Skip to main content

Full text of "A History of the Personal Computer"

See other formats


The Historical Background. 

1/2 Part 1 The Historical Backeroiind 

Figure 1.1: A griq»hical history of early computer technology (1937-1959). 

Chapter 1 Development of the Computer 

computer technology. It is a cursory overview of the 
early significant developments. It starts with the 
beginning of relay technology in the late 1930s and 
concludes with the integrated circuit technology of 
1959. The review of prior technology sets the stage for 
a discussion of personal computing and the 
microcomputer. It also forms the basis for a historic 
appreciation of the power and capabilities of today's 
personal computer, as compared to those early digital 

1.1 ... Original Digital Computers 

Relay Technology 

The original computers used mainly relay 
technology. Examples of these are those built by Howard 
Aiken in the USR, Bletchley Park in England, Konrad Zuse 
in Germany and those at Bell Telephone Laboratories in 
the USA. 

relay computers starting with the Complex Number 
Calculator. George Stibitz and S. B. Williams designed 
the calculator using 450 telephone relays and 10 
crossbar switches. The machine became operational in 
January 1940 and Stibitz demonstrated it from a remote 
location, Dartmouth College in Hanover, New Hampshire in 
September 1940. Following this Bell Laboratories 
developed the Models III, IV, V and VI computers. Bell 
completed the Model V, also known as the Bell 
Laboratories General Purpose Relay Calculator in 1946. 
It contained over 9,000 relays. An addition or 
subtraction required 0.3 seconds, a multiplication 0.8 

1/4 Part I The Historical Background 

Konrad Zuse designed a completely mechanical 
computer, the 21 in 1938. Zuse then developed a hybrid 
relay/ mechanical unit named the 22. R successful 

the German Aeronautical Research Institute for a more 
advance 23 machine. 2u3e completed the 23 computer in 
December 1941 using 2,300 relays. After the Second World 
War Zuse formed his own company that produced relay, 

Bletchley Park in England built a number of 

second World War for cryptoanalysis security work. M. H. 
A. Newman and C. E. Wynn-Williams completed a hybrid 
relay-vacuum tube machine called the Heath-Robinson in 
April 1943. The British Secrets Act has restricted the 
availability of technical information for the computing 
machines developed at Bletchley Park. 

Howard H. Aiken and IBM developed the Automatic 
Sequence Control Calculator (AECC) . IBM built the 
computer at its Endicott laboratory during the period of 
1939-43. It became operational in January 1943 and IBM 
presented the computer to Harvard University in August 
1944. It had 2,200 counter wheels, 3,300 relays, was 51 
feet long by 8 feet high and had a weight of 
approximately 5 tons. Addition or subtraction took about 
0.3 seconds, multiplication 6 seconds and division could 
require 16 seconds. Harvard renamed it the Mark I and 
during the next decade Aiken developed the Mark II, III 

Mechanical relay technology had inherent spes 
limitations as applied to computation. This led to tl 
use of vacuum tube technology that provided significai 
speed improvements . 

Vacuum Tube Technology (USA) 

John V. Atanasoff developed the first electron: 

College during the period of 1938-42. Atanaso" 
confirmed the design concepts but did not use ti 

computer in a practical sense. He developed the desic 

Development of the Computer 1/5 

computer the RBC (Rtanasoff -Berry Compute r) in 

Clifford E. Berry. 

The first electronic computer used for 

calculations and solving practical problems, was ENIRC 
(Electronic Numeric Integrator and Calculator) . J. 
Prespec Eckert and John W. Mauchly developed the 
computer at the Moore School of Electrical Engineering, 
University of Pennsylvania during the period of 1943-46. 
ENIAC was a huge machine, having over 18,000 vacuum 
tubes and 1, 500 relays that consumed 17 4 kilowatts of 
power. Rn addition or subtraction calculation took 200 
microseconds, a multiplication 2,800 microseconds and a 
division 24,000 microseconds. The clock rate was 60-125 
kHz. The U.S. Rrmy Ordnance Department funded the 
computer development for use in calculating ballistic 

Other early computers developed and built during 
the late 1940s and early 1950s using vacuum tube 
technology were EDVAC, BINRC, UNIVRC, Whirlwind and IAS. 
During the early development of the digital computer 
there was a close liaison between research institutions 
in the USA and the United Kingdom. This resulted in 
significant research and development of vacuum tube 
computers in England. 

