invented by a former Pennsylvanian steel worker called Frederick
Winslow Taylor. He believed that factory managers should analyse every task to reduce it to the
minimum number of essential movements, the one best way for a specialised worker to complete it.
His methods were attacked as inhumane and he died in 1915 a disappointed man.
But during the First World War, when labour was short and productivity at a premium, a new
generation of disciples picked up his ideas. Factories were invaded by eager young men with
clipboards and stopwatches carrying out time and motion studies.
Technological change was led by the United States. In that fast-growing country,
labour was scarce, and there was a great enthusiasm for machines that could increase productivity.
The typewriter, patented in America in 1868, made it quicker and easier to communicate and keep
records. Devices that we would hardly think of as inventions today, such as index cards and
vertical filing cabinets, revolutionised record-keeping. Adding and calculating machines relieved
clerks of the necessity to be accurate calculators themselves.
Efficiency had long been a priority at Lyons and Company, where the factories were organised very
much according to the principles of Taylorism. Each was laid out to handle the particular kind
of... cake, pie, bun, loaf or bread roll in which it specialised. Under the direction of a planning
office, every operation was time and motion studied to arrive at a fair, efficient estimate of the
number of staff required and the time entailed in making the product, which in turn partly
determined its selling price.
In 1923, the Lions Company secretary, George Booth, persuaded the board to add some intellectual
rigour to the company's management by recruiting some bright young minds. One of the five young men
who constituted this new class of management trainee was John Simmons, who had just received a
first-class degree in mathematics from Cambridge. Born in Ceylon,
now Sri Lanka, in 1902, John Simmons was the son and grandson of missionaries.
Unobtrusive in appearance, he always dressed soberly in a jacket and tie. On his arrival,
he was put to work in a department where black-coated clerks still stood at Dickensian high desks,
entering figures in huge ledgers by hand.
A quiet, austere and intellectually exacting man, Simmons found his spiritual home at Lyons.
He was to remain with the company for 45 years. He immediately began to apply his analytical skills
to the task of increasing efficiency. He streamlined and simplified, extended the use of office
machines wherever they made economic sense, analysed the works of the clerks and then told them how
to do it better. After a few years, the board agreed to let him set up a Department of Systems
Research, a team of analysts who would investigate inefficiencies and bottlenecks in the company's
office systems and propose solutions.
Everything was specified, from the value of the energy needed to bake a loaf of bread to the
thickness of the jam spread on the Swiss rolls. The task of checking actual performance against
these standards produced useful information, but was short on job satisfaction for the dozens of
calculator operators, who were doing nothing but multiplying, adding, subtracting and writing down
the answers by hand.
In 1935, Systems Research received a plea for help from the wholesale bakery sales department,
which supplied bread and cakes directly to shops all over the country. They found that clerks were
drowning in paper, copies of invoices and packing notes, all of which needed to be filed.
By using an early microfilm camera, called a recordac, to make the only record of customers'
orders, they were able to use the same paper order for pricing and valuation,
then as a packing list, and eventually to return it to the customers and invoice, leaving nothing
to file. It was the first commercial use of microfilm anywhere in the world.
When the Second World War came, Lyons and company shared the indomitable spirit of London,
serving tea throughout the Blitz in their surviving tea shops. Seventy were destroyed by bombs.
And entertaining soldiers on leave with the gaiety of their corner houses. Part of their
headquarters, Cadby Hall, became a depot where volunteers packed rations for the troops.
Most of the Lyons staff who had been on active service returned to their old jobs after the war.
But there were no more nippies. Labour shortages had forced Lyons to convert the tea shops to self
-service cafeterias. So in 1946, with an exhausted economy and severe currency restrictions in
force, it was absurd to think that Lyons might buy an electronic brain from America.
And yet here, in John Simmons' office, were two of his most trusted lieutenants,
Oliver Standingford and Raymond Thompson, babbling excitedly about this new device they'd heard of,
a calculator made of glass valves. It would be thousands of times faster than any mechanical
design, as it would have no moving parts. All of its operations would be carried out by the
movement of electrons in wires. The two men wanted permission to go to America to investigate this
amazing machine. Simmons consulted his mentor, the 78-year-old company secretary George Booth,
who expressed the indulgent view that youth should be given its head, even if that head contains
unusual ideas.
So in the spring of 1947, Thompson and Standingford boarded a ship for the five-day crossing of
the Atlantic. However, once there, they found nothing to match the systems that had been put in
place at Lyons by Simmons and his teams. But at last, they met Dr Herman Goldstein,
who during the war had worked at the University of Pennsylvania on a machine called ENIAC.
ENIAC had been designed to calculate firing tables for artillery. It would tell the gunners how
high to aim their weapons, given a target at a certain range, calculated on the basis of the weight
of the shell, its velocity on leaving the muzzle, the wind speed and direction, and the air
temperature and density.
