As far as I understand it, for a character to be common to all words being scanned, it has to occur at least once in each word. If @ARGV contains 3 words to be scanned, for example, then we are only interested in characters than occur a minimum of once per word for all three words, or three times. That's why when a character is found in a given word for the first time, it's "score" is incremented by 1 to bring the cumulative score to the word number +1 (since arrays start at 0). Other occurences of the same character in the same word are effectively ignored since we don't care. Only characters that score 3 are returned. I think this theory does satisfy the original requirements actually, even if the code itself is buggy...