Why don't we put the file characteristics into the token set that represents the file: file name, extension, size, lines, last edited, character encoding, and so on