VacHutu. Tube Technology (England) 

Between 1946 and 1948 the University of 
Manchester built an electronic computer to test the 
concept of electrostatic memory storage. F. C. Williams 
developed this memory concept using the cathode ray 

The first general purpose electronic computer in 
the United Kingdom was the Electronic Delay Storage 

Automatic Calculator (EDSAC) . Maurice V. Milkes 
developed EDSAC at Cambridge University between 1947 and 
1949. EDSAC was the first computer to use the stored 
program concept. 

Ferranti Ltd . , did additional developmental work 
on the University of Manchester computer for commercial 

1/6 Part I The Historical Background 

production. They completed the 

affected computer development in England and the USA. 
Bell Telephone Laboratories developed the transistor in 
1947/48. Utilization of solid state technology resulted 
in significant cost savings and reliability 

Solid State Technology 

England. T. Kilburn of the University of Manchester 
designed the experimental computer that was operational 
in November 1953. In the early 1950s, Bell Telephone 
Laboratories received s^n Air Force contract to build a 
special computer called TRRDIC (TEAnsistor Digital 
Computer) . Bell built TRADIC and had it operating in 
early 1954. 

Most early computer development had occurred in 
the USA and England at university research institutions. 
However a major USA corporation providing tabulating 
equipment worldwide entered the computer field. That 
company was IBM and they became the dominant supplier of 

1.2 ... IBM — International Business Machines 

Thomas J. Watson 3r., left National Cash Register 
Company and became General Manager of the Computing- 
Tabulating-Recording Company in May 1914 . The company 
name changed to International Business Machines 
Corporation (IBM) in 1924. Tabulating machines were the 
basis for the initial growth of the corporation. These 

and sorting machines to electric accounting machines. 
These machines met the diverse needs of business, 
industry and some fields of science into the 1940s. 

The first entry of IBM into the field of 
computers was the collaborative effort of the company 

Development of the Computer 1/7 

with Howard Aiken of Harvard University. IBM approved 
construction of the Automatic Sequence Control 
Calculator (ASCC) in 1939 and completed it as described 
in Section 1.1 in January 1943. 

in electronic technology. Utilization of this new 
technology resulted in the development of the 603/4 
Electronic Multiplier in September 1946. IBM initiated 
construction of the "Super Calculator," the SSEC 
(Selective Sequence Electronic Calculator) in 1945 and 
dedicated it in January 1948, It had 21,400 relays, 
12,500 vacuum tubes and operated until January 1952. 

With the start of the Korean war in June 1950, 
IBM initiated steps to assist in the war effort with the 
development of the Defense Calculator. This new computer 

of germai 

production 701 sy 
important project was the joint development of the SAGE 
(Semi-Automatic Ground Environment) computer with MIT 
between 1952 and 1956. The company announced their last 
vacuum tube computer system, the 705 Model III, in 
September 1957. In October 1957 the company issued a 
memo stating a policy to use solid-state technology on 
all new computer developments. 

IBM demonstrated a transistorized version of the 
Type 604 calculator in October 1954 and the first 
transistorized product was the 608 calculator that 
shipped in December 1957. These two machines did not 
have a stored program. Consequently IBM called them 
calculators, not computers. IBM initiated a project in 
late 1955 for a supercomputer utilizing transistors and 
the latest technology that became known as the Stretch 
system. Then in October 1958 they announced the 7090 
Data Processing System using the Stretch technology. The 
first delivery of a 7090 system was in November 1959. 


. Some 



ie wers 

; bi 



, e 






ic tape 



and ut: 






ime k 

nown as 




ng Mach 



the it 






ed. Tl- 

.e t 



ipped in 



r 1952. 


■ ther 

rt I The Historical Background 

Computer development at IBM and other 
institutions was very dependent on technology advances. 
Of particular importance were the advances in memory 
capabilities and solid state technology. 

l.S ... Technology 


The following is relative to the computer 
internal memory and not auxiliary storage units such as 
magnetic drums, disks or tape drives. Memory was a 
crucial technology in the expanding use of the computer. 

