Notes for Martin Campbell-Kelly and William Aspray Computer: A History of the Information Machine
Key concepts: conditional branch, human/computer interaction, stored-program computer, technology, technology intercept strategy, time-sharing, virtual memory.
Related theorists: Babbage, Edmund C. Berkley, Burks, Vannevar Bush, Perry Crawford, De Prony, Gee, Goldstine, Grace Murray Hopper, Kemeny, John Mauchly, von Neumann.
PREFACE TO THE SLOAN TECHNOLOGY SERIES
Books of this genre, serious attempts at narrating minimally biased history of evolution of state of the art best practices, form the foundation of critical programming and philosophy of computing studies; follow them with insider perspective of software management and software architect informed by substantial professional experience, including Brooks and Lammers.
(vii) Technology
is the application of science, engineering, and industrial
organization to create a human-built world.
(vii) The aim of the
series is to convey both the technical and human dimensions of the
subject: the invention and effort entailed in devising the
technologies and the comforts and stresses they have introduced into
contemporary life.
Acknowledgments
(ix)
This book has its origins in the vision of the Alfred P. Sloan
Foundation that it is important for the public to understand the
technology that has so profoundly reshaped Western society during the
past century.
Introduction
(2)
Today, research scientists and atomic weapons designers still use
computers extensively, but the vast majority of computers are
employed for other purposes, such as word processing and keeping
business records. How did this come to pass? To answer this question,
we must take a broader view of the history of the computer as the
history of the information machine.
(3) The old technologies had
three shortcomings: They were too slow in doing their calculations,
they required human intervention in the course of a computation, and
many of the most advanced calculating systems were special-purpose
rather than general-purpose devices.
(3) The basic function
specifications of the computer were set out in a government report
written in 1945, and these specifications are still largely followed
today. . . . One is the improvement in components, leading to faster
processing speed, larger information storage capacity, improved price
performance, better reliability, less required maintenance, and the
like.
(3-4) The second type of innovation was the mode of
operation . . . high-level programming languages, real-time
computing, time-sharing, networking, and graphically oriented
human-computer interfaces.
(4) We have organized the book in four
sections. The first covers the way computing was handled before the
arrival of electronic computers. The next two sections describe the
mainframe computer era, roughly from 1945 to 1980, with one section
devoted to the computer's creation and the other to its evolution.
The final section discusses the origins of the new computing
environment of the personal computer and the Internet.
Predicts deeper understanding of computers than their broad definition of information machines through emergence of synthetic historical scholarship epitomized by texts and technology studies.
(6) Our work falls in the present generation of scholarship based on the broader definition of the information machine, with strong business and other contextual factors considered in addition to technical factors. We anticipate that within the next decade, a new body of historical scholarship will appear that will enable someone to write a new synthetic account that will deepen our understanding of computers in relation to consumers, gender, labor, and other social and cultural issues.
PART ONE
Before the Computer
1 When Computers Were People
(9)
The electronic computer can be said to combine the roles of the human
computer and the human clerk.
(9-10) However, logarithmic and
trigonometric tables were merely the best-known general-purpose
tables. By the late eighteenth century, specialized tables were being
produced for several different occupations: navigational tables for
mariners, star tables for astronomers, life insurance tables for
actuaries, civil engineering tables for architects, and so on. All
these tables were produced by human computers, without any mechanical
aid.
(10) When Astronomer Royal Maskelyne died in 1811—Hitchins
had died two years previously—the Nautical
Almanac “fell
on evil days for about 20 years, and even became notorious for its
errors.”
CHARLES
BABBAGE AND TABLE-MAKING
(10-11)
During this period Charles Babbage became interested in the problem
of table-making and the elimination of errors in tables. . . .
Realizing that Cambridge (and England) had become a mathematical
backwater compared to continental Europe, Babbage and two fellow
students organized the Analytical Society, which succeeded in making
major reforms of mathematics in Cambridge and eventually the whole of
England.
(11) In 1819 Babbage made the first of several visits to
Paris, where he made the acquaintance of a number of leading members
of the French scientific academy, such as the mathematicians
Pierre-Simon Laplace and Joseph Fourier, with whom he formed lasting
friendships. It was probably during this visit that Babbage learned
of the great French table-making project organized by Baron Gaspard
de Prony. This project would show Babbage a vision that would
determine the future course of his life.
(11) It was by far the
largest table-making project the world have ever known, and de Prony
decided to organize it much as one would organize a factory.
(12)
De
Prony organized
his table-making “factory” into three sections. The first section
consisted of half a dozen eminent mathematicians—including Adrien
Legendre and Lazare Carnot—who decided on the mathematical formulas
to be used in the calculations. Beneath them was another small
section—a kind of middle management—that, given the mathematical
formulas to be used, organized the computations and compiled the
results ready for printing. Finally, the third and largest section,
which consisted of sixty to eighty human computers, did the actual
computation. The computers used the “methods of differences,”
which required only the two basic operations of addition and
subtraction, and not the more demanding operations of multiplication
and division. Hence the computers were not, and did not need to be,
educated beyond basic numeracy and literacy. In fact, most of them
were hairdressers who had lost their jobs because “one of the most
hated symbols of the ancient regime was the hairstyles of the
aristocracy.”
