Divide the inventory program into two parts in order to separate the initialization and processing part from the user I$/$O part. Use the command line to supply the inventory file name to the initialization$/$processing part of the program. The initialization$/$processing part then reads the file to populate the data and then closes the file which completes the initialization. The initialization$/$processing part then creates two pipes. After this it forks to create another process and with the second process it executes the I$/$O part. The I$/$O part performs all interactions with the user. You will need to create protocols on the pipes between the processing part and the I$/$O part. For instance, on a $$find record$$ the I$/$O part might first write an integer, say $3,$ to communicate to the processing part that it needs to find a record. After that the I$/$O places information on the pipe that is necessary for the processing part to perform the find. The processing part finds the record and sends the data to the I$/$O part $($maybe sending the record index too$).$ The I$/$O part would than display the record and ask the user something like $$what next?$$ This continues until the user chooses $$save and exit$$ or $$exit no save$.$ When an exit message is sent, sleep the I$/$O process for $10$ seconds to $$ensure$$ the message is received before terminating.