The first electronic digital computer developed 
by Atanasoff during the period of 1538-42 used 

had two drums and each drum had 32 bands of 5 

capacity of 1,60 bits, or in more familiar terms 200 

The ENinc computer used vacuum tubes for memory 
storage. It used the decimal system and had 20 
accumulators for storage of variables. Each accumulator 
could hold a 10 place number. Each place number was a 10 
stage ring counter corresponding to the digits 1 to 5. 
Therefore it required 100 unique flip-flop vacuum tubes 
(plus 2 for sign) in each accumulator to store a 10 
place number. By taking certain liberties we could say 
the system had 100 unique bits in each of the 20 
accumulators, for a total of 2,000 bits or 250 bytes in 

tubes used for memory. A typical delay line used 

filled tube and could store 1, 000 bits. Eckert and 
Mauchly applied the concept on the EDVAC computer during 
the period of 1945-51. Other early computers such as 

EDSAC, BINAC for Northrop Aircraft, Inc., and the 
initial UNIVAG computers also used the acoustic delay 
line concept 

The Princeton University, Institute for Advance 

Development of the Computer 1/9 

utilizing the cathode cay tube for the IAS computer in 
1945. However F. C. Williams developed the first 
functional implementation of electrostatic memory on the 
University of Manchester test computer between 1946-48. 

the middle of 194 8 Williams was able to demonstrate a 
unit with a capacity of several thousand bits. 
Subsequently it was operational on the IRS computer and 
the IBM 701 System in 1952. 

Jay W. Forrestor evaluated magnetic core memory 
at the Massachusetts Institute of Technology (MIT) 
between 1949 and 1951. Ferrite cores were operating at 
MIT and on the IBM 405 Recounting Machine in 1952. The 
first computer application of ferrite cores was on the 
Whirlwind computer at MIT in 1953. The preceding 
computer evolved into the SRGE (Semi Automatic Ground 
Environment) computers for a USA national air defense 
system. MIT selected IBM to build the computers. This 
defense program added significant impetus to the 
development of ferrite core technology. 

Each of these stages of memory development 

also enabled the use of larger and more complex 
software. Associated developments in secondary memory 
storage complemented the advances in internal memory. 

Seeonibuy Memory (Storage) 

The first electronic computers used paper cards 
for storage. Rtanasoff's computer used 8 . 5-by-l 1-inch 
cards and ENIAC used IBM cards. 

R. D. Booth of the United Kingdom was an early 
developer of various forms of memory storage. Booth 
experimented with thermal, rotating disk-pin and 
magnetic drum memories after the Second World War . By 
May of 1948 he had installed and demonstrated a working 
magnetic drum memory in the Automatic Relay Computer 
(ARC) . In the USA a company called Engineering Research 
Associates Inc . , (ERA) presented papers describing 

1/10 Parti The Historical Background 

conceived by Jacob Rabinow at the National Bureau of 
Standards (tJBS) in 1952. In early 1952, IBM established 
an advanced research development laboratory in San Jose, 
California that began looking at magnetic disk storage 
as an inexpensive fast data retrieval system. The 
requirements evolved from advance development of card- 
related applications and a United States Air Force 

requesting a change from batch to random access in the 
method of updating file systems. Extensive research was 
conducted on disk materials and coatings, various types 
of movable heads and electronic control systems for data 

IBM demonstrated the first magnetic disk drive 
assembly with movable read-write-heads in May 1955 and 
described the concept at the Western Joint Computer 
Conference in February 1956. The assembly consisted of a 
stack of fifty coated aluminum disks, 2 4 inches in 
diameter, rotating at 1,200 revolutions per minute. Each 
disk surface contained 100 concentric recording tracks 
that provided a total storage capacity of 5 million 
characters (the equivalent of 50,000 IBM cards) . The 
system was subsequently announced as the IBM 350 Disk 
Storage system. The preceding and a scheme for automatic 
addressing of data, was incorporated into the IBM 305 
RAMAC {Random Access Method of Accounting and Control) 
system announced in September 1956. IBM announced a 14 

5mory techn. 

ology wei 



the transi; 

=tor and 

.de signifii 

;ant imp] 

Development of the Computer 1/11 

Transistor and Integrated Circuit 

William B. Shockley was the director of a team of 
researchers that developed the transistor at Bell 
Telephone Laboratories in New Jersey, USA. The first 
transistor was operational in December 1947. John 

on a single semiconductor chip was first developed by 
Jack St. Clair Kilby at Texas Instruments, Inc., in 
1958. Kilby constructed the integrated circuit in 
germanium. Each component on the integrated circuit 
required precise and laborious interconnection by hand. 