Brought to life by teams of engineers working 16-hour days, it was a monster.
It weighed over 30 tonnes and incorporated almost 18,000 valves. It could complete each step in a
calculation in less than a millionth of a second.
Goldstein described all this and finished by giving Thompson and Standingford a list of everyone he
knew about in the United States who was doing serious work on electronic computing. Then,
enjoying their astonishment, he dropped his bombshell. And, of course, there's the mathematical
laboratory in Cambridge, England. The two men had come 3,000 miles to find out that a computer was
already under construction a couple of hours' drive away from Lyons' headquarters.
Thompson and Standingford returned to New York and the boat home in a state of intellectual
euphoria. While some might have used the cruise home on the Queen Elizabeth as an opportunity to
relax, they lost no time in starting to write the first draft of the report they would be
presenting to the Lions board. In it, they expressed their enthusiasm for an electronic calculating
machine. a device that could store data and instructions, that could perform sequences of
calculations on its stored material automatically, comparing words or figures in its memory and
reacting to differences, and printing out results. They explained how it could be used for sales
invoicing, for typing form letters and for managing the payroll. The information it needed,
employees, names and rates of pay, for example, could all be stored on magnetic wire or teleprinter
tape and used again and again, while each week a separate input tape or wire would tell it what
hours had been worked. A third instruction wire containing the programme would be played into the
computer, the calculations would be performed and the computer would then print the pay slips.
When Thompson and Standingford returned from their trip to the United States in June 1947, they
arranged to visit the Cambridge University Maths Lab. Its director was Dr Maurice Wilkes,
who had been to America the year before, had seen ENIAC at work and had drawn up plans to build
such a machine for himself. It would be an expensive project, but to his delight Wilkes
was offered thousands of valves free from government surplus stores. The Ministry of Supply was
equally delighted to have disposed of them with so little trouble.
The first problem that Wilkes tackled was how to store programmes and data. ENIAC had the memory of
a goldfish. Each time the engineers wanted to run a new calculation, they had to set up the
programme afresh by plugging wires into sockets, a process that could take a whole day.
Wilkes wanted to try a different approach to the memory problem, an idea that had been proposed by
ENIAC's chief designer but never implemented. It was to keep the stored numbers circulating as
trains of pulses until they were needed. To do that, he had to slow them down.
Now, sound travels through liquids much more slowly than electricity passes through wires. Pulses
of electricity can be converted into pulses of sound by causing a quartz crystal to vibrate.
The resulting ultrasound signal can then be transmitted much more slowly through a tube of a dense
liquid such as mercury. When the pulse train reaches the end of the tube,
it can be converted back into an electrical signal, trigger another quartz vibration and be sent
back through the tube to slow it even more. The trick can be repeated as many times as required,
so that a single 5-foot-long tube can hold more than 30 numbers at a time.
This makes mercury delay line storage massively more efficient than the valve storage used on
ENIAC.
Looking at a mercury delay line today, it seems scarcely credible that it could have been part of a
computer. It looks as if it ought to belong to a previous era of heavy engineering.
But heavy or not, it had to be a miracle of precision. Working back from the number of pulses he
wanted to keep in circulation, Wilkes calculated that each line, or tube,
had to be exactly 5 feet 4 inches long, precise to 1,000th of an inch.
The steel tubes each held half a tonne of mercury.
After several attempts, Wilkes had a memory circuit running satisfactorily by February 1947.
He called his computer the Electronic Delay Storage Automatic Calculator,
EDSAC.
When Thompson and Standingford arrived from Lyons, EDSAC was still 12 to 18 months from completion,
but Wilkes gave them to understand that it was only lack of funds that hampered his progress. Being
businessmen, they asked him how much, and he said £2,000. The two men went away and put the
finishing touches to their report. In it, they set out what they saw as the monumental opportunity
facing lions. No one else in Britain, they emphasised, had realised the far-reaching possibilities
of electronic machines. One of the first to see the draft report was John Simmons.
Here, it seemed, was a description of the machine he had dreamed about for almost 20 years.
Somehow, the company must be persuaded to take it on. Here, for the first time,
he told the board, there is a possibility of a machine that will be able to cope at almost
incredible speed with any variation of clerical procedure. He estimated that the cost of building a
computer could be recovered within two years by the savings it would achieve. Lyons must build its
own machine.
The board considered the report on 20 October 1947. Three days later,
the chairman, Harry Salmon, told Simmons that he was prepared to recommend a donation of £3,000 to
the Cambridge laboratory, with no strings attached. Furthermore, if and when the Cambridge machine
had been shown to work, he would spend up to £60,000 on building a machine for Lyons and Company.