(12) Although the Bureau was producing
mathematical tables, the operation was not itself mathematical. It
was fundamentally the application of an organizational technology,
probably for the first time outside a manufacturing or military
context, to the production of information.
(13) The laborers in
Adam Smith's imaginary pin-making factory would soon be replaced by a
pin-making machine. Babbage decided that rather than emulate de
Prony's labor-intensive and expensive manual table-making
organizations, he would ride the wave of the emerging mass-production
technology and invent a machine for making tables.
(13)
Conceptually the Difference Engine was very simple: It consisted of a
set of adding mechanisms to do the calculations and a printing
part.
(14) Unfortunately, the engineering was more complicated
than the conceptualization.
(14) Babbage was now battling on two
fronts: first, designing the Difference Engine; and second,
developing the technology to build it.
(15) By 1833, Babbage had
produced a beautifully engineered prototype Difference Engine that
was too small for real table-making and lacked a printing unit, but
showed beyond any question the feasibility of his concept.
(15)
Raising the specter of the Analytical Engine was the most spectacular
political misjudgment of Babbage's career; it fatally undermined the
government's confidence in his project, and he never obtained another
penny.
CLEARING
HOUSES
(16)
The Bankers' Clearing House was an organization that processed the
rapidly increasing number of checks being used in commerce.
(18)
Babbage clearly recognized the significance of the Bankers' Clearing
House as an example of the “division of mental labor,” comparable
with de Prony's table-making project and his own Difference
Engine.
(18) Alongside this physical and highly visible transport
infrastructure grew a parallel, unseen information infrastructure
known as the Railway Clearing House, which was modeled very closely
on the Bankers' Clearing House.
(18) Another important example was
the Central Telegraph Office founded in 1859 to overcome the problem
that it was not economical to have telegraph lines connecting every
town in the land with every other.
Packet switching resembles telegraphy system of sorters, pigeon holes, and messengers; does Hayles discuss in How We Think?
(18-19)
Row upon row of telegraphists communicated messages with all parts of
the nation and abroad. Messenger boys constantly scuttled through the
rows of telegraphists collecting telegrams as they arrived,
delivering them to a team of women sorters. The sorters
placed the telegrams in pigeon holes—one
for each of the hundreds of destination towns. More messenger boys
emptied the pigeon holes and delivered the telegrams to the
telegraphists for onward transmission.
(19) By the 1860s another
class of data-processing bureaucracy was beginning to develop in
association with the “thrift movement.” . . . It was this
relative prosperity combined with the thrift movement that created
the market for savings banks and industrial insurance companies in
the second half of the nineteenth century.
(20) Yet here was an
extremely sophisticated organizational technology that can still be
seen underlying the structure of any modern corporation.
HERMAN
HOLLERITH AND THE 1890 CENSUS
(20)
The population census was established by an Act of Congress in 1790
to determine the “apportionment” of members of the House of
Representatives.
(21) Over 21,000 pages of census reports were
produced for the 1880 census, which took some seven years to process.
This unreasonably long time provided a strong motive to speed up the
census by mechanization or any other means that could be
devised.
(22) While Hollerith was not a deep thinker like the
polymath Babbage, he was practical where Babbage was not. Hollerith
also had a strong entrepreneurial flair, so that he was able to
exploit his inventions and establish a major industry.
(22)
Hollerith's key idea was to record the census return for each
individual as a pattern of holes on punched paper tape or a set of
punched cards, similar to the way music was recorded on a string of
punched cards on fairground organettes of the period. It would then
be possible to use a machine to automatically count the holes and
produce the tabulations.
(24) A punching clerk—doing what was
optimistically described as “vastly interesting” work—could
punch an average of seven hundred cards in a six and a half hour day.
Female labor was heavily used for the first time in the census, which
a male journalist noted “augurs well for its conscientious
performance” because “women show a moral sense of responsibility
that is still beyond the average.” Over 62 million cards were
punched, one for each citizen.
Hollerith census machines used tabulator and sorter for punched cards.
(24-25) Each census machine consisted of two parts: a tabulating
machine, which could count the holes in a batch of cards, and the
sorting box, into which cards were placed by the operator ready for
the next tabulating operation. . . . When the press was forced down
on the card, a pin meeting the solid material was pushed back into
the press and had no effect. But a pin encountering a hole passed
straight through, dipped into a mercury cup, and completed an
electrical circuit. This circuit would then be used to add unity to
one of forty counters on the front of the census machine. The circuit
could also cause the lid of one of the twenty-four compartments of
the sorting box to fly open—into which the operator would place the
card so that it would be ready for the next phase of the
tabulation.
(25) As each card was read, the census machine gave a
ring of a bell to indicate that it had been correctly sensed.
Hollerith machines sabotaged by workers to provide a break.
(26) The trouble was usually that somebody had extracted the mercury
from one of the little cups with an eye-dropper and squirted it into
a spittoon, just to get some un-needed rest.
(26) The Census
Bureau used the Hollerith system not only to reduce the cost of the
census, but also to improve the quality and quantity of information,
and the speed with which it was produced.
