
 The Electronic Encyclopedia (TM)
 (C) 1988 Grolier Electronic Publishing, Inc.
 
 computer
 
     A computer is an apparatus built to perform routine calculations with 
speed, reliability, and ease. Three important types are: DIGITAL COMPUTERS, 
which function internally and perform operations exclusively with digital 
(discrete) numbers (a type on which most recent progress has centered and which
is the focus of much of this article); ANALOG COMPUTERS, which use continuously
variable parts exclusively for internal representation of magnitudes and to 
accomplish their built-in operations;
and the less-common HYBRID COMPUTERS, which use both continuously variable 
techniques and discrete digital techniques in their operation.
 
     Digital, analog, and hybrid computers are conceptually similar in that 
they all depend on outside instructions; in practice, however, they differ most
noticeably in the means they provide for receiving new programs to do new 
calculating jobs. Digital computers receive new programs quite easily via 
manual instructions or, in modern versions, via automatic means. For analog or 
hybrid computers, however, reprogramming is likely to involve partial 
disassembly and reconnection of mechanisms and components. Because analog 
computers are assemblies of physical apparatuses arranged so as to enact the 
specific type of mathematical relationship for which solutions are to be 
computed, the choice of a new relationship may require a new assembly. To the 
extent that analog machines can be considered programmable, their program is 
rebuilt into their structure for each job.
 
     History of Computers
 
     Historically, the most important early computing instrument is the ABACUS,
which has been known and widely used for more than 2,000 years. It is simply a 
wooden rack holding parallel wires on which beads are strung. When these beads 
are manipulated (moved along the wire) according to "programming" rules that 
the user must memorize, all ordinary arithmetic operations can be performed. 
Another computing instrument, the ASTROLABE, was also in use about 2,000 years 
ago for navigation.
 
     Blaise PASCAL is widely credited with building the first "digital 
calculating machine" in 1642. It performed only additions of numbers entered by
means of dials and was intended to help Pascal's father, who was a tax 
collector. In 1671, Gottfried Wilhelm von LEIBNIZ invented a computer that was 
built in 1694; it could add and, by successive adding and shifting, multiply. 
Leibniz invented a special "stepped gear" mechanism for introducing the addend 
digits, and this mechanism is still in use. The prototypes built by Leibniz and
Pascal were not widely used but remained curiosities until more than a century 
later, when Tomas of Colmar (Charles Xavier Thomas) developed (1820) the first 
commercially successful mechanical calculator that could add, subtract, 
multiply, and divide. A succession of improved "desk-top" mechanical 
calculators by various inventors followed, so that by about 1890 the available 
built-in operations included accumulation of partial results, storage and 
reintroduction of past results, and printing of results, each requiring manual 
initiation. These improvements were made primarily to suit commercial users, 
with little attention given to the needs of science.
 
          Babbage.
 
     While Tomas of Colmar was developing the desk-top calculator a series of 
very remarkable developments in computers was initiated in Cambridge, England, 
by Charles BABBAGE, a mathematics professor. Babbage realized (1812) that many 
long computations, especially those needed to prepare mathematical tables, 
consisted of routine operations that were regularly repeated; from this he 
surmised that it ought to be possible to do these operations automatically. He 
began to design an automatic mechanical calculatin'B	achine, which he called a 
"difference engine," and by 1822 he had built a small working model for 
demonstration. With financial help from the British government, Babbage started
construction of a full-scale difference engine in 1823. It was intended to be 
steam-powered; fully automatic, even to the printing of the resulting tables; 
and commanded by a fixed instruction program.
 
     The difference engine, although of limited flexibility and applicability, 
was conceptually a great advance. Babbage continued work on it for 10 years, 
but in 1833 he lost interest because he had a "better idea"--the construction 
of what today would be described as a general-purpose, fully 
program-controlled, automatic mechanical digital computer. Babbage called his 
machine an "analytical engine"; the characteristics aimed at by this design 
show true prescience, although this could not be fully appreciated until more 
than a century later. The plans for the analytical engine specified a parallel 
decimal computer operating on numbers (words) of 50 decimal digits and provided
with a storage capacity (memory) of 1,000 such numbers. Built-in operations 
were to include everything that a modern general-purpose computer would need, 
even the all-important "conditional control transfer" capability, which would 
allow instructions to be executed in any order, not just in numerical sequence.
The analytical engine was to use PUNCHED CARDS (similar to those used on a 
Jacquard loom), which were to be read into the machine from any of several 
reading stations. The machine was to operate automatically, by steam power, and
would require only one attendant.
 