Harry Salmon, the chairman of J. Lyons and Company, looked up as the clerk came apologetically into
the boardroom to hand him a note. It was from his comptroller, John Simmons, and reported that the
Cambridge Maths Lab engineers had telephoned that morning, the 8th of May 1949,
to say that their experimental computer, EDSAC, had run its first successful programmes.
Simmons was now asking the board to authorise the building of the Lyons computer. There was not
much need for discussion. Mr Harry had given his word two years before that Lyons would fund the
computer if the Cambridge experiment worked. The heads around the table nodded. Mr Harry unscrewed
the cap of his fountain pen, wrote on a piece of paper and passed it to the clerk. The clerk
brought the note back to Simmons. It bore the single word, yes.
John Simmons had been preparing the ground. In his mind, the project had already begun. He first
addressed himself not to the engineering aspects of the enterprise, but to the subject he had
developed into an art form at Lyons, analysing the requirements of each clerical job and planning
how best to meet them. Simmons thought it important enough to put someone onto this full time.
He turned to the department he had created to solve Lyons' clerical problems, the Systems Research
Office, and to the man currently running it, another former management trainee of the 1930s,
David Caminer.
During the war, Caminer had fought with the 8th Army in North Africa and had lost a leg.
He had the reputation of being not a comfortable person to work for. He did not suffer fools and
drove his staff relentlessly, although no more relentlessly than he drove himself. So Caminer and
his assistant began to put together the first pilot programme of clerical work, which they hoped to
test on the Cambridge machine, EDSAC, as soon as it was working.
They chose a section of the payroll because it would be a good test of the computer's capabilities.
It involved both data that did not change from one week to the next, such as an employee's name,
national insurance number, tax rate and rate of pay, and variable quantities, such as hours worked
and any bonuses due. It had to print out a payslip showing gross pay, deductions for tax and net
pay. It had to calculate down to the last penny the number of coins of each denomination that would
be needed to make up the pay packets of all the employees in the system. It had to carry forward
totals for the tax year to date. And most importantly, it had to work reliably to a regular weekly
deadline. This was almost certainly the first commercial programme ever run on a computer.
Unfortunately, the first version never worked. It stopped after two minutes, with no indication of
what had gone wrong. It quickly became clear that the size of the programme and the volume of data
required for an application such as payroll would be beyond the capacity of EDSAC's memory.
Any Lyons computer would need to have a store that was twice as large. David Caminer began a long
-lasting battle with the mathematicians to get them to realise how much greater the memory demands
of a clerical system would be. Mathematical jobs were simple, whereas the clerical jobs were
terribly complicated, he points out. For an item of sale in cakes, the computer had not only to
work out the value of those cakes and how much profit was going to be made on them, but also to
make sure that the chap who bought the cakes paid up for them. And from the same data we had to
know how many cakes the bakers were going to make, how much they were allowed to spend on
ingredients, how much energy they would use. And the final step was to compare what they had
actually spent with what they were supposed to spend and to what extent the company was making the
profit it had budgeted for. All that had to be built in. If the machine had storage capacity
sufficient for all we wanted to put into it, we could have done that without a great deal of worry.
But of course it had negligible storage space. So how to use the storage space successively to
accommodate all these things became a very nice problem, which I think we solved quite
conclusively. We were able to do things with tiny stores that now take megabytes,
gigabytes.
Well before EDSAC had fully proved its worth, Simmons had persuaded Mr Harry that it was time that
Lyons thought about engaging an expert for the building of our machine. Accordingly,
he placed an advertisement in the scientific journal Nature, and within a few days he had an
application on his desk from a Cambridge physicist; 29-year-old John Pinkerton.
Pinkerton had worked on radar research during the war and afterwards studied ultrasonics at
Cambridge. There, he had attended a lecture by Professor Douglas Hartree, who had been to see ENIAC
operating in America. It may well be, Hartree had said, with what now seems like understatement,
that the high-speed digital computer will have as great an influence on civilisation as the advent
of nuclear power.
Pinkerton's job interview at Cadby Hall lasted from 10 in the morning until 5.30 in the afternoon.
The Lyons managers explained that they wanted someone to build a version of EDSAC, which as
Pinkerton already knew was not yet working. He thought they were perhaps a little mad,
but it was going to be great fun.
Pinkerton is now widely regarded as the finest computer engineer of his generation. He combined a
profound understanding of the theoretical background to electronics, with a skilled pair of hands.
Most important of all, he was an effective team leader, setting high standards by example rather
than exhortation.
He joined Lyons at the beginning of 1949 and set up a workshop above the gatehouse in a part of the
Cadby Hall site which had formerly been a Catholic seminary. There he began to experiment with
electronic circuits built from second-hand components and made a number of trips to Cambridge to
learn all he could about building a computer.