AMERICA'S
LOVE AFFAIR WITH OFFICE MACHINERY
(26)
In the closing decades of the nineteenth century, office equipment,
in both its most advanced and its least sophisticated forms, was
almost entirely an American phenomenon.
(27) The fact is that
America was gadget-happy and was caught by the glamor of the
mechanical office. . . . This attitude was reinforced by the rhetoric
of the office systems movement.
(27-28) Systematizers set about
restructuring the office—introducing typewriters and adding
machines, designing multipart business forms and loose-leaf filing
systems, replacing old-fashioned accounting ledgers with machine
billing systems, and so on. The office systematizer was the ancestor
of today's information-technology consultant.
(28) There is thus
an unbroken line of descent from the giant office-machine firms of
the 1890s to the computer makers of today.
2 The Mechanical Office
(29)
To understand the development of the computer industry, and how this
apparently new industry was shaped by the past, one must understand
the rise of the office machine giants in the years around the turn of
the century. This understanding is necessary, above all, to
appreciate how IBM's managerial style, sales ethos, and technologies
combined to make it perfectly adapted to shape and then dominate the
computer industry.
(29) Today, we use computers in the office for
three main tasks. There is document preparation . . . information
storage . . . financial analysis and accounting.
(29) These were
precisely the three key office activities that the business machine
companies of the late nineteenth century were established to serve.
THE TYPEWRITER
(30)
Hence the major attraction of the typewriter was that typewritten
documents could be read effortlessly at several times the speed of
handwritten ones.
(33) Selling for an average price of $75,
typewriters became the most widely used business machines, accounting
for half of all office appliance sales.
(34) By the 1920s office
work was seen as predominantly female, and typing as universally
female.
(34) But now we can see that it is important to the
history of computing in that it pioneered three key features of the
office machine industry and the computer industry that succeeded it:
(1) the perfection of the product and low-cost manufacture; (2) a
sales organization to sell the product; and (3) a training
organization to enable workers to use the technology.
THE RANDS
(35)
Vertical filing systems used only a tenth of the space of the
box-and-drawer filing systems they replaced, and made it much faster
to access records.
THE FIRST ADDING MACHINES
(36)
The first commercially produced addign machine was the Arithmometer
developed by Thomas de Colmar of Alsace at the early date of
1820.
(37) The bookkeeper of the 1880s was trained to add up a
column of four-figure amounts mentally almost without error, and
could do so far more quickly than using an Arithmometer.
(38)
William S. Burroughs made the first successful attack on the second
critical problem of the office adding machine: the need to print its
results.
Social need for adding machines with progressive, withholding tax law, as Social Security Act would require punched-card machinery.
(39) A major
impetus to development of the U.S. adding machine industry occurred
in 1913 with the introduction of a new tax law that adopted
progressive tax rates and the withholding of tax from pay.
(39)
Between the two world wars Burroughs moved beyond adding machines and
introduced full-scale accounting machines.
THE NATIONAL CASH REGISTER
COMPANY
(41) [John H.]
Patterson understood that the cash register needed constant
improvement to keep it technically ahead of the competition. In 1888
he established a small “inventions department” for this purpose.
This was probably the first formal research and development
organization to be established in the office machine industry, and it
was copied by IBM and others.
(42) NCR's greatest legacy to the
computer age, however, was the way it shaped the marketing of
business machines and established virtually all of the key sales
practices of the industry.
THOMAS WATSON AND THE FOUNDING OF
IBM
(49) The
“rent-and-refill” nature of the punched-card business made IBM
virtually recession-proof. Because the punched-card machines were
rented and not sold, even if IBM acquired no new customers in a bad
year, its existing customers would continue to rent the machines they
already had, ensuring a steady year-after-year income.
(49) The
second source of IBM's financial stability was its punched-card
sales.
(50) Technical innovation was the third factor that kept
IBM at the forefront of the office machine industry between the two
world wars.
(51) Under the Social Security Act of 1935, it became
necessary for the federal government to maintain the employment
records of the entire working population of 26 million people. IBM,
with its overloaded inventories and with factories in full
production, was superbly placed to benefit.
3 Babbage's Dream Comes True
(54)
The idea of the Analytical Engine came to Babbage when he was
considering how to eliminate human intervention in the Difference
Engine by feeding back the results of a computation, which he
referred to as the engine “eating its own tail.” Babbage took
this simple refinement of the Difference Engine, and from it evolved
the design of the Analytical Engine, which embodies almost all the
important functions of the modern digital computer.
(55) In the
Analytical Engine, numbers would be brought from the store to the
arithmetic mill for processing and the results of the computation
would be returned to the store.
Is the dark age of mechanical digital computing attributed to failure of Babbage the historical fictional pivot for Steam Punk?
(59-60) As it was, Babbage's failure undoubtedly contributed to what L. J. Comrie called the “dark age” of digital computing. Rather than follow where Babbage had failed, scientists and engineers preferred to take a different, nondigital path—a path that involved the building of models, but which we now call analog computing.
THE TIDE PREDICTOR AND OTHER ANALOG
COMPUTING MACHINES
(60-61)
The most important analog computing technology of the nineteenth
century, from an economic perspective, was the mechanical tide
predictor. . . . In 1876 the British scientist Lord Kelvin invented a
useful tide-predicting machine.
