History and Exit Codes            The Bourne Shell            Back to the introduction

---

Differences Between Unix Shells

There are many different shells that can be used in the Unix environment. Each shell has its own features, strengths, and weaknesses. The C Shell is very useful and powerful in its own right, but there are other Shells that add functionality that can make working in the Unix environment more enjoyable. In this section, we will discuss two popular Unix Shells briefly, and highlight a few of the differences between them.

The TENEX C Shell After the C Shell was introduced, there was an effort to fix many of the bugs in it. This effort grew as people began to add extra features to the C Shell, such as command line editing, TENEX-style filename completion and various other features. The end result was a new shell called tcsh, which stands for the TENEX C Shell. Despite being superior to csh the TENEX C Shell was never adopted by major Unix system vendors. As an example, Sun's Solaris Operating Environment 6 ships with csh as the default shell.

The Bourne Again Shell

With the birth of the GNU project came the beginning of bash. The GNU project was created to develop a complete Unix-like operating system out of free software. Thus, a freely available POSIX compliant shell was needed. The Bourne Again Shell was written to fill the void. It was based on the Bourne Shell language, but also "borrowed" features found in the C and TC Shells. Bash was adopted for the Linux operating system and is now one of the most popular shells in use today.

Here is a short list of useful features that can be found in the various Unix shells:

• Showing your working directory at the prompt. Although DOS does this by default, most of the earlier Unix shells do not. This function saves time as you do not have to type the pwd command to see where you are in the Unix system.
• One keystroke filename completion. Consider creating a directory with a long name such as in the command:

  % mkdir thisIsAReallyLongDirectoryName_2
  

  and then wanting to copy all C++ implementation files into it. On a shell without filename completion you would be forced to type the lengthy command:

  % cp *.cpp thisIsAReallyLongDirectoryName_2/
  

  However, with filename completion you could type:

  % cp *.cpp thisIs[COMP]
  

  where COMP is the key used for completion ([ESC] in csh, [TAB] in tcsh/bash). The shell will then complete the name of the file in the current directory whose name matches the thisIs at the beginning. All that is left to type is a carriage return. You might be surprised how much this feature can speed up file management.
• Following symbolic links invisibly. To illustrate this, create a link to the /usr/include directory in your home directory with the following command:

  % ln -s /usr/include MyInclude
  

  Switch to the C Shell by typing csh at the prompt. Now follow the link typing cd MyInclude. Print out the working directory by using the pwd command. You will be presented with output similar to the following:

  /usr/include
  

  When you followed the link you were taken directly to /usr/include. This feature of Unix systems can often be annoying. For example, a user might create a link to another directory, follow the link and work inside it. The user could then decide that he/she needs a copy of a file found at the same level as where the link was created. It might be because you must use the cd command to follow a link, but novice Unix users seem to think that a link is just like any other directory. Thus, copying said file should be the simple command cp ../filename .. However, unknown to the novice user, the working directory was /usr/include. The cp command looks for file filename in the /usr directory, fails to find the file, and complains. This traditional style of Unix linking can cause complications in software development.

  "But filename was just 'up' one directory. Shouldn't '../filename' work?"
  "No, sorry. You will have to use '../../prevost2/a1a1/filename'."

  Return to your home directory (cd ~) and exit the C Shell. Now switch to the Bourne Again Shell by typing bash at the prompt. Again, follow the MyInclude link and print out the working directory. You should observe something similar to the following:

  /prevost2/a1a1/MyInclude
  

  The link was followed by the Bourne Again Shell invisibly. The link was treated as if it were an ordinary directory. This seems more logical and as a result can make working in bash a little easier. If you follow a link in bash you only need to cd .. to return to your point of origin.

There are many other shells that have not been mentioned. If you are interested, use your favourite search engine to find information about the Korn Shell (ksh), the Z Shell (zsh), rc, and es.

---

History and Exit Codes            The Bourne Shell            Back to the introduction

---

Copyright (C) 2000 - 2002, The University of British Columbia