After the historic board meeting in May 1949 that gave the project the final green light,
work started in earnest on building the machine in a huge room on the second floor of the WX block
at Cadby Hall that Simmons had earmarked long before.
Always a stickler for the exact use of language, Simmons insisted that the word computer should be
used only to refer to the arithmetic unit, the part of the machine that carried out calculations.
The system as a whole, he suggested, was the electronic office.
Three months later, he noted in his diary that his pet project now had a name. It seems to be
generally agreed that it would be appropriate to christen the electronic machine project LEO,
the Lyon's Electronic Office.
The LEO development team formally came together in May 1950, with Caminer and his assistant moving
from systems research to join Pinkerton and the engineers. It took very little time before the
demands of designing and programming the new machine became all-absorbing.
Computers in the late 1940s were big, expensive and time-consuming to build and to run,
but essentially they consisted of exactly the same elements as a modern PC. An input device,
a store or memory, a processing unit and an output device. The first decision Pinkerton made to
diverge from the EDSAC design was that there would have to be twice as much memory,
64 mercury delay lines instead of 32. There was another problem with EDSAC.
With its 3,000 valves, it was not very reliable. When it broke down, the engineers would simply
replace valves one by one until they found the dud, a tedious and time-consuming process.
Pinkerton decided that for ease of maintenance, he would make the computer out of interchangeable
units. Each unit could hold up to 28 valves with their associated circuitry.
The units slid easily in and out of metal racks, holding up to 12 units each.
If a unit failed, it was the work of a few moments to remove the whole unit and replace it with
another.
The engineering team very quickly realised that there was so much work to do designing and testing
units that, like the Mercury delay lines, they would have to be manufactured by an outside
supplier. As it happened, one of the team, Ernest Kaye, had a friend who was the director of Camper
and Nicholson, boat builders on the south coast. I was visiting one day, he recalls,
and he said, You wouldn't know anyone who wanted any electrical work done, would you? We've got
teams of electricians, and during tough times they're sitting twiddling their thumbs. So the boat
builders became part-time computer engineers.
Given the calibre and dedication of Pinkerton and his team, the task of producing what was
essentially a copy of EDSAC was not intrinsically difficult, and they completed it within two
years. The real challenge for the engineers was to build the additional parts of the system that
EDSAC did well enough without. What LEO needed were high-speed methods of delivering data and
programmes to the machines and of producing the results. Originally, they had thought the answer
might lie in recording the input and output data in binary form onto spools of magnetised wire,
a technology that had been used for some time to record sound for broadcasting. They aimed to
develop the necessary equipment in partnership with a company that had relevant experience and
chose the communications company Standard Telephone and Cables. STC were keen to be involved.
Its engineers suggested using magnetic tape. and undertook to develop tape drives that would load
data into the computer at high speed, converting it from decimal to binary notation in the process,
and similar output devices that would link to printers or card punchers. Lyons was assured that
these would be delivered by May 1951, when the whole system was due to go into operation.
By the end of 1950, it was clear that the devices STC had promised were far from ready.
There were two problems. The first was mechanical. The company had not discovered a way to stop and
start a spool of tape running at high speed with sufficient accuracy. More seriously,
the decimal binary converter was simply unreliable. STC was using its contract with
Lyons to fund a series of experiments, and the experiments were failing disastrously.
Pinkerton decided that in the short term they should use tried and tested technology,
paper tape or punched card readers, as the basis of slower speed inputs and outputs.
The programmers were desperate to run something to demonstrate the worth of the machine to the
company. David Caminer chose a job known as bakery valuations. It involved calculating the value of
the week's output of bread, cakes and pies from the Cadby Hall bakeries, taking into account
materials, labour and indirect costs such as the power to run the ovens, calculating the value of
bakery products leaving the Cadby Hall site in terms of factory costs, retail price and profit
margin, and calculating the value of the products held in stock. This process conventionally
occupied 50 hours of clerk's time per week.
Camina specified the requirements of the job and drew up a flow chart to show how the different
parts of it related to one another. Camina believes Lyons was the first company to use charts in
that way. The programme ran successfully for the first time on 5th September 1951,
but only as an experiment. Less than three months later, LEO took over bakery valuations from the
clerks who had previously done the work, and became the first computer in the world to run a
routine office job.
Using the slow speed inputs and outputs, all that Lyons had available for the first year or two,
it did the job in less than five hours, plus eight hours to punch the data onto paper tape.
The successful Baker Evaluations job demonstrated that a computer could be relied on to work week
in, week out. Not exactly without fail, but with failures kept to a manageable level.
And they knew what needed to be done to build up to the speeds that would be necessary to make
bigger jobs, such as the payroll worth doing. It was at least the end of the beginning.
[Presenter] A Computer Called Leo by Georgina Ferry is read by Maggie Tagney and abridged by Peter Everett.