(61) This was the overriding
problem with analog computing technologies—a specific problem
required a specific machine.
(61) A whole new generation of analog
computing machines were developed in the 1920s to help design the
rapidly expanding U.S. electrical power system.
(63) [Vannevar]
Bush's differential
analyzer, built between 1928 and 1931, was not a general-purpose
computer in the modern sense, but it addressed succh a wide range of
problems in science and engineering that it was far and away the
single most important computing machine developed between the wars.
A WEATHER-FORECAST FACTORY
(65)
At the end of the war in 1918, [Lewis Fry] Richardson returned to his
work at the Meteorological Office and completed his book, which was
published in 1922. In it he described one of the most extraordinary
visions in the history of computing: a global weather-forecast
factory.
SCIENTIFIC COMPUTING SERVICE
(67)
[Leslie John] Comrie's great insight was to realize that one did not
need special-purpose machines such as differential analyzers; he
thought computing was primarily a question of organization. For most
calculations, he found that his “calculating girls” equipped with
ordinary commercial calculating machines did the job perfectly
well.
(68) Perhaps Comrie's major achievement at the Nautical
Almanac Office was to bring punch card machines—and therefore,
indirectly, IBM—into the world of numerical computing.
THE HARVARD MARK I
(72)
The following month, December 1937, [Howard Hathaway] Aiken polished
up his proposal, in which he referred to some of Babbage's
concepts—in particular, the idea of using Jacquard cards to program
the machine—and sent it to IBM. This proposal is the earliest
surviving document about the Mark I, and it must remain something of
an open question as to how much Aiken derived from Charles Babbage.
In terms of design and technology, the answer is clearly very little;
but in terms of a sense of destiny, Aiken probably derived a great
deal more.
(73) Unfortunately, the calculator was incapable of
making what we now call a “conditional branch”--that
is, changing the progress of a program according to the results of an
earlier computation. This made complex programs physically very long.
. . . If Aiken had studied Babbage's—and especially
Lovelace's—writings more closely, he would have discovered that
Babbage had already arrived at the concept of a conditional branch.
First book on digital computing by Aiken on Mark I, although Mauchly memorandum credited as real starting point, and von Neumann the first on the stored-program computer.
(74)
After the dedication and the press coverage, there was intense
interest in the Harvard Mark I from scientific workers and engineers
wanting to use it. This prompted Aiken and his staff to produce a
500-page Manual of
Operation of the Automatic Sequence Controlled Calculator,
which was effectively the first book on digital computing ever
published.
(76) Babbage could never have envisioned that one day
electronic machines would come onto the scene with speeds thousands
of
times faster than he ever dreamed. This happened within two years of
the Harvard Mark I being completed.
PART TWO
Creating the Computer
4 Inventing the Computer
THE MOORE SCHOOL
THE ATANASOFF-BERRY COMPUTER
ECKERT AND MAUCHLY
Must appreciate how beliefs about likely processing speeds influence speculation on design and uses of computer technologies.
(86) By August 1942 [John] Mauchly's ideas on electronic computing had sufficiently crystallized that he wrote a memorandum on The Use of High Speed Vacuum Tubes for Calculating. In it he proposed an “electronic computer” that would be able to perform calculations in 100 seconds that would take a mechanical differential analyzer 15 to 30 minutes, and that would have taken a human computer “at least several hours.” This memorandum was the real starting point for the electronic computer project.
ENIAC AND EDVAC: THE STORED-PROGRAM CONCEPT
(88)
All together, in addition to its 18,000 tubes, the ENIAC would
include 70,000 resistors, 10,000 capacitors, 6,000 switches, and
1,500 relays.
(88) The ENIAC was Eckert and Mauchly's project.
Eckert, with his brittle and occasionally irascible personality, was
“the consummate engineer,” while Mauchly, always quiet, academic,
and laid back, was “the visionary.”
(91) In short, there were
three major shortcomings of the ENIAC: too little storage, too many
tubes, too lengthy reprogramming.
(91-92) Von Neumann was
captivated by the logical and mathematical issues in computer design
and became a consultant to the ENIAC group to try to help them
resolve the machine's deficiencies and develop a new design. This new
design would become known as the “stored-program computer,” on
which virtually all computers up to the present day have been based.
All this happened in a very short space of time.
(92) From this
point on, while construction of the ENIAC continued, all the
important intellectual activity of the computer group revolved around
the design of ENIAC's successor: the EDVAC, for Electronic Discrete
Variable Automatic Computer.
(92-93) Goldstine has likened the
stored-program concept to the invention of the wheel: It was
simple—once he had thought of it. This simple idea would allow for
rapid program set-up by enabling a program to be read into the
electronic memory from punched cards or paper tape in a few seconds;
it would be possible to deliver instructions to the control circuits
at electronic speeds; it would provide two orders of magnitude more
number storage; and it would reduce the tube content by 80 percent.
But most significantly, it would enable a program to treat its own
instructions as data. Initially this was done to solve a technical
problem associated with handling arrays of numbers, but later it
would be used to enable programs to create other programs—laying
the seeds for programming languages and artificial intelligence.
