Bit
A bit is the smallest unit of information. Information in binary digital computing devices is stored and processed in the form of bits. One bit corresponds to either a 1 or a 0 in the binary number system, or designates one of two possible logical values or states, such as "true" or "false." The word "bit," a contraction of the words "binary digit," was introduced in 1946 by John Tukey, an American statistician and pioneering computer scientist.
A bit can be physically represented within a computer or other device as a voltage condition (e.g., high or low), or as the presence or absence of a pit in the aluminum film layer of a CD or CD-ROM, or as the state of magnetization of a tiny area of a magnetic disk or tape, or as any other distinct pair of physical states, whether electronic, chemical, mechanical, or other. Inside an electronic computer, bits are ultimately translated into high and low voltage states, regardless of the form of their storage on disk or elsewhere, for ease and speed of manipulation. These bits can be interpreted in any number of ways; for example, they could be used as a binary number (high voltage = 1, low voltage = 0), a decision state (high voltage = ON, low voltage = OFF), or as a binary logical state (high voltage = TRUE, low voltage = FALSE).
All modern computer information processing is accomplished by vast numbers of transistors switching on and off in a concerted effort, each reacting solely to the voltages (which represent bit values) on its input lines. (A transistor can be thought of as an electronic valve that either permits current to flow through itself or does not; these two states of flow are the transistor's "switched on" and "switched off" states.) For example, an Intel Pentium processor chip contains about 3.1 million transistors.
One bit conveys little information, but when 8 bits are grouped together to form a byte, many types of useful information can be encoded. A single byte can record a decimal digit (number from 0 to 9), letter of the alphabet, or other symbol. Interpreted as a positive integer number, a byte may take on any value from 00000000 (decimal 0) to 11111111 (decimal 255). In other words, a byte can take on 28 (2 raised to the eighth power) possible states, and these 28 = 256 states can, if one wishes to interpret them as numbers, be assigned to the numbers from 0 to 255. Bytes may be combined into groups of bytes called words. No computer is built to process information bit by bit, or even byte by byte. Rather, computers process words, 32 bits being a common word length. Since each bit comprising a word is symbolized by a voltage, and each voltage requires a separate conductor, this requires that bundles of 32 conductors (say) always run together from point to point on the processor chip or between the processor and memory, and that memory cells or other bit-processing units usually be lined up in groups of 32 to handle the word length. This leads to one way of classifying computer processors: a computer with a 32-bit word length can be classified as a "32-bit machine." Another way of classifying computers is according to the number of bits they devote to addressing locations in memory; a 16-bit address space, for instance, means that a single address can specify any single memory location out of 216 = 65,536 locations. To speak loosely of a machine as "32-bit" or "16-bit" might thus refer either to the width of its data path or to the width of its address field. Larger word lengths make computers faster (more bits processed per clock tick) but more difficult to build (more wires and wider registers to crowd onto the chip); longer address lengths enable computers to support larger programs, but require longer word lengths.
The physical sense of the word bit should be distinguished from its information-theoretic sense, for 1 bit physically resident in a computer does not necessarily represent 1 bit of actual information. Say that a certain bit stored in the computer represents the answer to some question: Yes or No. If we already know the answer to our question with 99.9% probability, then we do not learn much by inspecting the answer bit; we do not acquire a bit of information by inspecting the bit of memory. Thus one may speak of the number of "bits" in a computer memory independently of how much actual information the memory contains; a disk drive might store 10 billion bits all set to random values, or to 0, and thus contain no information at all.
This is the complete article, containing 780 words
(approx. 3 pages at 300 words per page).
