Hi,
Yes, PL is parallel load - a logic low level on this input is what loads the shiftregister with the 8 databits. This way you know that all bits you shift in is from the "same moment in time". The Shiftin command does not handle this pin for you, you need to pulse it "manualy".

Correct, Q7 is the serial output of the shiftregister.

CE is chip enable or clock enable if you like, pull it low to use the CP pin as a positive clock input or tie the CP high and use the CE pin as a negative clock input.

Note that the '165 does not "send" the data out when you pulse the PL pin. It is the Shiftin command that "pulls" the data out of the shiftregsiter. Shiftin generates the clock pulses the '165 uses to shift the databits thru the shiftregister.