Some initial concepts

• shells: bash

  (GUBI)

  • what is it?
    • text front end alternative to GUI
    • program runner
• resources $\rightarrow$ features (ALAN) There are some extra concepts that are important to understand how Linux works, and to effectively use the shell.

  A single directory tree: no device names

  In the Windows world each device has a letter identification. So it is common to have the main hard disk under C:, the cd-rom driver under D: and so on. The complete address of a file is something like ``C:My Documents
  Research
  thesis.doc''. Linux takes a different approach. There are no device letters, in fact the user is unaware of the actual location of his/her files. All directories and files are under a single root directory. The structure of the directory tree is controlled by the administrator.The complete address of any file at the root directory, for example: ``/home/employee1/Research/thesis.doc''. This has many advantages. In particular, users can be spared of any changes in the system's device. for example, if a project turns out to use lots of disk space, we can buy a new hard disk and assign to this new disk the home directory of the project, copying all the old information into the new disk.

  The PATH environment variable: finding programs

  As in Windows, programs can be installed in any directory in Linux. However, in Windows programs not installed in specific folders can only be called if you locate the folder and click in the program icon. In the linux bash shell, you can tell the system which directories can contain programs. When the user type a command the system automatically searches in the directories specified. The places where bash looks for programs are stored in the PATH. In all linux installation some directories are already part of the path. These are generally /usr/bin and /usr/local/bin. Almost all standard Linux applications are located in these two directories. However, any user can add new directories to his/her path and run other programs without having to search for them. You can even download a new version of a program in the system and run your version by default. You only have to include the directory where you stored the program in the path BEFORE the system directories.

  auto completion

  Another nice feature of the bash shell is auto-completion. Is is normally very tiresome to type complete file addresses or even program names. The bash shell has a special feature called ``auto completion'': when the user starts typing and presses the <TAB> key, the systems tries to complete the typing. If it is a program name (first word to be typed), bash searches the directories in the path to find which programs start with the letters typed. Otherwise bash looks in the partial path specified by the user to search for completions. If there is only one completion, pressing the tab key will cause the appropriate word to appear, if there is more than one completion, a sound will be emitted. Pressing <TAB> the second time will cause all possible completions to appear.

  job control: fg, bg, kill

  When using command line, it is important to understand the concept of ``Process''. A process is a program running. There can be many processes running the same program (for example, when you click a the Windows explorer icon in Windows). A processes can be in two states: running and suspended. When a process is suspended no work is performed and no work is lost. A running process just executes the program tasks normally. Processes can be killed, that is, the user can determine that the processing will end (this is equivalent to click in the X icon of a window).

  history

  The shell maintains a list of all the commands that the user has typed. This list is called the ``history''.

  data flow (STDIN, STDOUT, PIPES)

  Linux programs, by default, get input from the keyboard and output data to the screen. The keyboard is the Standard Input, or STDIN, and the screen is the Standard Output, or STDOUT. All basic shell commands use this principle. A nice feature of command line interactions is that you can have the output of a program in STDOUT be fed directly to the STDIN of another program, creating a unix Pipe.. For example, suppose you have a program that reads words and outputs synonyms for those words, and a program that reads words and counts the total number of words read. If you connect the two programs together, you can type words, and have a final count of the total number of synonyms. HORRIVEL!!!

  protection - chmod

  Falar direto no learning by doing.

  wrap up