Information technology, generations, types, architecture and memory notes

Information Technology (IT) is the use of computers, storage, software and telecommunication to create, process, store, secure and exchange electronic data. Information is processed data that means something to a user. Technology is the set of tools that does that processing. A computer is the central tool of IT, not the whole of IT. Looking up roll 2145, printing a notice, or reading an OMR sheet are IT jobs: a machine turns data into information a person can act on.

Sinha’s working definition of a computer is an electronic device that accepts data as input, processes it according to a stored set of instructions (a program), and gives information as output, usually keeping both data and instructions in storage. Those four jobs are the IPO cycle plus storage. Input brings data in (keyboard, mouse, scanner). Process is the CPU’s work: arithmetic, comparison, moving bytes. Output sends the result out (monitor, printer, speaker). Storage holds data and programs so that the next step does not start from nothing. On a lab desktop: you type roll 2145 (input); the CPU matches it in a file (process); the screen shows Anil Sharma, B.A. Semester III (output); the hard disk keeps the file (storage).

Data and information are not the same word. Data are raw, unorganised facts and figures: 2145, 36, Bikaner. Information is data processed so that it is meaningful, timely and useful: “Anil Sharma, roll 2145, has to score at least 36 marks out of 100.” The same digits can be data in one line and information in the next. GIGO — garbage in, garbage out — is the limit: a computer does not know that a wrong roll was typed; it will process that wrong roll with perfect speed.

IT is built from named parts. Hardware is the body you can touch: cabinet, CPU chip, keyboard, mouse, monitor, printer, cables. Software is the set of programs that tell the hardware what to do: system software (the operating system, later in Unit 3) and application software (a word processor, a browser). Firmware is a small program sitting in ROM, used when the machine starts. People (sometimes called liveware) use the system. Data is what is processed. A network joins machines so that information can move; the types of network are Unit 4. A smartphone that sends a mark to a server is IT. A bank’s mainframe that clears a payment is IT. The lab mouse by itself is only hardware.

Sinha lists the characteristics that made computers spread. Speed: millions of instructions in a second — a mark-list that would take a clerk an afternoon. Accuracy: the same sum is the same every time if the data are right. Diligence: a machine does not tire after the hundredth OMR sheet. Versatility: the same CPU can total marks, draw a letter, or play a clip, once the right program is loaded. Storage: a disk holds more than a cupboard of paper, and can fetch a file in seconds. Reliability: electronic parts fail less often than a tired clerk; the same job can run all night. Automation: once programmed, the machine follows the steps without a person repeating each one. The same books name the limits. A computer has no intelligence of its own and no feelings. It cannot decide a new rule that was never programmed. It needs power, instructions, and correct data.

A generation of computers is a stage of electronic machines that share one main switching technology and, with it, a typical size, speed, language and use. The word does not mean “every new laptop is a new generation.” Five generations are named in Sinha, Jain and Rapidex. Before them sit only calculating aids — abacus, Pascal’s adder, Babbage’s Analytical Engine — which are ancestors, not the five generations of this unit. The five generations begin when the switch inside the machine becomes electronic.

First generation (about 1940–1956). The switch is the vacuum tube (thermionic valve). Machines are huge, hot, and hungry for power; a tube fails often. Memory is magnetic drum or delay line. The language is machine language — only 0 and 1. Input is punched card or paper tape; output is print. ENIAC (1946) and UNIVAC I (1951) are the classroom names. A first-generation machine filled a room. It could not sit on a lab table.

Second generation (about 1956–1963). The switch is the transistor. Transistors are smaller, cooler and more reliable than tubes. Memory is magnetic core. Languages rise to assembly and the first high-level languages: FORTRAN and COBOL. Batch processing is the common way of work: a pile of jobs, one after another. IBM 1401 is a typical name. Size falls from a room toward a large cabinet, still not a personal desk machine.

Third generation (about 1964–1971). The switch is the integrated circuit (IC) — many transistors on one silicon chip. Magnetic core continues; keyboards and monitors become ordinary. Operating systems appear; multiprogramming and time-sharing let more than one program (and more than one user) share the CPU. High-level languages spread. IBM System/360 is the classroom mainframe of this generation. Minicomputers also belong here in size: a department can own a machine, not only a national lab.

