Encyclopedia of Computer Science and Technology

operating system 353tion programmers with a way to invoke system services, andgives users a way to control programs and organize files.DevelopmentThe earliest computers were started with a rudimentary“loader” program that could be used to configure the systemto run the main application program. Gradually, a moresophisticated way to schedule and load programs, link programstogether, and assign system resources to them wasdeveloped (see job control language and mainframe).As systems were developed that could run more thanone program at a time (see multitasking), the duties of theoperating systems became more complex. Programs had tobe assigned individual portions of memory and preventedfrom accidentally overwriting another program’s memoryarea. A technique called virtual memory was developed toenable a disk drive to be treated as an extension of themain memory, with data “swapped” to and from the diskas necessary. This enabled the computer to run more and/or larger applications. The operating system, too, becamelarger, amounting to millions of bytes worth of code.During the 1960s, time sharing became popular particularlyon new smaller machines such as the DEC PDPseries (see minicomputer), allowing multiple users to runprograms and otherwise interact with the same computer.Operating systems such as Multics and its highly successfuloffshoot UNIX developed ways to assign security levelsto files and access levels to users. The UNIX architecturefeatured a relatively small kernel that provides essentialprocess control, memory management, and file system services,while drivers performed the necessary low-level controlof devices and a shell provided user control. (See unix,kernel, device driver, and shell.)Starting in the late 1970s, the development of personalcomputers recapitulated in many ways the earlier evolutionof operating systems in the mainframe world. Earlymicrocomputers had a program loader in read-only memory(ROM) and often rudimentary facilities for entering, running,and debugging assembly language programs.During the 1980s, more complete operating systemsappeared in the form of Apple DOS, CP/M, and MS-DOS forIBM PCs. These operating systems provided such facilitiesas a file system for floppy or hard disk and a command-lineinterface for running programs or system utilities. Thesesystems could run only one program at a time (althoughexploiting a little-known feature of MS-DOS allowed additionalsmall programs to be tucked away in memory).As PC memory increased from 640 kB to multiple megabytes,operating systems became more powerful. Apple’sMacintosh operating system and Microsoft Windows couldmanage multiple tasks. Today personal computer operatingsystems are comparable in sophistication and capability tothose used on mainframes. Indeed, PCs can run UNIX variantssuch as the popular Linux.ComponentsWhile the architecture and features of operating systemsdiffer considerably, there are general functions common toalmost every system. The “core” functions include “booting”the system and initializing devices, process management(loading programs intro memory assigning them a share ofprocessing time), and allowing processes to communicatewith the operating system or one another (see kernel).Multiprogramming systems often implement not only processes(running programs) but also threads, or sections ofcode within programs that can be controlled separately.A memory management scheme is used to organize andaddress memory, handle requests to allocate memory, freeup memory no longer being used, and rearrange memory tomaximize the useful amount (see memory management).There is also a scheme for organizing data created orused by programs into files of various types (see file). Mostoperating systems today have a hierarchical file system thatallows for files to be organized into directories or foldersthat can be further subdivided if necessary. In operatingsystems such as UNIX, other devices such as the keyboardand screen (console) and printer are also treated like files,providing consistency in programming. The ability to redirectinput and output is usually provided. Thus, the outputof a program could be directed to the printer, the console,or both.In connecting devices such as disk drives to applicationprograms, there are often three levels of control. At thetop level, the programmer uses a library function to opena file, write data to the file, and close the file. The libraryitself uses the operating system’s lower-level input/output(I/O) calls to transfer blocks of data. These in turn aretranslated by a driver for the particular device into the lowlevelinstructions needed by the processor that controls thedevice. Thus, the command to write data to a file is ultimatelytranslated into commands for positioning the diskhead and writing the data bytes to disk.A typical operating system processes user commands or actionsusing an interface (such as a shell). Both user commands andrequests from application programs communicate with the operatingsystem through the application Programming Interface (API),which provides services such as file, memory, process, and networkmanagement.