Invitation to study contested history of development of electronic computer as Hayles does cybernetics.
(93) During these meetings, Eckert
and Mauchly's contributions focused largely on the delay-line
research, while von Neumann, Goldstine, and Burks concentrated on the
mathematical-logical structure of the machine. Thus there opened a
schism in the group between the technologists (Eckert and Mauchly) on
the one side and the logicians (von Neumann, Goldstine, and Burks) on
the other, which would lead to serious disagreements later on.
(93)
Von Neumann designated these five units as the central control, the
central arithmetic part, the memory, and the input and output
organs.
(93) Another key decision was to use binary to represent
numbers.
(94) By the spring of 1945, the plans for EDVAC had
evolved sufficiently that von Neumann decided to write them up. His
report, entitled A First
Draft of a Report on the EDVAC,
dated 30 June 1945, was the seminal document describing the
stored-program computer. . . . Von Neumann's sole authorship of the
report seemed unimportant at the time, but it later led to him being
given sole credit for the invention of the modern computer.
Compare urge to disseminate bootstrapping knowledge to create computers as will of technological unconscious to later cycle in which money making trumped openness that made corporations and wealthy individuals rivaling the largest established human and machine ensembles; today an intellectual deoptimization settled on path of least resistance default philosophies of computing dominate.
(95) Although originally intended for internal circulation to the Project PY group, the EDVAC Report rapidly grew famous, and copies found their way into the hands of computer builders around the world. This was to constitute publication in a legal sense, and it eliminated any possibility of getting a patent. For von Neumann and Goldstine, who wished to see the idea move into the public domain as rapidly as possible, this was a good thing; but for Eckert and Mauchly, who saw the computer as an entrepreneurial business opportunity, it was a blow that would eventually cause the group to break up.
THE ENGINEERS VERSUS THE LOGICIANS
THE MOORE SCHOOL LECTURES
(98)
Apart from its gargantuan size, the feature the media found most
newsworthy was its awesome ability to perform 5,000 operations in a
single second.
(98) The Moore School Lectures, as the course later
became known, took place over eight weeks from 8 July to 31 August
1946. The list of lecturers on the course read like a Who's
Who of
computing of the day. It included luminaries such as Howard Aiken and
von Neumann, who made guest appearances, as well as Eckert, Mauchly,
Goldstine, Burks, and several others from the Moore School who gave
the bread-and-butter lectures.
MAURICE
WILKES AND EDSAC
(100)
[F. C.] Williams decided that the key problem in developing a
computer was the memory technology. . . . With a single assistant to
help him, he developed a simple memory system based around a
commercially available cathode-ray tube. With guidance from Newman,
who “took us by the hand” and explained the principles of the
stored-program computer, Williams and his assistant built a tiny
computer to test out the new memory system. . . . The date was
Monday, 21 June 1948, and the “Manchester Baby Machine”
established incontrovertibly the feasibility of the stored-program
computer.
(102) From the outset, he [Maurice Wilkes] decided that
he was interested in having
a
computer, rather than trying to advance computer engineering
technology. He wanted the laboratory staff to become experts in using
computers—in programming and mathematical applications—rather
than in building them.
(104) On 6 May 1949, a thin ribbon of paper
containing the program was loaded into the computer; half a minute
later the teleprinter sprang to life and began to print 1, 4, 9, 16,
25 . . . The world's first practical stored-program computer had come
to life, and with it the dawn of the computer age.
5 The Computer Becomes a Business
Machine
(105) What really
happened in the 1950s, as suggested by the title of this chapter, is
that the computer was reconstructed—mainly by computer
manufacturers and business users—to be an electronic
data-processing machine rather than a mathematical instrument.
(106)
There were actually three types of firms that entered the computer
industry: electronics and control equipment manufacturers, office
machine companies, and entrepreneurial start-ups.
“MORE THAN OPTIMISTIC”: UNIVAC
AND BINAC
(109) The most
ambitious feature of the UNIVAC was the use of magnetic-tape storage
to replace the millions of punched cards used in the Census Bureau
and other businesses.
(110) The Prudential's computer expert was a
man named Edmund C. Berkley,
who was to write the first semipopular book on computers, Giant
Brains,
published in 1949.
(111) The first American stored-program
computer to operate, the BINAC was never a reliable machine.
IBM:
EVOLUTION, NOT REVOLUTION
(112)
IBM had several electronics and computer development projects in the
laboratories in the late 1940s; what delayed them being turned into
products was the uncertainty of the market.
(113) IBM
institutionalized its attitude to electronics and computers through
the slogan “evolution not revolution.” By this, it meant that it
would incorporate electronics into existing products to make them
faster, but they would not otherwise be any different.
(115)
Computer historians have often failed to realize the importance of
the [IBM Card Programmed Calculator] CPC, not least because it was
called a “calculator” instead of a “computer.” Watson
insisted on this terminology because he was concerned that the latter
term, which had always referred to a human being, would raise the
specter of technological unemployment.
UNIVAC COMES TO LIFE
Hopper foremost female computer professional and promoter of advanced programming techniques.
