Encyclopedia of Computer Science and Technology

shell 429have achieved their greatest proliferation and elaborationwith UNIX, the operating system developed by Ken Thompsonand Dennis Ritchie starting in 1969 and widely used foracademic, scientific, engineering, and Web applications.UNIX shells serve the same basic purposes as the MS-DOS shell: interactive control of the operating system andthe ability to run stored command scripts. However, theUNIX shells have considerably more complex syntax andcapabilities.Part of the design philosophy of the UNIX system wasto place the core operating system functions in the kernel(see Kernel and UNIX). This modular design meant thatUNIX, unlike most other operating systems, did not have tocommit itself to a particular form of user interface or commandprocessor. Accordingly, a number of such processors(shells) have been developed, reflecting the programmingstyle preferences of their originators.The first shell to be developed was the Bourne Shell,named for its creator, Steven R. Bourne, who developed itat Bell Labs, the original home of UNIX. The Bourne shellimplemented some basic ideas that are characteristic ofUNIX: the ability to redirect input and output to and fromfiles, devices or other sources (using the < and > characters),and the ability to use “pipes” (the | character) to connectthe output of one command to the input of another.The next major development was the C shell (csh). TheBourne shell used a relatively simple and clean syntaxdevised by its creator. As the name suggests, the C shell(developed at the University of California, Berkeley) takesits syntax from the C programming language, which wasby far the most commonly used language on UNIX systems.One logical reason for this choice was that C programmerscould quickly learn to write scripts with the C shell. TheC shell also added support for job control (that is, movingprocesses between foreground and background operation)and in general was easier to use for interactively controllingprograms from the command line.UNIX users sometimes used both shells, since the simplerand more consistent syntax of the Bourne shell is generallythought to be better for writing scripts. (The twoshells also reflected the split in the UNIX world betweenthe version of the operating system provided by AT&T andthe variant developed at UC Berkeley.)David Korn at AT&T then decided to combine the bestfeatures of both shells. His Korn shell (Ksh) kept the betterscripting language features from the Bourne shell but addedjob-control and other features from the C shell. He alsoadded the programming language concept of functions (seeprocedures and functions), allowing for cleaner organizationof code.Another popular shell, BASH (Bourne Again Shell) wasdeveloped by the Free Software Foundation for GNU, anopen-source version of UNIX. BASH and Ksh share mostfeatures and both are compatible with POSIX, a standardspecification for connecting programs to the UNIX operatingsystem.The surge of interest in open-source UNIX in recentyears (see Linux) has brought a new generation of shellusers. Although modern Linux distributions provide a fullgraphical user interface (GUI)—indeed, a choice of them—such tasks as software installation and configuration ofteninvolve entering shell commands. Experienced users canalso find, copy, move, or otherwise manipulate batches offiles more efficiently in the shell than in using windows andmouse movements.Shell ScriptsRegardless of the version of the shell used, shell scriptswork in the same basic way. A shell script is a text filecontaining commands to the shell. The commands can usecontrol statements (see branching statements and loop)and invoke both the shell’s internal features and the manyhundreds of utility programs that are available on UNIXsystems (see scripting languages).Once the script is written, there are two ways to executeit. One way is to type the name of the shell at the commandprompt, followed by the name of the script file, as in:$ sh MyScriptAlternatively, the chmod (change mode) command can beused to mark the script’s file type as executable, and the firstline of the script then contains a statement that invokes theshell, which will parse the rest of the script. The script cannow be executed simply by typing its name at the commandprompt (or it can be included as a command in anotherscript).Here is a simple example of a shell script that prints outvarious items of information about the user and the currentsession on a UNIX system:#! /sbin/shecho My username: `whoami`echo My current directory: `pwd`echoecho My disk usage:du -kechoecho System status:uptimeif test -f log.txt; thencat log.txtelse echo Log file not foundfiThe first line tells UNIX which shell to use to interpretthe script (in this case the Bourne shell, sh, will beexecuted). The echo command simply outputs the text thatfollows it to the screen. “whoami” is a UNIX command thatprints the user’s name. The script takes advantage of aninteresting UNIX feature: The whoami command is put in“backquotes” (` `). This inserts the output of the whoamicommand (the user name) in place of that command, andthe resulting text is output by the echo command.The du command gives the user’s disk usage, while theuptime command gives some statistics about how manyusers are on the system and how long the system has beenrunning. Finally, the if statement at the end of the script