What I do to expand the ports is to use 4 output ports to control a 74hc138 decoder and it will select each of 74HC373 data latches. Now you can have a total of 64 outputs off of 4 processor outputs. So you will run your data bus, say 8 data output ports to the D inputs of all the '373s and you can latch them up individually with unique data for each one by selecting one at a time through the '138 "one of eight decoder" chip by attaching Y0 through Y7 to "OC" output control of each '373.