(121) The programming team was initially led by Mauchly, but was later run by a programmer recruited from the Harvard Computation Laboratory, Grace Murray Hopper, who would become the driving force behind advanced programming techniques for commercial computers and the world's foremost female computer professional.
Election night role of mock up UNIVAC predicting outcome for Eisenhower was key introduction of computers to general public.
(123) The appearance of the UNIVAC on election night was a pivotal moment in computer history. Before that date, while some people had heard about computers, very few had actually seen one; after it, the general public had been introduced to computers and had seen at least a mock-up of one. And that computer was called a UNIVAC, not an IBM.
IBM'S BIG PUSH
(125)
However, this failed to take account of the fact that computers were
hot news, and business magazines were buzzing with stories about
electronic brains for industry and commerce. Cost effectiveness was
no longer the only reason, or even the most important reason, for a
business to buy a computer.
(126) Instead of the mercury delay
lines chosen for the UNIVAC, IBM decided to license the Williams Tube
technology developed for the Manchester University computer in
England.
(126) Another advantage over the UNIVAC was that IBM's
computers were modular in construction—that is, they consisted of a
series of “boxes” that could be linked together on
site.
(126-127) Although it was still a laboratory prototype at
this time, IBM mounted a crash research program to develop core
memory into a reliable product.
(127) Yet it was not the
large-scale 700 series that secured IBM's leadership of the industry,
but the low-cost Magnetic Drum Computer.
Mastery of marketing and long term planning by IBM through dissemination of their computers in higher education to produce the next generation of workers trained on them; good example of social factor influencing history more so than the technological capabilities of the devices, a topic developed with respect to real time processing.
(127) With an
astute understanding of marketing, IBM placed many 650s in
universities and colleges, offering machines with up to a 60 percent
discount provided courses were established in computing. The effect
was to create a generation of programmers and computer scientists
nurtured on IBM 650s, and a trained workforce for IBM's products. It
was a good example of IBM's mastery of marketing, which was in many
ways more important than mastery of technology.
(128) In the
memorable phrase of the computer pundit Herb Grosch, in losing its
early lead to IBM, Remington Rand “snatched defeat from the jaws of
victory.”
THE COMPUTER RACE
(129)
Thus, by the end of the 1950s the major players in the computer
industry consisted of IBM and a handful of also-rans: Sperry Rand,
Burroughs, NCR, RCA, Honeywell, GE, and CDC. Soon journalists would
call them IBM and the seven dwarves.
6
The Maturing of the Mainframe:
The Rise and Fall of IBM
THE
BREAKTHROUGH MODEL 1401
(131)
The turning point for IBM was the announcement of the model 1401
computer in October 1959. The 1401 was not so much a computer as a
computer system.
(133)
IBM also had to find a technological solution to the programming
problem. The essential challenge was how to get punched card-oriented
business analysts to be able to write programs without a huge
investment in retraining them and without companies having to hire a
new temperamental breed of programmer. The solution that IBM offered
was a new programming system called Report Program Generator
(RPG).
(133) IBM developed entire program suites for the
industries that it served most extensively, such as insurance,
banking, retailing, and manufacturing. These application programs
were very expensive to develop, but because IBM had such a dominant
position in the market it could afford to “give” the software
away, recouping the development costs over tens of hundreds of
customers.
(134) How did IBM get its forecast so wrong? The 1401
was certainly an excellent computer, but the reasons for its success
had very little to do with the fact that it was a computer. Instead,
the decisive factor was the new type 1403 “chain” printer that
IBM supplied with the system.
IBM AND THE SEEN DWARVES
REVOLUTION,
NOT EVOLUTION: SYSTEM/360
(137-138)
The biggest problem, however, was not in hardware but in software.
Because the number of software products IBM offered to its customers
was constantly increasing, the proliferation of computer models
created a nasty gearing effect. Given m
different
computer models, each requiring n
different
software products, a total of m
x
n
programs
had to be developed and supported. . . . For IBM, a compatible range
promised to be the electronic Esperanto that would contain the
software problem.
(139) The decision to produce a compatible
family was not so clear-cut as it appears in hindsight, and there was
a great deal of agonizing at IBM.
(140) The New Product Line was
one of the largest civilian R&D projects ever undertaken.
(141)
Simply keeping the machine designs compatible between the
geographically separate design groups was a major problem.
(141)
All told, direct research costs were around $500 million. But ten
times as much again was needed for development—to tool up the
factories, retrain marketing staff, and re-equip field
engineers.
(142) However, all internal debate about the
announcement strategy effectively ceased in December 1963 when
Honeywell announced its model 200 computer. . . . Honeywell 200s
could run IBM programs without reprogramming and, using a
provocatively named “liberator” program, could speed up existing
1401 programs to make full use of the Honeywell 200's power.
(143)
The computer industry and computer users were stunned by the scale of
the announcement. While an announcement from IBM had long been
expected, its tight security had been extremely effective, so that
outsiders were taken aback by the decision to replace the entire
product line.
(144) System/360 has been called “the computer
that IBM made, that made IBM.” The company did not know it at the
time, but System/360 was to be its engine of growth for the next
thirty years.
THE
DWARVES FIGHT BACK
(144-145)
But in technological terms, System/360 was no more than competent. .