The producer in Bristol is Frances Burns. Tomorrow, news leaks out about LEO,
the electronic brain that'll do your thinking for you.
People who owned Lion's tea shops. It's the 1950s, and LEO,
the Lyon's electronic office, is more than a dream. And the man in charge of the far-sighted
project is beginning to see that the machine that computes the company payslips has vast potential.
The reader is Maggie Tagney.
[MT] For all his public austerity, John Simmons soon allowed himself to recognise that the LEO project
was momentous and that he personally had set in motion events that would change history.
Raymond Thompson, who was responsible for the day-to-day management of the project, began to
believe that its success could have commercial implications beyond improving Lyon's accounting and
management practices.
For fear of giving away information potentially useful to other computer manufacturers, he
discouraged his colleagues from publicising the progress of the computer too soon.
But as news of LEO's commissioning leaked out, it was the scientific computer users who began to
queue up to ask for time on the machine. After all, LEO was a precious resource.
It was one of only three working computers in Britain at the time. The Ministry of Supply offered
£300 for work on the problem of calculating shell trajectories. Thompson replied that although the
sum was too low for the work involved, Lyons would accept the offer because it would provide us
with experience in operating the machine on a different kind of work. Each night,
LEO would be encircled by a red tape while it ran the defence-related programmes. Only those who
had signed the Official Secrets Act were allowed near the computer as long as they were running.
Within a year, LEO was also working on weather forecasting for the Meteorological Office,
tables of annuity values for the Institute of Actuaries and guided missile trajectories for de
Havilland. It was not what LEO was designed to do, but at least it meant that the computer was
running and beginning to earn its keep.
At last, Simmons began to think about introducing LEO to the wider world. He thought Lyons should
make a public statement before American companies started making claims about the suitability of
their own systems for office work. Unfortunately, he had not anticipated the ultimate failure of
the subcontractor, STC, to build reliable tape decks to provide the high-speed inputs and outputs.
Without them, LEO could make no claim as an economical alternative for large-scale clerical work.
The team fell back on the latest electromechanical equipment, paper tape and punched cards,
and the magnetic tape experiment was eventually abandoned altogether.
LEO was eventually pronounced finished on Christmas Eve 1953. Not only was the computer complete,
with smart blue doors concealing the racks of valves, but the room was decorated and ready for use.
The Lyon's board threw a party to celebrate.
The first job to benefit from the completed machine was the payroll. To calculate each worker's
wages and print the payslip, the computer took one and a half seconds, twice as fast as they'd
hoped. Within a few months, nearly 10,000 Lyons workers had their pay calculated by LEO,
and it always produced the payslips on time.
A week after the payroll programme began running, in February 1954, Lyons held a press conference
and demonstration to introduce LEO to the world. The response was enthusiastic.
From the evening news, puzzled, LEO the brain will do your thinking for you,
to the new statesman, into the industrial field has come a vast new array of workers, the slave
electrons which do not pay union dues. The press gave generous space to the new development.
Most opted for the gee whiz angle, focusing on Leo's perceived capacity to outdo the human brain.
In a remarkably prescient aside, Richie Calder of the New Statesman speculated that if it were
possible to replace valves with transistors, much more compact and reliable, but then still at an
early stage of development, an electronics enthusiast can conceive a cub LEO as a desk computer.
The coverage of LEO for the first time brought into millions of homes the idea that computers could
do the kinds of jobs that ordinary people did every day. This inspired conflicting reactions.
LEO's senior engineer, Ernest Kay, remembers describing the project to a friend of his parents,
happily outlining the work LEO could do and how it could make the drudgery of clerical work a thing
of the past. She listened very carefully and at the end of it she said,
But we must destroy it.
What of the Lyons employees whose jobs might have been at risk? The workforce was not unionised,
but there was an active staff association and the Lyons Board took a paternalistic attitude to
those it employed. For his part, John Simmons felt it was essential that if the computer was to be
accepted throughout the organisation, staff should feel it was there to help them, not to supersede
them. He approached the problem with characteristic humanity. From the moment the board gave the go
-ahead to build the computer, he took the workforce into his confidence, briefing the staff on the
LEO project and holding demonstrations for everyone from board members to clerical staff.
Simmons had the board's permission to promise that no one would be sacked when the computer began
working, and he kept his word. In practice, the computer took so long to come into operation that
normal levels of staff turnover made a smooth transition from manual to computer working possible.
At the same time, new jobs were created in data preparation and computer operation.
Simmons' insistence on making staff feel that they were part of the new development paid dividends
when he decided to introduce the computer to the most visible but also the most conservative part
of the Lyon's empire, the tea shops. The tea shop chain was struggling to regain the profitability
it had enjoyed in the 1930s. Self-service had become the norm. Formica had long since replaced
marble on the tabletops, and a Lyons tea shop interior now resembled a works canteen more closely
than an elegant dining room. This impression was heightened by the poor quality of the food.
