See perldoc -q sort. Use capturing parens to extract the different chunks and compare the corresponding chunks. Use a Schwartzian transform so you don't wind up pulling items apart multiple times.