. . For example, the proprietary electronics technology that IBM had
chosen to use, known as Solid Logic Technology (SLT), was halfway
between the discrete transistors used in second-generation computers
and the true integrated circuits of later machines.
Lack of time-sharing support in System/360 major design flaw.
(145) Perhaps the most serious design flaw in System/360 was its
failure to support time-sharing—the fastest-growing market for
computers—which enabled a machine to be sued simultaneously by many
users.
(146) RCA was the only mainframe company to go
IBM-compatible in a big way.
(146) Hence, the second and less
risky strategy to compete with System/360 was product
differentiation—that is, to develop a range of computers that were
software-compatible with one another but not compatible with
System/360. This was what Honeywell did.
(146) The third strategy
to compete with IBM was to aim for niche markets that were not well
satisfied by IBM and in which the supplier had some particular
competitive advantage.
THE FUTURE SERIES
(148)
In June 1970 the evolutionary successor to the 360 range was
announced as System/370. The new series offered improved price
performance through updated electronic technology: True integrated
circuits were used in place of the SLT modules of System/360, and
magnetic-core storage was repalced with semiconductor memory. The
architecture was also enhanced to support more effectively
time-sharing and communications-based on-line computing. The
technique of “virtual memory” was
also introduced, using a combination of software and hardware to
increase the amount of working memory in the computer (and therefore
permit the execution of much larger programs).
(148) While the
life of the 360-370 line was being extended by these modest
improvements, IBM's main R&D resources were focused on the much
more challenging Future Series.
(149) FS was indeed a remarkably
advanced architecture, and this entailed its own problems. Even today
there is no major computer range that has the functionality of FS as
it was proposed in the early 1970s.
(149) But more than anything,
it was the growing importance of the existing software investment of
IBM and its users that made a new architecture less feasible by the
day and sealed the fate of FS.
Suggest the contradiction of the personal computer and Internet age is the knowledge gap between problem solving by use and by programming resulting from hysteresis of a generation of users acculturated to closed source, interface level competencies.
(150) The maturing of the mainframe has produced one of the great technological contradictions of the twentieth century: As semiconductor components improve in size and speed by a factor of two every year or two, they are used to build fossilized computer designs and run software that is groaning wit age. The world's mainframes and the software that runs on them have become like the aging sewers beneath British city streets, laid in the Victorian era. Not a very exciting infrastructure, and largely unseen—but they work, and life would be totally different without them.
THE DECLINE OF THE IBM EMPIRE
(151)
IBM's current malaise has frequently been attributed to the rise of
the personal computer, but the facts do not support this contention.
The scene was set for IBM's downfall in the mid-1970s, with the
commodification of the mainframe—that is, by the mainframe becoming
an article that could be produced by any competent manufacturer. This
had the inevitable result of reducing the generous profit margins IBM
had enjoyed for sixty years.
(153) Only IBM guaranteed a complete
solution to business problems, and an IBM salesman was all too likely
to remind a data-processing manager that no one ever got fired by
hiring from IBM. This was a patronizing attitude that came close to
condescension, and often resulted in a love-hate relationship between
IBM and its customers.
PART THREE
Innovation and
Expansion
7 Real Time: Reaping the
Whirlwind
(157)
No previous office technology had been able to achieve this speed,
and the emergence of real-time computers had the potential for
transforming business practice.
JAY
FORRESTER AND PROJECT WHIRLWIND
(159)
[Perry] Crawford was perhaps the first to appreciate real-time
digital control systems, long predating the development of practical
digital computers. As early as 1942 he had submitted a master's
thesis on Automatic
Control by Arithmetic Operations,
which discussed the use of digital techniques for automatic
control.
(161) As did everyone leading an early computer project,
Forrester soon discovered that the single most difficult problem was
the creation of a reliable storage technology.
(163) Crawford
foresaw a day when real-time computing would control not just
military systems but whole sectors of the civilian economy—such as
air traffic control.
THE
SAGE DEFENSE SYSTEM
(167)
The value to the nation of the core memory spin-off could by itself
be said to have justified the cost of the entire Whirlwind project.
Distributed control with SAGE Direction Centers; subindustry grew to develop and implement its basic technologies.
(167) The SAGE system that was eventually constructed consisted of a
network of twenty-three Direction Centers distributed throughout the
country.
(168) The real contribution of SAGE was thus not to
military defense, but through technological spin-off to civilian
computing. An entire subindustry was created as industrial
contractors and manufacturers were brought in to develop the basic
technologies and implement the hardware, software, and
communications.
SABRE: A REVOLUTION IN AIRLINE RESERVATIONS
Technology intercept strategy joined system planning and expected advances; compare to military planning of scheduled breakthroughs discussed by Edwards.
(172-173)
Simultaneously an idealized, integrated, computer-based reservation
system was planned, based on likely future developments in
technology. This later become known as a “technology
intercept” strategy.
The system would be economically feasible only when reliable
solid-state computers with core memories became available. Another
key requirement was a large random-access disk storage unit, which at
that time was a laboratory development rather than a marketable
product.
(175) The airline reservations problem was unusual in
that it was largely unautomated in the 1950s and so there was a
strong motivation for adopting the new real-time technology.