Rationing remained in force until 1954, and the greatest selling point of a Lyons tea shop meal,
that it was good value for money, no longer applied. To make up for the shortage of skilled cooks,
Lyons opted to prepare all the foods centrally in the highly mechanised Cadby Hall kitchens,
and deliver it in individual frozen portions to the shops, where meals could be quickly reheated.
Although they also sold ready meals in packs to the public, the term they coined for their frozen
food products, 'frood', somehow never caught on.
There was also competition from newer styles of restaurants offering fast food. Lyons itself opened
the doors to the hamburger by buying the rights to the Wimpy franchise in the United Kingdom. By
the end of the 1960s, wimpy hamburger restaurants outnumbered tea shops by more than two to one.
Simmons and his team saw LEO as an opportunity not only to take the drudgery out of managing the
tea shops, but also to help them improve their performance. They agreed that after the payroll,
writing a programme for the ordering and distribution of goods for the tea shops should be a
priority. They began by leaving their offices and observing the round-the-clock operation that
went on in the manufacturing and delivery sections of the Cadby Hall site.
The next step was to drive around London visiting tea shop manageresses and asking them what might
make their work easier. The worst crime a manageress could commit,
it transpired, was to order more than she needed. As a result, manageresses tended to order too
little. Helping them to predict their needs more accurately would obviously help the business.
The solution they found was years ahead of its time. Each afternoon, at a set time,
a data entry clerk, sitting at her card punch, would telephone the tea shop manageress.
Using standard product codes, the manageress would tell the clerk what changes she wanted to make
to her standard order. The clerk... wearing a headset to free her hands, would punch the changes
onto cards as she heard them. The stack of cards would then be sorted, delivered to the computer
room and run in parallel with the standard orders. Not only were the manageresses saved from their
tedious daily form filling, but their orders were moved several hours closer to the delivery time.
LEO, meanwhile, recorded the sales value of the goods delivered to each shop, so that they could be
compared with the takings at the end of each month. For a range of different products,
LEO could print out the 10 best and 10 worst performing shops. Managers would then be in a position
to investigate the factors that were affecting performance. In this and other ways,
LEO was designed not just as a fast data processor, but as a management tool.
Soon, the daily reports filed by manageresses to their supervisors began to include peons of praise
to LEO.
Unfortunately, the higher levels of management seemed less appreciative. The tea shops division was
possibly the most conservative part of an essentially conservative organisation. The other problem
was that LEO did not, in the early days at least, save money in direct clerical costs.
Costs were, if anything, higher. Even with the standard orders, the demand for data input was high.
At the same time, as long as there was only one computer, a number of human clerks had to be
available in case LEO ever failed. But the idea of placing a computer at the heart of an ordering
and distribution system was wholly novel, and tea shop's distribution caught the public
imagination. The juxtaposition of the high-tech electronic machine with the homely comfort of a
cup of tea and a penny bun... made for an intriguing picture. Visitors,
treated to the full tour of the operation, wrote up their experiences in admiring terms.
One remarked that each girl had time to manipulate a cigarette, a cup of tea or a powder compact in
addition to the headphones and card punch and added that the general atmosphere was one of
nonchalant super efficiency.
John Simmons could have stopped there and the Lyons' operation might have remained no more than a
quirky footnote in the history of computing. But his evangelising instincts drove him further.
As early as 1950, he began to plant in the minds of the Lyons' board the idea that their business
might diversify from catering into making computers for sale.
From the moment the members of the board had accepted the principle of the computer, they had
agreed that a second machine would have to be built as backup for the first. Computers were
unreliable. You needed at least two in case one of them broke down. Now it was clear that the
second machine could be substantially improved. Simmons's thinking was entirely logical.
If the LEO team were to build a better machine, why not build two, making LEO one obsolete?
And if two, then why not build more? to be leased or sold to others.
The family was understandably hesitant. The investment required would be considerable at a time
when the profitability of the core catering business was in question and the size of the market
completely unknown. Simmons persisted, however, and finally persuaded the board to agree that the
computer operation should as soon as possible be incorporated as a separate subsidiary company,
LEO Computers Ltd. The company was duly
set up in November 1954.
LEO was about to take its first steps out of the protected environment in which it had been
nurtured, into a world of cutthroat competition in which only the fittest would survive.
But in the early 1950s, Lyons appeared to be the only business in the country that had even thought
about using computers. The computer specialists themselves promulgated the view that computers were
specialist tools for scientific users. The Cambridge pioneer Douglas Hartree opined that the
country's needs could be met by no more than half a dozen computers administered by skilled
mathematicians. It is a quote that has been gleefully reproduced in every computer history since.
