Architecture (562409), страница 19
Текст из файла (страница 19)
Another type of storage media, called a flash memory, traps small amounts of electric charge in “wells” on the surface of a chip. Side effects of this trapped charge, such as the electric field it creates, are later used to read the stored value. To rewrite to flash memory, the charges in the wells must first be drained. Such drives are useful for storing information that changes infrequently.
| IV | HISTORY |
The earliest mechanical information storage devices were music boxes of the 18th century that encoded sequences of notes as pins on a revolving drum. In the early 1800s Joseph Marie Jacquard used paper cards with information recorded as holes punched in them to control weaving looms. This idea of punched-card storage was later used by British mathematician and inventor Charles Babbage in the first programmable computer. The holes on each card allowed an arm to pass through and activate a mechanism on the other side. In the census of 1890, American inventor Herman Hollerith used punched cards to hold data. These cards were then read by machines in which rows of electrical contacts sensed when a hole was present. In the 1940s the first electronic computers used punched cards and rolls of paper tape with punched holes for storing both programs and data. Before magnetic media became popular, various memory devices such as cathode ray tubes and mercury-delay lines were used to store information. The first use of magnetic memory devices came in the late 1940s in the form of magnetic tapes and drums, and then magnetic cores, in which small doughnuts of magnetic material each stored one bit of information.
In the 1970s the first hard disks appeared, with platters as large as four feet in diameter. Shortly thereafter, the International Business Machines Corporation (IBM) invented floppy disks as a mechanism to load micro-programs into their mainframe computers. CD-ROMs were introduced in the early 1980s to store music in digital form for high-fidelity playback. These technologies underwent explosive growth with their use in mass-market personal computer systems.
| V | FUTURE TECHNOLOGIES |
Although magnetic and CD-ROM technologies continue to increase in storage density, a variety of new technologies are emerging. Redundant Arrays of Independent Disks (RAIDs) are storage systems that look like one device but are actually composed of multiple hard disks. These systems provide more storage and also read data simultaneously from many drives. The result is a faster rate of data transfer to the CPU, which is important for many very high speed computer applications, especially those involving large databases of information.
Several experimental technologies offer the potential for storage densities that are thousands or millions of times better than is possible today. Some approaches use individual molecules, sometimes at superconducting temperatures, to trap very small magnetic fields or electrical charges for data storage. In other technologies, large two-dimensional data sets such as pictures are stored as holograms in cubes of material. Individual bits are not stored at any one location, but instead are spread out over a much larger area and mixed in with other bits. Loss of information from any one spot thus does not cause the irreplaceable loss of any one bit of information.
Mainframe Computer
Mainframe Computer, a high-level computer designed for the most intensive computational tasks. Mainframe computers are often shared by multiple users connected to the computer via terminals. The most powerful mainframes, called supercomputers, perform highly complex and time-consuming computations and are used heavily in both pure and applied research by scientists, large businesses, and the military.
Data Processing
| I | INTRODUCTION |
Data Processing, in computer science, the analysis and organization of data by the repeated use of one or more computer programs. Data processing is used extensively in business, engineering, and science and to an increasing extent in nearly all areas in which computers are used. Businesses use data processing for such tasks as payroll preparation, accounting, record keeping, inventory control, sales analysis, and the processing of bank and credit card account statements. Engineers and scientists use data processing for a wide variety of applications, including the processing of seismic data for oil and mineral exploration, the analysis of new product designs, the processing of satellite imagery, and the analysis of data from scientific experiments.
Data processing is divided into two kinds of processing: database processing and transaction processing. A database is a collection of common records that can be searched, accessed, and modified, such as bank account records, school transcripts, and income tax data. In database processing, a computerized database is used as the central source of reference data for the computations. Transaction processing refers to interaction between two computers in which one computer initiates a transaction and another computer provides the first with the data or computation required for that function.
Most modern data processing uses one or more databases at one or more central sites. Transaction processing is used to access and update the databases when users need to immediately view or add information; other data processing programs are used at regular intervals to provide summary reports of activity and database status. Examples of systems that involve all of these functions are automated teller machines, credit sales terminals, and airline reservation systems.
| II | THE DATA-PROCESSING CYCLE |
The data-processing cycle represents the chain of processing events in most data-processing applications. It consists of data recording, transmission, reporting, storage, and retrieval. The original data is first recorded in a form readable by a computer. This can be accomplished in several ways: by manually entering information into some form of computer memory using a keyboard, by using a sensor to transfer data onto a magnetic tape or floppy disk, by filling in ovals on a computer-readable paper form, or by swiping a credit card through a reader. The data are then transmitted to a computer that performs the data-processing functions. This step may involve physically moving the recorded data to the computer or transmitting it electronically over telephone lines or the Internet.
Once the data reach the computer, the computer processes it. The operations the computer performs can include accessing and updating a database and creating or modifying statistical information. After processing the data, the computer reports summary results to the program’s operator.
As the computer processes the data, it stores both the modifications and the original data. This storage can be both in the original data-entry form and in carefully controlled computer data forms such as magnetic tape. Data are often stored in more than one place for both legal and practical reasons. Computer systems can malfunction and lose all stored data, and the original data may be needed to recreate the database as it existed before the crash.