Fourth generation (1971 to the present for personal machines). The switch is the microprocessor — a whole CPU on one chip (Intel 4004, 1971), built with LSI and then VLSI. Semiconductor RAM replaces core. The microcomputer and the IBM PC (1981) put a computer on a desk. GUI, mouse and local networks ride this generation. The lab PC, the laptop, the phone in a bag — all fourth-generation machines in the switching sense. A new model of the same microprocessor family is still fourth-generation hardware, not a new generation.

Fifth generation (present and still developing). The aim is machines that take a step toward artificial intelligence: parallel processing, ULSI chips, huge memory, and an interface closer to natural language. The fifth generation is not “the phone you bought last year.” Fifth generation is named by the attempt to make the machine reason, see, or speak with less rigid programming. AI as a full topic is Unit 5; here the generation-label is enough: ULSI, parallelism, AI as the intended work.

This unit’s types are types by size, speed and who can share the machine. They are not analog versus digital (almost every machine named here is digital). They are not “desktop versus laptop” as two of the four names. The four names are micro, mini, mainframe, super. Learn them as a ladder: cheapest and smallest at the micro end; fastest and most expensive at the super end.

A microcomputer is built around a single microprocessor and is meant, in the classic definition, for one person at a time. The lab desktop, a laptop, a tablet and a smartphone are all microcomputers. Cost is lowest of the four. Speed and storage are enough for typing, browsing, a spreadsheet, a college assignment. A microcomputer can sit on a network and talk to a mainframe; it does not stop being a micro because it is online. “Personal computer (PC)” is the older office name for a desktop micro. A phone is a micro, not a mini: mini is the next rung, not a synonym for small.

A minicomputer is a midrange, multi-user machine. Historically DEC’s PDP and VAX, and later departmental servers, sat here: tens or a few hundreds of users, more storage and I/O than a PC, less than a mainframe. A college might once have put student records on a mini while each lab still had micros for typing. The word “mini” is relative to mainframe, not to a phone. Mini means medium size, multi-user, department-level.

A mainframe is a large, very reliable system for an organisation that must run many jobs and many terminals at once: a bank, the railways, a census, an examination board. I/O capacity is huge — thousands of screens, printers, tape and disk — because the work is not one student’s file but lakhs of records. A mainframe is not “a very big PC.” It is built for volume, security and uptime: many transactions, all day. The result server that prints a whole university’s marks is mainframe-scale work, even if the student sees only a micro in the lab.

A supercomputer is the fastest class, built for scientific and engineering jobs that a mainframe is not shaped for: weather and climate models, nuclear simulation. The measure is FLOPS (floating-point operations per second), not only MIPS. Architecture is massively parallel: thousands of processors working together. India’s PARAM series (C-DAC) is the classroom name. Cray is the older international name. The contrast with a mainframe is the job: a mainframe processes huge numbers of business records; a supercomputer runs a few enormous scientific calculations at the highest speed. A supercomputer is not used to type a letter. A microcomputer is not used to forecast a cyclone over the Thar.

Architecture is the named arrangement of parts and the paths among them. The stored-program (von Neumann) model has five functional units: input, output, memory, ALU (arithmetic and logic unit), and control unit. Instructions and data live in the same memory. The CPU fetches an instruction, decodes it, executes it, then fetches the next. That loop is the fetch–decode–execute cycle.

A bus is a shared set of wires that carry signals among CPU, memory and I/O. Three buses are named. The data bus carries the actual bytes (an instruction, a number, a character). The address bus carries the location: which memory cell or which device. The control bus carries commands and status: read, write, interrupt, clock. Width of the data bus (8, 16, 32, 64 bits) is how many bits move together in one step.

The CPU (Central Processing Unit) is the brain of the system. In the fourth generation it is a microprocessor on one chip, seated on the motherboard. It has two large inner parts plus registers. The ALU does arithmetic (add, subtract, multiply, divide) and logic (AND, OR, NOT, compare, shift). The control unit (CU) does not add numbers. It directs: it reads the instruction, decides which part of the ALU or memory must work, and times the steps with the system clock. Clock speed is given in GHz on a modern micro; older books still say MHz. A faster clock means more cycles per second.