[Presenter] But how will the pioneering scientists who made it make out in the wider world of international
business? The reader is Maggie Tagney.
[MT] With the launch of LEO Computers Limited, John Simmons felt he could begin the task of persuading
British business to take computers seriously. A less pragmatic motive for a new and risky business
venture, it is hard to imagine. But Simmons was quite explicit about his educational aims.
LEO Computers was founded, he said, to build computers and to be the means of placing our team's
practical knowledge and experience at the disposal of others.
LEO II, designed by John Pinkerton, differed very little from LEO I in its architecture and
engineering principles. It still read its inputs from paper tape and relied on mercury delay lines
for storage and glass valves for its electronics. Pinkerton was able to make considerable
improvements in its performance, however. By using shorter delay lines,
the bulky memory units took up much less space. A mere 50 feet by 25 was enough for the whole
installation. This and other improvements produced a machine that operated four times as fast as
LEO I.
Begun in 1954, LEO II did not come into operation until May 1957.
twice as long as they'd planned. It was the first sign that the cottage industry approach that had
built LEO I was inadequate for a manufacturer hoping to compete for sales in the wider
marketplace.
Lyons had set up LEO Computers as a commercial manufacturer, but failed to provide anything like
enough resources. The engineering team still numbered fewer than two dozen, and they were torn in
two directions. While trying to build a machine that could do anything LEO I could do,
and do it quicker, they were also having to research and develop new devices that future customers
might want, such as fast printers that could handle the letters of the alphabet as well as numbers,
and document readers that would remove the need for punched tapes or cards.
Visiting the United States, the team leader, Raymond Thompson, discovered that IBM had computers in
production with a new form of memory, called the magnetic core store. This consisted of thousands
of cores, each not much bigger than the head of a pin, made of a ceramic called ferrite.
Electrical connections to each of the cores either magnetised or demagnetised the individual
elements, representing the binary zeros and ones of the saved information.
Magnetic core stores had first been devised as part of a government project to boost America's air
defences, so IBM was benefiting from billions of dollars of Cold War funding.
The LEO engineers set out to design a magnetic core store for their new machines, although the
first seven were built using the old Mercury delay line systems. Meanwhile,
the programmers stayed busy, as well as payroll and tea shops distribution,
there was now a programme for bakery rail orders, which handled the orders from travelling salesmen
for cakes to be distributed by rail to grocery shops around the country. There was a programme for
tea blending that valued the dozens of varieties of tea that Lyons included in its various blends,
Red Label, Green Label and so on. The ice cream division required a number of programmes to match
supply and demand around the country. They even had a programme that incorporated weather
forecasts.
In 1955, the annual conference of the Office Management Association was devoted to the subject of
electronics in the office. Although interested, the audience was inclined to be sceptical.
There were chuckles when one lecturer said... the idea of a computer as an electronic brain.
Think of it as an expensive moron, infinitely slow and infinitely stupid,
which can make mistakes faster than you ever dreamed possible.
John Simmons came away feeling that somehow they had failed to make the impact he had hoped for.
Not long after the conference, a writer in the association's journal sneered. A potential computer
user needs to have some confidence in his own judgment if he is to buy his computer from a tea
shop. Nevertheless, with the arrival of its first two orders from outside Lyons,
LEO had the opportunity to show the business world that it was serious. Through family connections
in the tobacco business, the Lyons chairman obtained an order from WD and HO Wills.
Soon afterwards, another order came through from the steel company, Stuarts and Lloyds.
With these early customers, LEO established an approach to sales that contrasted vividly with the
hit-and-run style of IBM. It stemmed directly from John Simmons' philosophy that a computer
installation was a system fully integrated into the operation of the business it served.
Once LEO obtained an order, a senior systems person from the company would spend up to three months
working closely with the customer, doing everything he could to find out how the existing system
worked and what the company really needed. Meanwhile, key members of the customer staff would be
selected and sent on a five-week programming course. Once trained, they would help to write the
initial programs for their company. Training, maintenance and operations staff was also part of the
package.
Orders were coming in, but they came slowly, partly because there was no real attempt to develop a
brand image for the computer. It was some years before LEO computers began to distribute anything
like the professional and persuasive sales literature used by the IBM reps. And by the time the
first few LEO IIs were delivered, the technology had moved on again, leaving them looking
increasingly old-fashioned. The second generation of computers would dispense with bulky and
temperamental valves in favour of smaller and more reliable transistors. The first of these second
generation machines, from IBM and the British company EMI, reached their customers in 1960.
They were smaller, faster and usually cheaper. than their valve-based equivalents.