Longer-established and more traditional businesses were much slower
to respond.
THE
UNIVERSAL PRODUCT CODE
(177)
Even greater benefits would come if manufacturers, wholesalers, and
retailers all used the same, universal product code. This would allow
the whole of the food industry to harmonize their computer hardware
and software development, in effect becoming one integrated
operation. This implied an extraordinary degree of cooperation among
all these organizations, and the implementation of the UPC was thus
as much a political achievement as a technical one.
(178) One of
the first was to agree upon a product code of ten digits—five
digits representing the manufacturer and five representing the
product.
(179) The bar code principle has diffused very rapidly to
cover not just food items but most packaged retail goods (note the
ISBN bar ode on the back cover of this book). The bar code has become
an icon of great economic significance.
(180) In a sense, there
are two systems co-existing, one physical and one virtual. The
virtual information system in the computers is a representation of
the status of every object in the physical manufacturing and
distribution environment—right down to an individual can of peas.
8 Software
New type of rhetoric required for computer programming.
(181) The fundamental difficulty of writing software was that, until computers arrived, human beings had never before had to prepare detailed instructions for an automaton—a machine that obeyed unerringly the commands given to it, and for which every possible outcome had to be anticipated by the programmer.
PROGRAMMING LANGUAGES: FORTRAN AND COBOL
Fascinating to find Hopper simultaneously crossing philosophy of computing and feminist discourse where she is explicitly grouped among those not claiming to be a feminist.
(187)
Probably no one did more to change the conservative culture of 1950s
programmers than Grace Hopper, fist programmer for the Harvard Mark I
in 1943, then with UNIVAC. For several years in the 1950s she
barnstormed around the country, proselytizing the virtues of
automatic programming at a time when the technology delivered a good
deal less than it promised.
(191) COBOL was particularly
influenced by what one critic described as Hopper's “missionary
zeal for the cause of English language coding.”
SOFTWARE ENGINEERING
(200)
Misconception of early programmers as long-haired men (male humans) when many were women, noting Grace Hopper evangelizing learning programming languages and four who started a business.
(201) The Garminsch conference began a major cultural shift in the perception of programming. Software writing started to make the transition from being a craft for a long-haired programming priesthood to becoming a real engineering discipline.
9
New Modes of Computing
(207)
Improved machines and software enabled more sophisticated
applications and many primitive batch-processing systems became real
time, but data processing still consisted of computing delivered to
naïve users by an elite group of systems analysts and software
developers.
(207) The personal computer grew out of an entirely
different culture of computing, which is the subject of this chapter.
This other mode of computing is associated with computer
time-sharing, the BASIC programming language, Unix, minicomputers,
and new microelectronic devices.
PART FOUR
Getting Personal
10
Shaping of the Personal
Computer
(233) Perhaps its
most serious distortion is to focus on a handful of individuals,
portrayed as visionaries who clearly saw the future and made it
happen: Apple Computer's Steve Jobs and Microsoft's Bill Gates figure
prominently in this genre. By contrast, IBM and the established
computer firms are usually portrayed as dinosaurs: slow-moving,
dim-witted, deservedly extinct. When it comes to be written, the
history of the personal computer will be much more complex than this.
It will be seen to be the result of a rich interplay of cultural
forces and commercial interests.
VISICALC
Importance of early computer games producing a new generation of programmers who developed understanding of HCI (Gee).
(250) Computer games are often overlooked in discussions of the personal-computer software industry, but they played an important role in its early development. Programming computer games created a corps of young programmers who were very sensitive to what we now call human/computer interaction. The most successful games were ones that needed no manuals and gave instant feedback.
11
The Shift to Software
(260)
Entrants to the new industry needed not advanced software-engineering
knowledge but the same kind of savvy as the first software
contractors in the 1950s: creative flair and the technical knowledge
of a bright undergraduate. The existing software companies simply
could not think or act small enough: their overhead costs did not
allow them to be cost-competitive with software products for personal
computers.
GRAPHICAL USER INTERFACE
GUI user-friendliness was the next step in broadening computer use following Kemeny vision of BASIC programming on time-sharing systems.
(264) For the personal computer to become more widely accepted and reach a broader market, it had to become more “user-friendly.” During the 1980s user-friendliness was achieved for one-tenth of computer users by using a Macintosh computer; the other nine-tenths could achieve it through Microsoft Windows software. Underlying both systems was the concept of the graphical user interface.
STEVE JOBS AND THE MACINTOSH
Product differentiation marketing strategy by Apple may have contributed to shift toward postmodern preferences Turkle articulates.
(274) John Sculley recognized that Apple had a classic business problem that called for a classic solution: product differentiation.
12
From the World Brain to the World Wide Web
To Campbelly-Kelley the history of the modern computer as information machine concludes with commerce, recreation, and socializing seeming to have replaced the initial excitement of access to knowledge that the Internet offered.
(284) However, the use of the Internet that has created the most excitement is its potential for giving ordinary people access to the world's store of knowledge through the “World Wide Web.”
Campbell-Kelly, Martin and William Aspray. Computer: A History of the Information Machine. New York: Basic Books, 1996. Print.