Registers are tiny, very fast storage cells inside the CPU. They are not RAM. RAM sits outside the CPU on the motherboard. Named registers: the program counter (PC) holds the address of the next instruction. The instruction register (IR) holds the instruction just fetched. The accumulator (and other general registers) hold data the ALU is working on. The MAR (memory address register) holds the address being read or written. The MDR / MBR (memory data / buffer register) holds the byte going to memory or coming from it. Because registers are few and on-chip, they are the fastest store in the hierarchy; cache and RAM are slower steps down.

The CPU is not the metal box. The cabinet holds motherboard, CPU, RAM sticks, disk, power supply and ports. The CPU is the chip (with its heat-sink) that processes. Storage holds; the monitor shows; the CPU executes.

Primary memory (main memory) is the store the CPU can address directly on the memory bus. It is semiconductor memory on this generation of machines. It is faster than disk and smaller. When a program runs, the bytes the CPU needs right now sit here, not only on the hard disk. This unit names three primary stores: RAM, ROM, and cache.

Units of measure: a bit is 0 or 1; a nibble is 4 bits; a byte is 8 bits (enough for one character in the older codes). In the binary convention Sinha uses, 1024 bytes = 1 KB; 1024 KB = 1 MB; then GB, TB. Do not mix these units with the disk’s job: the disk keeps the file overnight; RAM holds the running copy.

RAM (Random Access Memory) is read-and-write memory. Any location can be reached in roughly the same time (that is the “random” in the name — not “haphazard”). RAM is volatile: when power goes, the contents go. That is why an unsaved assignment on the lab PC vanishes in a cut. Two common kinds: DRAM (dynamic) stores each bit as charge on a tiny capacitor; the charge leaks, so the chip must refresh thousands of times a second. DRAM is cheap and is used as main memory. SRAM (static) uses flip-flops, needs no DRAM-style refresh, is faster and dearer, and is used in cache. Capacity of main RAM is quoted in GB on a modern micro.

ROM (Read-Only Memory) is non-volatile. Contents stay when power is off. The CPU can read it in ordinary use; the user does not type into it as into RAM. Firmware lives here: the small start-up program (BIOS on a PC) that tests hardware and loads the operating system from disk. “Read-only” does not mean the chip was never written; it means ordinary programs do not treat it as working RAM. Kinds: Mask ROM is written at the factory and cannot be changed. PROM is blank when bought and can be programmed once with a PROM programmer. EPROM has a quartz window; ultraviolet light erases it, then it can be programmed again. EEPROM is erased electrically, without removing the chip or using UV. Flash is a form of EEPROM used in BIOS chips and in pen drives. The pen drive itself is secondary storage; the flash technology is shared.

Cache memory is a small, very fast SRAM store between the CPU and main RAM. The CPU is faster than RAM; without cache it would wait. Cache keeps copies of the bytes the CPU is using now and is likely to use next (locality of reference). Levels: L1 smallest and fastest, on the chip; L2 larger; L3 larger still, often shared among cores. Cache is not a substitute for RAM and not a disk. Speed order: registers, then cache, then RAM, then secondary. The memory between CPU and RAM is cache.

Secondary memory (auxiliary / backup storage) is non-volatile store that is not on the CPU’s main memory bus in the same way RAM is. It is cheaper per bit, larger, and slower. The running program is copied from here into RAM; results that must survive a power cut are written back here. If primary memory is the desk you work on, secondary memory is the cupboard. The cupboard keeps the assignment overnight. The desk is cleared when the lab closes.

Two ways of reaching data must be kept apart. Sequential access means you pass records in order: to reach file 500 you read 1–499 first. Magnetic tape is sequential. Direct access (also called random access on disk) means the head can go to a location without reading everything before it. Disk and flash are direct.

Magnetic disk — HDD. A hard disk drive stores bits as magnetic spots on spinning platters. Each platter face is divided into concentric tracks; each track is cut into sectors (a sector is the smallest addressable block, often 512 bytes in older books). A cylinder is the same track number on all platters, reached without moving the head arm in or out. Read/write heads float above the surface. Access is direct. Capacity today is in TB. It is the classic “C: drive” of a lab PC. Moving parts mean it can fail mechanically. Floppy disk is the older small magnetic disk (1.44 MB in the last common size); it is history in the lab, still in the book list.

