The process substitution

The process substitution is a handy way to feed the of multiple commands/processes to the input of another process. The standard way to manage a process substitution goes along with the following syntax:

>(list_of_commands)
<(list_of_commands)

Mind the space between <,>, and the parentheses; there is no space at all:

zarrelli:~$ wc -l <(ps -fj)
5 /dev/fd/63

In this example, the output of ps -fj has been given as an input to wc -l, which counted 5 lines in the output. Notice/dev/fd/63.

This is the file descriptor used by the process substitution to feed the results of the process inside the parentheses to another process. So, file descriptors in /dev/fd are used to feed data, and this is useful, especially for those commands that cannot take advantage of pipes, because they expect data to be read from a file and not fed from the standard input. A classic example of a multiprocess feed as follows:

zarrelli:~$ mkdir "test 1"
zarrelli:~$ mkdir "test 2"
zarrelli...