Independently in 1959, Robert H. Noyce at 
Fairchild Semiconductor Corporation also developed a 
monolithic concept of integrating components on a single 
silicon chip. However, Noyce had also developed a 
practical method to interconnect the components. This 
new technology deposited an insulating layer on the 

by a photo resist technique. Then by depositing a 
vaporized metal layer, interconnection of the components 
was achieved in a practical manner. Noyce's concept had 
been facilitated by the invention of the planar process 
by Jean Hoerni, a technique to make a flat, or planar 
surface for transistors. The founding of Fairchild 
Semiconductor in 19 57, had been financed by Fairchild 
Camera and Instrument Corporation, who completed its 
ownership of the new company in 1959. 

interconnecting the components and the planar process 
were just as important as the integrated circuit 
concept . This would be crucial in the ensuing patent 
litigation between the two inventors and respective 
companies. In resolving the coverage of the patent 
applications between the two companies, an agreement 
resulted in Jack Kilby and Robert Noyce being declared 

Technology improvements and memory advances in 
particular resulted in increased utilization of the 

1/12 Parti The Historical Background 

computer. However, it is software that facilitated the 
practical application of the computer. 

1.4 ... Software 

Development of the stored program concept had a 
direct relationship to advances in memory technology. 
The size, speed of access, reliability and cost of 
memory were crucial factors in the evolution of 

The BegmtUng 

The initial computers did not have a stored 
program, or what became known later as software. The 
first practical electronic computer ENIRC, used forty 
plug-boards, that required a configuration of wire 
connections for each instruction. Later changes utilized 
three panels, each containing 1,200 ten-position 

John von Neumann described the concept of storing 

first computer to function with a 
stored program was EDSRC at Cambridge University, 
England in 1949. 

Grace M. Hopper was an early pioneer in the 
development of initial programming languages. Her 
initial work started on the Mark 1 at Harvard University 
in 1544. Then at the Eckert-Mauchly Computer 
Corporation. The first compiler was R-0 and ran on the 
UNIVRC computer in 1952. This compiler formed the basis 
for other variations such as ARITH-MftTIC, MflTH-MATIC and 
FLOW-MftTIC (1955-58) which assisted in the later 
development of COBOL. 

One of the earliest applications of computer 
graphics was on the MIT Whirlwind computer, that was 
operational by 1953. This was an interactive system that 

applied to the SRGE defense system built by IBM. 

Development of the Computer 1/13 


In 1954 IBM established a project directed by 
John W. Backus to develop a compiler for the Model 704 

language FORTRftN (FORmula TRRNslation) that IBM finished 
in April 1957. The language had a notation oriented to 
mathematicians and scientists. Improvements in the 

language followed with the subsequent release of 
versions II, III and IV. 


By 1957 a number of people had become concerned 
that a common programming language designed for 

government also stated a concern about the proliferation 
of different compilers from the various computer 
manufacturers. In 1559 the Defense department initiated 

formed called CODRSYL (Committee On DAta SYstem 
Languages). The committee coordinated the development of 
a new language that became COBOL (common Business 
Oriented Language) . CODASYL formed other committees, to 
work on the definition and further development of the 
language. These developments resulted in the 1960 
release of COBOL. 


Artificial intelligence researchers started using 
computers and developed game software such as chess and 
checkers between the late 1940s and the 1950s. There 
were many researchers in this field of artificial 
intelligence. Early chess playing programs were 
developed by Claude E. Shannon in 1949 and Alex 
Bernstein in 1957. Arthur L. Samuel of IBM developed 

William Higinbotham and associate Dave Potter 
developed the first video game at Brookhaven National 
Laboratory in Upton, New York in 1958. Called Merlin, it 
simulated a game of tennis using an analog computer, 
oscilloscope display and paddle-type controllers for 

1/14 Parti The Historical Background 

u1:ilization of computers. As the number of users 
increased, new periodicals were published and new 
associations formed to disseminate information and 
facilitate user interaction. 