Optical disk. CD (about 700 MB), DVD (4.7 GB and up), Blu-ray (tens of GB) store bits as pits read by laser. CD-ROM is read-only; CD-R is write-once; CD-RW is rewritable. Useful for distributing a notes pack or a film; too slow and small to be the only disk of a modern PC.

Flash and SSD. A USB pen drive, an SD card, and a solid-state drive use flash memory: no spinning platter, silent, shock-resistant, faster random access than HDD. The pen drive in a student’s bag is secondary storage you can carry. An SSD inside the cabinet is secondary storage that has replaced many HDDs. Flash here is a device class of secondary memory; it is not RAM. RAM forgets at power-off; a formatted pen drive does not.

Magnetic tape. Sequential access. Cheap for backup of a mainframe or a college server. Not for the file you open every minute in the lab.

An input device sends data and instructions into the computer. An output device sends results to the user or to another machine. Some devices do both (touch screen, modem). Pointing devices are a special class of input — they move a pointer on the screen — and have the next lesson. This lesson is the rest of the I/O set.

Keyboard. The standard character input. The common layout is QWERTY (named from the first six letter keys). Keys include alphanumeric, function (F1–F12), control, and a numeric pad. It does not draw the pointer; it types codes. Scanner makes a digital picture of a page. Microphone is sound input; a webcam is image and video input (the camera as hardware is I/O; teleconferencing as a use is Unit 4).

Three readers must stay distinct. A barcode reader shines light on the printed bars of a book or a shop item and turns the pattern into a number. MICR (Magnetic Ink Character Recognition) reads the special magnetic-ink digits at the bottom of a bank cheque. OCR (Optical Character Recognition) turns printed letters into editable text — a scanned page of a notice, not a tick in a bubble. OMR (Optical Mark Recognition) reads marks on a printed sheet: you darken A, B, C or D and a machine scores the page. OMR is an input device in this unit and also the method of this exam.

Monitor (VDU — Visual Display Unit). Soft-copy output. Older labs used CRT (cathode-ray tube): a heavy glass tube, the picture drawn by an electron beam. Current labs use LCD or LED flat panels: thin, cooler, lower power. Soft copy is on screen and vanishes when you switch off; hard copy is on paper.

Printers make hard copy. By how they mark the page: impact printers (dot-matrix, daisy-wheel) strike the page through a ribbon and can print carbon copies; non-impact printers (inkjet, laser) do not strike. By how much they print at once: a character printer prints one character; a line printer prints a whole line (older mainframe rooms); a page printer (laser) composes a full page. Laser is common in an office; inkjet is cheaper to buy and dearer per page in ink. A plotter is not a printer of paragraphs: it draws large line-work (maps, building plans) with pens or similar tools. Speakers and headphones are sound output. A projector throws the screen onto a wall for a class.

The CPU is not an output device. The hard disk is storage, not an input device as its home in this unit: it can feed RAM and take results from RAM, but it sits under secondary memory. The mouse is input (and a pointing device). The printer is output. The touch screen is both. Monitor = soft copy; printer = hard copy.

A pointing device is an input device that controls a pointer (cursor) on the screen so that you can select, drag and draw without typing coordinates. It belongs under input, but it is listed on its own because almost every fourth-generation micro with a GUI depends on one. The control unit still receives codes; the difference is the human gesture: move, click, tap, drag — not a letter key.

The mouse is the default pointing device of a lab PC. It sits beside the keyboard. Motion on the desk becomes motion of the arrow. Buttons click (select, open, menu); a wheel scrolls. Older mice used a rubber ball (mechanical); current lab mice are optical or laser, reading the surface with a sensor, no ball to clog. A trackball is a mouse turned over: the ball stays in a fixed housing and the thumb rolls it — useful where desk space is short. A touchpad is the flat pointing surface on a laptop. A joystick is a stick on a base, built for games and some lab controls, not for typing an essay. A light pen points at a CRT screen; rare in a modern LCD lab, still in book lists. A digitizing tablet with a stylus is for drawing maps and diagrams. A touch screen is both a pointing device and a display: a finger or stylus on the glass is the pointer. Phones and many kiosks are touch-screen micros.

A pointing device is not every input device. A keyboard is input but not pointing. A scanner is input but not pointing. A printer is output. The monitor shows the pointer; it does not point. The CPU and RAM are not pointing devices. The mouse does not store the assignment; the disk does. The mouse only tells the CPU where on the screen the next click is.