Computer

a device or set of devices for the mechanization and automation of data processing (computations).

Modern computers are subdivided into three classes according to the method of representing data: (1) analog computers, in which data are represented in the form of continuously changing variables expressed by physical quantities (the angle of rotation of a shaft, the intensity of an electrical current, voltage, and so on); (2) digital computers, in which data are represented in the form of discrete values of variables (numbers) expressed by a combination of discrete values of some physical quantity; and (3) hybrid computers, various units of which represent data by one or the other method.

Historically, digital computing devices appeared first—for example, abacuses and their numerous precursors. In the 17th century the French scientist B. Pascal, and later the German mathematician G. W. von Leibniz, built the first digital computers. The first computer suitable for practical use was Thomas de Colmar’s adding machine (1820). V. T. Odhner’s adding machine, which became very widespread, was built in 1874. In the early 20th century tabulators appeared for performing various statistical, bookkeeping, and financial-banking operations.

The idea of building a general-purpose digital computer belongs to Professor C. Babbage of Cambridge University. In 1833 he designed a computer whose features are close to those of modern devices. The plan was ahead of the needs of the day and technical possibilities for realization.

The development of the theory of relay-switching circuits and experience in using telephone and punched card equipment made it possible during the 1930’s to undertake the development of a computer with program control. At first electromagnetic relays were used. The first such machine, the Mark I, was built in 1944 in the United States. The first electronic digital computer, the ENIAC (Electronic Numerical Integrator and Computer) was built in 1946, also in the United States.

In the Soviet Union, the MESM (Small Electronic Calculator) electronic digital computer was developed in 1950 at the Academy of Sciences of the Ukrainian SSR under the direction of Academician S. A. Lebedev. The MESM marked the beginning of work in the field of mathematical electronic machine building in the USSR. In subsequent years a number of digital computers that differed in productivity and technical concept were built in the USSR to satisfy the needs of the national economy (for example, the BESM, Strela, M-20, M-220, Minsk, Ural, and Mir).

The first continuous-operation devices appeared in the 16th and 17th centuries. These include the slide rule and nomograms for navigational calculations. In the mid-19th century, very simple mechanical integrators appeared. Work on analog computers developed significantly at the turn of the 20th century. Machines for solving differential equations, electromechanical integrating machines, and others were developed. In the USSR, the beginning of analog computer development dates back to 1927; it is associated with the work of S. A. Gershgorin, M. V. Kirpichev, I. S. Bruk, V. S. Luk’ianov, and others. During the 1950’s and 1960’s, several types of analog computers were built, many of which have found wide application.

The development of electronic computers is closely linked to achievements in the field of electronics. The first electronic computers used vacuum tubes; it is customary to call these computers first-generation machines. The development of semiconductor radio electronics made possible the transition to designing second- and third-generation computers. They are characterized by a more complex logic diagram and by software, which is a programmed extension of the hard-ware. The technology for manufacturing second-generation computers differed little from that used for the manufacture of first-generation computers; vacuum tubes were replaced by semiconductor triodes (transistors) and diodes. Third-generation computers are built with integrated circuits, which contain dozens of transistors, resistors, and diodes in a single module. The transition to producing computers with integrated circuits demanded an almost complete revision of the technology of production of electronic computers.

The theory of mathematical modeling is the basis for constructing analog computers. Using analogies among phenomena that differ in their physical nature, the analog computer models the processes being calculated. A large part of the analog computer equipment consists of linear and non-linear decision elements. In electronic analog computers these are operational DC amplifiers (integrator, amplifier, inverter), units of coefficients, typical nonlinearities, delays, and so on. To solve a specific problem, the units of an analog computer are interconnected in the necessary combinations. Output data of the analog computer are obtained from readings on the displays at the terminal points of the circuit. The analog computer is characterized by high speed of operation, simplicity in the linkage with the object being investigated, the possibility of easily changing the parameters of the problem under investigation both during its preparation and during the solution process, comparatively low precision, and a limited class of solvable problems.

Solving a problem on digital computers involves the sequential performance of arithmetic operations on numbers that correspond to quantities representing the initial data. The numbers are usually represented in the form of an aggre-gate of mechanical, pneumatic, or electrical pulses and are fixed by elements that can each assume several stable states that correspond strictly to one digit of the number. Before solution on the digital computer, the problem is broken down into a series of simple sequential operations and their order is set—that is, a program of computations is drawn up.

Digital computers are subdivided into three classes according to their method of control: those with manual control, those with a rigid program; and general-purpose types. Digital computers with manual control include keyboard computers, adding machines, and lever-operated computers. Modern desk digital computers are manufactured almost entirely with electronic components. The computing process is controlled manually, which results in low computation speed. The digital computer with manual control is a means of mechanizing calculation operations and is only suitable for solving very simple problems with a limited number of computations.

Rigid-program digital computers include tabulators, specialized machines oriented to solving a narrow range of problems (for example, on-board computers), and so on. In these computers the computing process is controlled automatically by a program set up on a switchboard or permanently built into the machine. The digital computer with a switchable program is a means for partial automation of the computing process and is rapidly being replaced by general-purpose digital computers. Computers with built-in programs are used in cases where simplicity, reliability, low cost, and small size and weight are needed, primarily for one-time use (for example, in missiles).

General-purpose computers with automatic program control are the most refined means of automating the labor-consuming processes of human mental activity. Modern general-purpose digital computers are a complicated automatic computing complex that includes a processor, an immediate-access memory, one or several large-capacity external memories, and data input-output units. The process of computation is controlled by the control device and the pro-gram of computations located in the computer memory. The loading of particular units, the coordination of their operation, and the control of the sequence of solving problems are done by program equipment. The set of programs that per-forms these and many other functions is called software. Algorithmic languages (ALGOL, FORTRAN, COBOL, and others) are used to describe problem-solving. The input of initial data and programs and the output of results in the form most convenient for the user are accomplished by the set of input-output devices that are part of the general-purpose digital computer. Initial data may be given in the form of graphs, digital and textual documents, images of the object being calculated (for example, an overall view of a building, a cross section of an airplane wing, and so on), audiovisual display, and others.

Digital computers are characterized by high productivity and precision in the results obtained and by algorithmic universality, which results from the fact that the rearrangement of the digital computer to solve a new problem involves only a change in the computation program and the initial data stored in the computer memory, without changing the design of the machine itself.

Hybrid computer systems consist of organically linked analog and digital computers. Data exchange between analog-action and discrete-action computers is accomplished through special convertors. A division of functions between machines is typical for a combined system: the analog computer is used to reproduce fast-occurring processes with limited precision of variables, and the digital computer is used for computations with greater precision and for statistical processing of results. The hybrid computing system combines high precision and speed, which are more difficult to obtain by means of just one of the computers.

the term adopted in foreign literature (chiefly English-language) for equipment that operates automatically either according to a preassigned program or according to sequential instructions to solve mathematical problems and problems of economics, statistics, production planning, management, and so forth. The term usually refers to electronic computers (in Russian, elektronnye vychislitel’nye mashiny).