The final step in the data-processing cycle is the retrieval of stored information at a later time. This is usually done to access records contained in a database, to apply new data-processing functions to the data, or—in the event that some part of the data has been lost—to recreate portions of a database. Examples of data retrieval in the data-processing cycle include the analysis of store sales receipts to reveal new customer spending patterns and the application of new processing techniques to seismic data to locate oil or mineral fields that were previously overlooked.
| III | HISTORY |
To a large extent, data processing has been the driving force behind the creation and growth of the computer industry. In fact, it predates electronic computers by almost 60 years. The need to collect and analyze census data became such an overwhelming task for the United States government that in 1890 the U.S. Census Bureau contracted American engineer and inventor Herman Hollerith to build a special purpose data-processing system. With this system, census takers recorded data by punching holes in a paper card the size of a dollar bill. These cards were then forwarded to a census office, where mechanical card readers were used to read the holes in each card and mechanical adding machines were used to tabulate the results. In 1896 Hollerith founded the Tabulating Machine Company, which later merged with several other companies and eventually became International Business Machines Corporation (IBM).
During World War II (1939-1945) scientists developed a variety of computers designed for specific data-processing functions. The Harvard Mark I computer was built from a combination of mechanical and electrical devices and was used to perform calculations for the U.S. Navy. Another computer, the British-built Colossus, was an all-electronic computing machine designed to break German coded messages. It enabled the British to crack German codes quickly and efficiently.
The role of the electronic computer in data processing began in 1946 with the invention of the ENIAC, the first all-electronic computer. The U.S. armed services used the ENIAC to tabulate the paths of artillery shells and missiles. In 1950 Remington Rand Corporation introduced the first nonmilitary electronic programmable computer for data processing. This computer, called the UNIVAC, was initially sold to the U.S. Census Bureau in 1951; several others were eventually sold to other government agencies.
With the purchase of a UNIVAC computer in 1954, General Electric Company became the first private firm to own a computer, soon followed by Du Pont Company, Metropolitan Life, and United States Steel Corporation. All of these companies used the UNIVAC for commercial data-processing applications. The primary advantages of this machine were its programmability, its high-speed arithmetic capabilities, and its ability to store and process large business files on multiple magnetic tapes. The UNIVAC gained national attention in 1952, when the American Broadcast Company (ABC) used a UNIVAC during a live television broadcast to predict the outcome of the presidential election. Based upon less than 10 percent of the election returns, the computer correctly predicted a landslide victory for Dwight D. Eisenhower over his challenger, Adlai E. Stevenson.
In 1953, IBM produced the first of its computers, the IBM 701—-a machine designed to be mass-produced and easily installed in a customer’s building. The success of the 701 led IBM to manufacture many other machines for commercial data processing. The sales of IBM’s 650 computer were a particularly good indicator of how rapidly the business world accepted electronic data processing. Initial sales forecasts were extremely low because the machine was thought to be too expensive, but over 1800 were eventually made and sold.
In the 1950s and early 1960s data processing was essentially split into two distinct areas, business data processing and scientific data processing, with different computers designed for each. In an attempt to keep data processing as similar to standard accounting as possible, business computers had arithmetic circuits that did computations on strings of decimal digits (numbers with digits that range from 0 to 9). Computers used for scientific data processing sacrificed the easy-to-use decimal number system for the more efficient binary number system in their arithmetic circuitry.
The need for separate business and scientific computing systems changed with the introduction of the IBM System/360 family of machines in 1964. These machines could all run the same data-processing programs, but at different speeds. They could also perform either the digit-by-digit math favored by business or the binary notation favored for scientific applications. Several models had special modes in which they could execute programs from earlier IBM computers, especially the popular IBM 1401. From that time on, almost all commercial computers were general-purpose.
One notable exception to the trend of general-purpose computers and programming languages is the supercomputer. Supercomputers are computers designed for high-speed precision scientific computations. However, supercomputers are sometimes used for data processing that is not scientific. In these cases, they must be built so that they are flexible enough to allow other types of computations.
The division between business and scientific data processing also influenced the development of programming languages in which application programs were written. Two such languages that are still popular today are COBOL (COmmon Business Oriented Language) and Fortran (FORmula TRANslation). Both of these programming languages were developed in the late 1950s and early 1960s, with COBOL becoming the programming language of choice for business data processing and FORTRAN for scientific processing. In the 1970s other languages such as C were developed. These languages reflected the general-purpose nature of modern computers and allowed extremely efficient programs to be developed for almost any data-processing application. One of the most popular languages currently used in data-processing applications is an extension of C called C++. C++ was developed in the 1980s and is an object-oriented language, a type of language that gives programmers more flexibility in developing sophisticated applications than other types of programming languages.
Development of computer science
Computer science as an independent discipline dates to only about 1960, although the electronic digital computer that is the object of its study was invented some two decades earlier. The roots of computer science lie primarily in the related fields of electrical engineering and mathematics. Electrical engineering provides the basics of circuit design—namely, the idea that electrical impulses input to a circuit can be combined to produce arbitrary outputs. The invention of the transistor and the miniaturization of circuits, along with the invention of electronic, magnetic, and optical media for the storage of information, resulted from advances in electrical engineering and physics. Mathematics is the source of one of the key concepts in the development of the computer—the idea that all information can be represented as sequences of zeros and ones. In the binary number system, numbers are represented by a sequence of the binary digits 0 and 1 in the same way that numbers in the familiar decimal system are represented using the digits 0 through 9. The relative ease with which two states (e.g., high and low voltage) can be realized in electrical and electronic devices led naturally to the binary digit, or bit, becoming the basic unit of data storage and transmission in a computer system.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