By that time, John Pinkerton had already been working for three years on a wholly original
successor to LEO 2. Using transistors and with the more compact magnetic core store,
LEO 3 would have eight times the storage capacity and ten times the speed of the most advanced LEO
2, while fitting into a much smaller space. Its most advanced feature...
the capacity to run more than one programme at a time. In fact, it could carry out up to 13
parallel operations simultaneously.
The history of computers does not bear out Emerson's dictum that if a man builds a better
mousetrap, the world will make a beaten path to his door. However good your computer,
you have to sell it. Making the first sale of LEO 3 in competition with rivals...
principally IBM and the largest of the British company's ICT, was a nerve-wracking business for
the LEO team, even though they knew they had a better product. The first customer was the tyre
manufacturer, the Dunlop Rubber Company. The chairman of Lyons happened to know the Dunlop
chairman, so it was the old boys' network rather than shrewd marketing that clinched the sale.
In spite of all this, by 1960 the staff of LEO Computers could well believe that they were entering
a golden age. Their numbers had increased to well over 400 in total.
There was work in plenty for the Service Bureau and the new product, LEO 3, was widely admired and
beginning to win orders in the teeth of competition from much larger rivals. Surely the new decade
could bring only further success.
The report that had arrived one morning in the summer of 1962 from the merchant bankers Lazards
made uncomfortable reading for John Simmons. It quoted the view that no company should stay in
computer manufacture unless they were prepared to commit themselves to up to £50 million of capital
expenditure by the end of the decade, and it concluded that it was doubtful that a small computer
business, carried on as a subsidiary of a very different business, could hold its own with the
giants in a rapidly growing market.
Lyons had consulted Lazards in the first place because LEO Computers was losing money. It had made
a net loss of £300,000 since its foundation in 1954. Within months discussions
began with Lazards about a possible merger with one or more other computer businesses. With no
inkling of what was going on in boardrooms and city offices, John Pinkerton had begun working on an
upgrade of the LEO 3 to make an even faster version called the 326. He had an experimental model
running by 1963, and the team of consultants began to win orders, often going head-to-head with
the best that IBM could offer, on the grounds that it was simply superior to anything else
available at the time.
Some of these orders, for the post office for example, were among the largest that had ever been
placed in Europe. At last it seemed possible that LEO Computers was poised to take not only a
technological but also a commercial lead. Yet throughout this period,
its parent company was plotting a strategic withdrawal that would leave LEO without the resources
to exploit its advantage.
English Electric, a large general electrical company whose computer division was based near
Sheffield, quickly emerged as the most likely target and in February 1963 agreed to merge its
computer business with LEO. John Pinkerton and David Caminer knew nothing about the merger,
apart from the odd whisper, until it was signed and sealed and ready to be announced to the press.
The Times called it a sensible piece of rationalisation. But the old LEO hands greeted the news
with disbelief and dismay. Two weeks later, they were informed that the new organisation would be
managed not by Thompson, but by the head of English Electric's computer division.
The same pattern was repeated throughout the company. At each level, a former English electric
manager was put in, either alongside or over the head of the corresponding LEO manager.
Perhaps the final indignity was the name given to the new company, English Electric LEO,
or EEL. No longer a proud beast of the savannah, but a legless,
crawling thing.
A year later, the members of the Lyons Board sold their remaining shares in EEL to English Electric
for slightly less than £2 million. This covered their spending on the whole LEO project,
and one can imagine the smiles of relief around the boardroom table. If any board member expressed
regret, it was not recorded.
EEL became a wholly owned subsidiary of English Electric, who merged it with Marconi to create
EELM, which later became part of International Computers Limited. And in 1998,
ICL was taken over by the Japanese Fujitsu Corporation.
Meanwhile, Lyons closed down its corner houses, its tea shops and hotels, and in 1977 was finally
taken over by Allied breweries. The Cadby Hall food factory was pulled down in 1984.
In the last few years, there have been several front-page stories about computer installations.
In every case, what makes the story is the fact that it has been a spectacular and expensive
failure for a well-known institution. The Passport Office, Cambridge University,
the Ministry of Defence, the Child Support Agency. In every case, a system bought at huge expense
has turned out not to be usable by the people who need it. The LEO experiment.
quixotic as it may have been in the context of a large catering company, is worth remembering for
much more than being the first. It is worth remembering because its architects never forgot what
the computer was for. It was a tool for business, and so it was their responsibility to make sure
it worked for business. Maybe that is the lesson that the designers of modern systems would do well
to remember.
[Presenter] A computer called Leo by Georgina Ferry. The book was abridged by Peter Everett.
The reader was Maggie Tagney, and the producer in Bristol was Francis Burns.

Provenance :
Donated to the LEO Computers Society by Hilary Caminer

Copyright
The video above may be subject to copyright of the original author and is present only on this website for non-profit educational purposes only. If you are the original copyright owner and wish to have the media removed from this website please contact us using the relevant contact email address listed on our contact page.