Introduction

Operating systems (OSs) work together with the central processing units (CPUs) of computers, as noted in the LIS 386.13 PowerPoint presentation entitled Introduction to Computing, The CPU uses the OS to

The OS uses the CPU to

The OS in your computer is probably either a version of Windows (e.g., XP, 98, Millennium Edition, 2000), or Macintosh System 9 or X (for 10, also known as Jaguar). You are probably also aware that an OS named Unix is widely used, e.g., on many computers that house Webpages or email servers. (If you are attending classes on the UT-Austin campus, you may have noticed machines running under Unix in the iSchool Information Technology Laboratory.) You may even be aware that a variety of Unix known as Linux has become widely popular in the last few years as an alternative to Windows on PC-type microcomputers.

You can think of the OS as being the manager of the computer system as whole, in the sense that it directs the operations of both the hardware and the other software, i.e., the non-OS software. This other software includes (1) applications programs, the programs like Microsoft Word and Netscape that perform the tasks that the user wants done, and (2) utility programs, programs like Telnet and WinZip that handle relatively simple tasks which, although necessary from time to time, are peripheral to the main goals of the user.

Components of an OS

An OS typically has four primary modules: Process Manager, Memory Manager, File Manager, and Device Manager.

Process Manager

The Process Manager handles the scheduling of activities within the CPU and the assignment to the CPU of individual tasks involved in an application or utility program. In the typical situation in modern computers, more than one application program and/or utility program will be running at any given moment, a situation known as "multiple tasking" (which is often shortened to "multitasking").

Strictly speaking, the tasks do not run simultaneously, for the CPU can handle only one task at a time. However, the Process Manager sees to it that the CPU works on the steps of a given task for only a few microseconds at a time before the Process Manager gives the CPU a different task for another short period of time. The Process Manager chooses tasks from among all the application and utility programs that the user is using concurrently, and rotates the CPU's services among the various tasks from the various programs. The CPU does this so rapidly and smoothly that the user perceives the programs as running simultaneously. If more than one user is using the computer concurrently, i.e., if there are multiple users (a situation that is often referred to as "multiuser" operation), then the Process Manager rotates the CPU's services among all the tasks of all the users, again usually doing this so rapidly and smoothly that each individual user perceives his or her own programs as running simultaneously.

The Process Manager module switches the CPU from one task to another according to complicated rules that depend on such factors as:

• the maximum length of time that the CPU is allowed to handle any one task at a stretch;
• the Process Manager's monitoring whether the CPU is actually doing useful work or is just waiting for something, e.g., for a particular datum to be delivered to it for a particular task, in which case the Process Manager will usually set the CPU to work on a different task till the needed datum arrives; and
• the Process Manager's monitoring inputs from the user(s) (e.g., from the user's clicking the mouse or touching a key) and signals from such sources as modems and disk drives, to which inputs the Process Manager responds according to further rules.

This description of CPU-task-switching merely scratches the surface of the complexities possible, but at least it may suggest to you why the writing of OSs is usually regarded as the pinnacle of the art of programming.

Memory Manager

The Memory Manager module controls the use of the CPU's primary memory resource, random-access memory (RAM). RAM operates at very high speeds; e.g., the RAM available today (2000) provides storage and retrieval of data in less than 100 nanoseconds. RAM consists physically of integrated circuit chips. RAM chips used to be extremely expensive but have dropped in price to current levels of around $1 per megabyte or less. The Memory Manager

checks the validity of each request for memory space, and, if it is a legal request . . . provides a portion [of RAM] not already in use. In a multiuser environment it sets up a table to keep track of who is using which section of memory. Finally, when the time comes to reclaim the memory, it "deallocates" it.

Of course, one of the Memory Manager's primary responsibilities is to preserve the space in [RAM] occupied by the operating system itself [to prevent that space] from being accidentally or intentionally altered. (Endnote 1)

File Manager

The File Manager module handles the files used by the system, especially those stored on the system's disk drive or drives. On each disk drive, files over some threshold length (typically, somewhere between 512 and 4,096 bytes) are broken up into pieces, each of which is that threshold length or less. The File Manager keeps track of where the first piece of each file is physically located (in terms of the hardware of the disk drive), and of where the subsequent links in the chain of pieces of that individual file are located. The File Manager keeps track of what physical locations are available for the storage of new files, and it handles the deletion of files that the user(s) or the application programs discard. It also controls access to read-only files and to hidden files, keeps track of whether the file has been backed up or not, and records the time and date of each change to a file.

Device Manager

The Device Manager

Historical Background

Early OSs

Early computers, such as those developed before and during World War II, had only minimal OSs. When first turned on, such computers relied on a few initial processes that were embodied in the hardware of the machine and sufficed to get the computer ready to receive inputs. The inputs came from such sources as toggle switches set by hand on a console, plugboards (cf. Introduction to Computing) wired in advance by humans, and signals from punched-paper-tape readers and punched-card readers. Programs consisted of instructions in the actual machine language; i.e., they consisted of sequences of 0s and 1s that directly activated the CPU and the other hardware devices. In this situation , it was hard to say which inputs constituted the OS and which the program.