     Babbage's computers were never completed. Various reasons are advanced for
his failure, most frequently the lack of precision machining techniques at the 
time. Another conjecture is that Babbage was working on the solution of a 
problem that few people in 1840 urgently needed to solve.
 
     After Babbage there was a temporary loss of interest in automatic digital 
computers. Between 1850 and 1900 great advances were made in mathematical 
physics, and it came to be understood that most observable dynamic phenomena 
can be characterized by differential equations, so that ready means for their 
solution and for the solution of other problems of calculus would be helpful. 
Moreover, from a practical standpoint, the availability of steam power caused 
manufacturing, transportation, and commerce to thrive and led to a period of 
great engineering achievement. The designing of railroads and the construction 
of steamships, textile mills, and bridges required differential calculus to 
determine such quantities as centers of gravity, centers of buoyancy, moments 
of inertia, and stress distributions; even the evaluation of the power output 
of a steam engine required practical mathematical integration. A strong need 
thus developed for a machine that could rapidly perform many repetitive 
calculations.
 
          Use of Punched Cards by Hollerith.
 
     A step toward automated computation was the introduction of punched cards,
which were first successfully used in connection with computing in 1890 by 
Herman HOLLERITH and James Powers, working for the U.S. Census Bureau. They 
developed devices that could automatically read the information that had been 
punched into cards, without human intermediation. Reading errors were 
consequently greatly reduced, work flow was increased, and, more important, 
stacks of punched cards could be used as an accessible memory store of almost 
unlimited capacity;
furthermore, different problems could be stored on different batches of cards 
and worked on as needed.
 
     These advantages were noted by commercial interests and soon led to the 
development of improved punch-card business-machine systems by International 
Business Machines (IBM), Remington-Rand, Burroughs, and other corporations. 
These systems used electromechanical devices, in which electrical power 
provided mechanical motion--such as for turning the wheels of an adding 
machine. Such systems soon included features to feed in automatically a 
specified number of cards from a "read-in" station;
perform such operations as addition, multiplication, and sorting; and feed out 
cards punched with results. By modern standards the punched-card machines were 
slow, typically processing from 50 to 250 cards per minute, with each card 
holding up to 80 decimal numbers. At the time, however, punched cards were an 
enormous step forward; they provided a means of input, output, and memory 
storage on a massive scale. For more than 50 years following their first use, 
punched-card machines did the bulk of the world's business computing and a 
considerable portion of the computing work in science.
 
          Automatic Digital Computers.
 
     By the late 1930s punched-card machine techniques had become so well 
established and reliable that Howard Hathaway AIKEN, in collaboration with 
engineers at IBM, undertook construction of a large automatic digital computer 
based on standard IBM electromechanical parts. Aiken's machine, called the 
Harvard Mark I, handled 23-decimal-place numbers (words) and could perform all 
four arithmetic operations; moreover, it had special built-in programs, or 
subroutines, to handle logarithms and trigonometric functions. The Mark I was 
originally controlled from prepunched paper tape without provision for 
reversal, so that automatic "transfer of control" instructions could not be 
programmed. Output was by card punch and electric typewriter. Although the Mark
I used IBM rotating counter wheels as key components in addition to 
electromagnetic relays, the machine was classified as a relay computer. It was 
slow, requiring 3 to 5 seconds for a multiplication, but it was fully automatic
and could complete long computations without human intervention. Mark I was the
first of a series of computers designed and built under Aiken's direction.
 
          Electronic Digital Computers.
 
     The outbreak of World War II produced a desperate need for computing 
capability, especially for the military. New weapons systems were produced for 
which trajectory tables and other essential data were lacking. In 1942, John P.
ECKERT, John W. MAUCHLY, and their associates at the Moore School of Electrical
Engineering of the University of Pennsylvania decided to build a high-speed 
electronic computer to do the job. This machine became known as ENIAC, for 
"Electrical Numerical Integrator And Calculator" (see ENIAC). The size of its 
numerical word was 10 decimal digits, and it could multiply two such numbers at
the rate of 300 products per second, by finding the value of each product from 
a multiplication table stored in its memory. ENIAC was thus about 1,000 times 
faster than the previous generation of relay computers.
 