1.5... Otiter Developments 


One of the first groups to be formed in the 
computer industry was the Rssociation for Computing 
Machinery (RCM) , founded during 1947/48. The Data 
Processing Management Association (DPMft) formed in 1945, 
is one of the largest groups with involvement in 
education and certification of data processing 
professionals. IBM customers were one of the first to 
form user groups. The 3HRRE group that formed in 1955 
had an initial orientation to scientific users. Rn IBM 
commercial users group called GUIDE (Guidance of Users 
of Integrated Data-processing Equipment) formed in 19 56. 
Then the International Federation for Information 
Processing (IFIP) formed in 1959. 


One of the earliest periodicals was the Digital 
Computer Newsletter started by the Office of Naval 
Research in 1949. Another early publication was the IBM 
Technical Newsletter series started in 1550. The ACM 
issued a quarterly Journal in 1954 that evolved into the 
Communications of the ACM in 1558. 

Datamation started in October 1957 as Research and 
Engineering (The Magazine of Datamation) . It started as 

publication in 1967. 


Vannevar Bush published a futuristic article 
entitled "As Me May Think" in the July 1945 issue of 
Atlantic Monthly [12, pp. 47-55] . The article described 
a "future device for individual use" called "Memex" in 

Development of the Computer 1/15 


ications . . . that 

may be consulted with exceeding 


and flexibility. " 

■ Another essential feature of 


was "associative i 

ndexing ... whereby any item may 

be c 

aused at will 

to select immediately and 


tically another." 

Bush's Memex was similar to 


concepts of an 

interactive personal computer 


ting a field of kn. 

owledge with hypertext links. 

Most of the act: 

.vities in computer development 

and u; 

;e related to lar 

ge mainframe computers. However 


s organizations we 

■re attempting to lower the cost 

of computeirs by introduc 

ing small computer systems. 

1.6 ... Small Computer Sy^ems 


MIT built the Memory Test Computer (MTC) between 
19 52 and 19 5 3 to test magnetic core memory planes for 
the Whirlwind 1 computer. Harlan Anderson and Kenneth 
Olsen designed MTC and would later co-found Digital 
Equipment Corporation (DEC) . The computer was a 16-bit 
unit built from standard Whirlwind plug-in circuit 
package forms. 

Following the completion of the MTC computer 
Olsen and Wesley A. Clark proposed the construction of a 
large transistorized computer, the TX-1 . Rejected by 
management, Clark designed a smaller 18-bit unit, the 
TX-0 in 1957. Then MIT developed the ARC (Average 
Response Computer) and a much larger computer, the TX-2 
in 1958 . One of the students using these computers was 
Charles E. Molnar. Molnar and Clark would later develop 
the first personal computer, the MIT LINC. 


small Personal Automatic Calculator (PAC) project, 
starting in the late 1940s. He then described details of 
an engineering model of the PAC project in December 
1954. This project evolved into the IBM 610 Auto-Point 
Computer announced in September 1957. The system 
consisted of three units. A floor-standing cabinet that 

1/16 Parti The Historical Background 

punches, plugboard and magnetic drum. The other two 
units were an operator keyboard for control and data 
entry and an electric typewriter for printed output. It 
was one of the last two vacuum tube computer models 
built by IBM and the company considers this to be the 
first "Personal Computer." IBM built about 180 units at 
a purchase price of 555,000. 


minicomputer market was Kenneth H. Olsen. After 
graduating from MIT he worked at the Lincoln 
Laboratories, then on the Whirlwind and SftGE computer 
systems between 1950 and 1957. Olsen co-founded Digital 
Equipment Corporation (DEC) with Harlan E. Anderson in 
August 1957, to provide low cost logic modules and 
computers for engineers and scientists. 

The first DEC minicomputer was the 18-bit PDP-1 
(Programmed Data Processor - One). DSC demonstrated the 
PDP-1 prototype at the Eastern Joint Computer Conference 
in December 1959. It had a cathode-ray tube display, 
keyboard and was the first small commercial interactive 
general-purpose computer . 


The general use of large commercial and 
scientific computers was firmly established by the end 
of the 1950s. However the major computer manufacturers 
did not commit sufficient resources to the development 
of small competitive computer systems. This resulted in 
other companies releasing small computers such as the 
Bendix Aviation Corporation G-15, the Librascope/General 
Precision LGP-30 in 1956 and the DEC PDP-1 announcement 
in 1959. 

These small systems were lowering the cost of 
computer technology. Although they targeted at 
scientific users, they were a part of the evolution to 
personal computing and the first personal computer in 
the 1960s.