By the late 1940s, computer specialists had realized that it made sense to input to a computer, when it was first turned on, a small program that was designed to handle the CPU and the hardware devices in certain fairly standard ways. Application programs could then rely on this program to handle the details of their communications with the CPU and the other devices. Though tiny compared to today's OSs, these small startup programs were the OSs for the computers of that era. The initial tiny OSs soon expanded to include modules that would handle application programs written in assembly languages (rather than machine languages) and in the first generation of procedural languages (e.g., ALGOL, COBOL, and FORTRAN). As computer hardware developed rapidly during the 1950s and 1960s, so did the OSs that the various computer manufacturers supplied for their machines. Each different manufacturer used its own OSs, and these were often different for different lines of computers from a single manufacturer.

Unix

Unix was first developed in 1969 by programmers working at Bell Laboratories (then a unit of Western Electric, the manufacturing arm of the American Telephone and Telegraph Company [also known as "Ma Bell"]; now a unit of Lucent Corporation, the successor of Western Electric). As described in UNIX Power Tools (Endnote 2):

UNIX is unique in that it wasn't designed as a commercial operating system meant to run application programs, but as a hacker's toolset, by and for programmers. In fact, an early release of the operating system went by the name PWB (Programmer's Work Bench).

When Ken Thompson and Dennis Ritchie first wrote UNIX at AT&T Bell Labs, it was for their own use, and for their friends and co-workers. Utility programs were added by various people as they had problems to solve. Because Bell Labs wasn't in the computer business, source code was given out to universities for a nominal fee. Brilliant researchers wrote their own software and added it to UNIX in a spree of creative anarchy that hasn't been equalled since. . .

Unix has become a major operating system (indeed, "[m]any hackers feel that Unix is the Right Thing—the One True Operating System" [Endnote 3]) because it was designed, from its birth, to be a multitasking, multiuser system. Because of its power and sophistication, and because its source code soon became widely available, Unix also became, early in its life, a multiplatform system (i.e., an operating system capable of being adapted to run on hardware from many different manufacturers).

Apple Computer Company, founded by Steve Wozniak and Steve Jobs in 1976, put its first microcomputer, the Apple I, on the market that same year. The Apple I, Apple II, Apple III, and the early Macintoshes, introduced in 1984, were much too small to store in their memory a useful version of Unix. Hence Apple developed its own Disk Operating System, initially often called Apple DOS; later versions became known as Macintosh System 1, 2, 3, etc. Some of the other microcomputers of the late 1970s and early 1980s, such as the Osborne I and the Radio Shack TRS-80, used a compact operating system called CP/M (the letters initially stood for Control Program/Monitor, and later for Control Program for Microcomputers), which Gary Kildall first was originally wrote in 1973. Other microcomputers of this late 1970s-early 1980s period, e.g., the Atari and the Commodore, used small operating systems of their own. All of the microcomputers and their operating systems of this period used, as the basic unit of storage and processing, a "computer word" consisting of 8 bits; i.e., these systems used a "word" equal to one byte.

The IBM PC (Personal Computer), introduced in 1981 and the first microcomputer to use a 16-bit computer word, was likewise too small to use Unix. In what must be one of the most costly mistakes ever made by a business, IBM, rather than developing in-house an operating system for the PC, arranged to obtain an operating system for the PC from a small Seattle company, Microsoft. The co-owner of Microsoft, Bill Gates, in turn arranged to buy a 16-bit version of the CP/M operating system, known as QD-DOS (Quick and Dirty Disk Operating System), that another small company, Seattle Computer Products, had nearly finished writing. Microsoft made minor improvements in QD-DOS and licensed it—Bill Gates being too smart to sell it—to IBM with the name MS-DOS. The IBM PC rapidly became a commercial success, and so did MS-DOS and Microsoft.

Although it was developed in the 1980s, more than 10 years after the inception of Unix, MS-DOS was a single-task, single-user OS. Because the Macintosh OS was a multitasking OS, it held an early advantage over MS-DOS. Not till the 1990s did Microsoft develop a capability for multitasking, with its Windows OS. Both the Macintosh OS and Windows (98 and ME) continue to be essentially single-user OSs, although Windows NT and Windows 2000 have some multiuser capability (they can handle small numbers of simultaneous users).

Linux

Linux is an operating system that is relatively compact but mimics Unix accurately. For all practical purposes, it can be thought of as a version of Unix. Its basic components were written in 1991 by Linus Torvalds. Linux immediately caught the fancy of many people in the computing community around the world, and it has been widely adopted.

Linux is . . . freely distributable [and was] originally developed by Linus Torvalds . . . as a student at the University of Helsinki in Finland. Linus now works for Transmeta Corporation, a startup in Santa Clara, California, and continues to maintain the Linux kernel, that is, the lowest-level core component of the operating system.

Linus released the initial version of Linux for free on the Internet, inadvertently spawning one of the largest software-development phenomena of all time. Today, Linux is authored and maintained by a group of several thousand (if not more) developers loosely collaborating across the Internet. Companies have sprung up to provide Linux support [and] to package it into easy-to-install distributions [as these packages are called] . . . . (Endnote 3)

An example of the freeware and shareware availability of much Linux software is an interesting library-automation system, koha, "the first open source integrated library system . . . . Made in New Zealand by Katipo Communications Ltd. and maintained by a team of volunteers from around the globe, the Koha system is a full catalogue, opac, circulation and acquisitions system . . . that [is] free."