     ENIAC used 18,000 standard vacuum tubes, occupied 167.3 sq m (1,800 sq ft)
of floor space, and consumed about 180,000 watts of electrical power. It had 
punched-card input and output and arithmetically had 1 multiplier, 1 
divider-square rooter, and 20 adders employing decimal "ring counters," which 
served as adders and also as quick-access (0.0002 seconds) read-write register 
storage. The executable instructions composing a pd-I5:i!embodied in the 
separate units of ENIAC, which were plugged together to form a route through 
the machine for the flow of computations. These c/|gecuyons had to be redone 
for each different problem, together with presetting function tables and 
switches. This "wire-your-own" instruction technique was inconvenient, and only
with some license could ENIAC be considered programmable; it was, however, 
efficient in handling the particular programs for which it had been designed. 
ENIAC is generally acknowledged to be the first successful high-speed 
electronic digital computer (EDC) and was productively used from 1946 to 1955. 
A controversy developed in 1971, however, over the patentability of ENIAC's 
basic digital concepts, the claim being made that another U.S. physicist, John 
V. ATANASOFF, had already used the same ideas in a simpler vacuum-tube device 
he built in the 1930s while at Iowa State College. In 1973 the court found in 
favor of the company using the Atanasoff claim.
 
          The Modern "Stored Program" EDC.
 
     Intrigued by the success of ENIAC, the mathematician John VON NEUMANN 
undertook (1945) a theoretical study of computation that demonstrated that a 
computer could have a very simple, fixed physical structure and yet be able to 
execute any kind of computation effectively by means of proper programmed 
control without the need for any changes in hardware. Von Neumann contributed a
new understanding of how practical fast computers should be organized and 
built; these ideas, often referred to as the stored-program technique, became 
fundamental for future generations of high-speed digital computers and were 
universally adopted.
 
     The stored-program technique involves many features of computer design and
function besides the one named; in combination, these features make 
very-high-speed operation feasible. Details cannot be given here, but a glimpse
may be provided by considering what 1,000 arithmetic operations per second 
implies. If each instruction in a job program were used only once in 
consecutive order, no human programmer could generate enough instructions to 
keep the computer busy. Arrangements must be made, therefore, for parts of the 
job program called subroutines to be used repeatedly in a manner that depends 
on how the computation progresses. Also, it would clearly be helpful if 
instructions could be altered as needed during a computation to make them 
behave differently. Von Neumann met these two needs by providing a special type
of machine instruction called conditional control transfer--which permitted the
program sequence to be interrupted and reinitiated at any point--and by storing
all instruction programs together with data in the same memory unit, so that, 
when desired, instructions could be arithmetically modified in the same way as 
data.
 
     As a result of these techniques and several others, computing and 
programming became faster, more flexible, and more efficient, with the 
instructions in subroutines performing far more computational work. Frequently 
used subroutines did not have to be reprogrammed for each new problem but could
be kept intact in "libraries" and read into memory when needed. Thus, much of a
given program could be assembled from the subroutine library. The all-purpose 
computer memory became the assembly place in which parts of a long computation 
were stored, worked on piecewm3Y and assembled to form the final results. The 
computer control served as an errand runner for the overall process. As soon as
the advantages of these techniques became clear, the techniques became standard
practice.
 
     The first generation of modern programmed electronic computers to :)ke 
advantage of these improvements appeared in 1947. This group included computers
using random access memory (RAM), which is a memory designed to give almost 
constant access to any particular piece of information. These machines had 
punched-card or punched-tape input and output devices and RAMs of 1,000-word 
capacity with an access time of 0.5 Greek mu sec (0.5 X 10 to the power of 
minus 6 sec); some of them could perform multiplications in 2 to 4 Greek mu 
sec. Physically, they were much more compact than ENIAC: some were about the 
size of a grand piano and required 2,500 small electron tubes, far fewer than 
required by the earlier machines. The first-generation stored-program computers
required considerable maintenance, attained perhaps 70% to 80% reliable 
operation, and were used for 8 to 12 years. Typically, they were programmed 
directly in machine language, although by the mid-1950s progress had been made 
in several aspects of advanced programming. This group of machines included 
EDVAC and UNIVAC (see UNIVAC), the first commercially available computers.
 
          Advances in the 1950s.
 
     Early in the 1950s two important engineering discoveries changed the image
of the electronic-computer field, from one of fast but often unreliable 
hardware to an image of relatively high reliability and even greater 
capability. These discoveries were the magnetic-core memory and the 
transistor-circuit element (see COMPUTER MEMORY).
 
