<< wdiff >> From man page of wdiff: wdiff compares two files, finding which words have been deleted or added to old_file to get new_file. A word is anything between whitespace. The output is collected and used to produce an annotated copy of new_file on standard output. Suitable annotations produce a nice display of word differences between the original files. Example: text-a I like FreeBSD. text-b I love FreeBSD. % wdiff -n text-a text-b I [-like-] {+love+} FreeBSD.