Create two programs: one program that reads the data from <>, splits the data into words, and then updates a DBM file noting the number of occurrences of each word; and another program to open the DBM file and display the results sorted by descending count. Run the first program on a few files and see if the second program picks up the proper counts.