     These new technical discoveries rapidly found their way into new models of
digital computers; RAM capacities increased from 8,000 to 64,000 words in 
commercially available machines by the early 1960s, with access times of 2 or 3
msec. These machines were very expensive to purchase or to rent and were 
especially expensive to operate because of the cost of expanding programming. 
Such computers were typically found in large computer centers--operated by 
industry, government, and private laboratories--staffed with many programmers 
and support personnel. This situation led to modes of operation enabling the 
sharing of the high capability available; one such mode is batch processing, in
which problems are prepared and then held ready for computation on a relatively
inexpensive storage medium, such as magnetic drums, magnetic-disk packs, or 
magnetic tapes. When the computer finishes with a problem, it typically "dumps"
the whole problem--program and results--on one of these peripheral storage 
units and takes in a new problem. Another mode of use for fast, powerful 
machines is called time-sharing. In time-sharing the computer processes many 
waiting jobs in such rapid succession that each job progresses as quickly as if
the other jobs did not exist, thus keeping each customer satisfied. Such 
operating modes require elaborate "executive" programs to attend to the 
administration of the various tasks.
 
          Advances in the 1960s.
 
     In the 1960s efforts to design and develop the fastest possible computers 
with the greatest capacity reached a turning point with the completion of the 
LARC machine for Livermore Radiation Laboratories of the University of 
California by the Sperry-Rand Corporation, and the Stretch computer by IBM. The
LARC had a core memory of 98,000 words and multiplied in 10 Greek mu sec. 
Stretch was provided with several ranks of memory having slower access for the 
ranks of greater capacity, the fastest access time being less than 1 Greek mu 
sec and the total capacity in the vicinity of 100 million words.
 
     During this period the major computer manufacturers began to offer a range
of computer capabilities and costs, as well as various peripheral 
equipment--such input means as consoles and card feeders; such output means as 
page printers, cathode-ray-tube displays, and graphing devices;
and optional magnetic-tape and magnetic-disk file storage. These found wide use
in business for such applications as accounting, payroll, inventory control, 
ordering supplies, and billing. CENTRAL PROCESSING UNITS (CPUs) for such 
purposes did not need to be very fast arithmetically and were primarily used to
access large amounts of records on file, keeping these up to date. By far the 
greatest number of computer systems were delivered for the more modest 
applications, such as in hospitals for keeping track of patient records, 
medications, and treatments given. They are also used in automated library 
systems, such as MEDLARS, the National Medical Library retrieval system, and in
the Chemical Abstracts system, where computer records now on file cover nearly 
all known chemical compounds.
 
          Recent Advances.
 
     The trend during the 1970s was, to some extent, away from extremely 
powerful, centralized computational centers and toward a broader range of 
applications for less-costly computer systems. Most continuous-process 
manufacturing, such as petroleum refining and electrical-power distribution 
systems, now use computers of relatively modest capability for controlling and 
regulating their activities. In the 1960s the programming of applications 
problems was an obstacle to the self-sufficiency of moderate-sized on-site 
computer installations, but great advances in applications programming 
languages (see SOFTWARE, COMPUTER) are removing these obstacles. Applications 
languages are now available for controlling a great range of manufacturing 
processes, for computer operation of machine tools, and for many other tasks.
 
     Moreover, a new revolution in computer hardware is well under way, 
involving miniaturization of computer-logic circuitry and of component 
manufacture by what are called large-scale integration, or LSI, techniques. In 
the 1950s it was realized that "scaling down" the size of electronic digital 
computer circuits and parts would increase speed and efficiency and thereby 
improve performance--if only manufacturing methods were available to do this. 
About 1960 photoprinting of conductive circuit boards to eliminate wiring 
became highly developed. Then it became possible to build resistors and 
capacitors into the circuitry by photographic means (see PRINTED CIRCUIT). In 
the 1970s vacuum deposition of transistors became common, and entire 
assemblies, such as adders, shifting registers, and counters, became available 
on tiny "chips." In the 1980s very large scale integration (VLSI), in which 
hundreds of thousands of transistors are placed on a single chip, is becoming 
increasingly common. Many companies, some new to the computer field, introduced
in the 1970s programmable MINICOMPUTERS supplied with software packages. The 
size-reduction trend continued with the introduction of personal computers (see
COMPUTER, PERSONAL), which are programmable machines small enough and 
inexpensive enough to be purchased and used by individuals. MICROPROCESSORS 
equipped with ROM, or read-only memory (which stores constantly used, 
unchanging programs), perform many process-control, testing, monitoring, and 
diagnosing functions, such as automobile-engine diagnosis and production-line 
inspection tasks.
 
     Progress in the area of software, however, has not matched the great 
strides in hardware. Software has become the major cost of many systems because
programming productivity has not increased very rapidly. New languages, such as
ADA, are being constructed to help alleviate this problem. Despite difficulties
with software, however, the cost per calculation of computers is rapidly 
decreasing, and their convenience and efficiency is expected to increase in the
foreseeable future.
 
     Julian Bigelow
