Ashley, I didn't answer the last part of your question. The autopilot can be used in basic mode (steer this course, maintain this height, hold this speed, climb at this speed etc.) or it can be told to follow the computer. The computer knows the aircraft performance, it knows the airline's routes and it knows the ground-based navigation aids. The aircraft has inertial navigation - you tell it where it is to start, and it keeps track of any movement from that point - which it updates automatically by searching out two ground-based aids to give a 90-degree cut in ranges. The computer can work out the wind and its effect and calculate where to start descending to meet height and speed requirements lower down.
All this leaves the pilot sitting open-mouthed, watching in amazement! He (or she) can, of course, override the computer's suggestions and revert to basic autopilot functions, or disconnect the thing completely and fly by hand. It's a great servant, but it can never be allowed to be the master.
Autoland comes in when visibility is too bad for a visual landing, but a pilot can see enough to ensure he's landing on a runway. Three autopilot channels fly the ILS, staring suspiciously at each other to confirm that they all do the same thing. At set heights above the ground it first locks the attitude, then executes a flare (raising the nose) for landing and closes the throttles. It then keeps straight along the runway on the ILS beam.
That's how it works, but autoland is only done in bad weather or for practice. Usually the pilot likes to demonstrate and hone his skill with what he hopes will be a "greaser".
End of multi-part answer!