ZERO physical 'fiddling' for aircraft. Primary inputs for an air target such as range, azimuth, elevation angle and closing rate all in analog form were provided by the fire control radar locked on the target. For air targets the computer was in 'automatic mode' and the operators could observe and check the dials and indicators for anything that looked abnormal.
Computers function was to plot a firing solution with continuous updates telling the gun mount where to point the gun to have the shell intersect the point where the aircraft would be when the shell reached it. In addition the computer sent a 'signal' to the ammo loader to set either a time delay detonation fuse (based on calculated time of flight) or, fuse "quick" (detonate on impact). If there was some 'error' occuring the weapons control officer could call for a 'spot' adjustment and one of the hand cranks would be turned to refine the solution.
The hand inputs were of PRIMARY value during shore bombardment. With nothing for a radar to lock on (sometimes used a beacon) the computer was on its own. The first firing solution was based upon range and azimuth to target using military map coordinates. Then a 'spot round' would be fired. Using radio communications, a 'spot' would be called back to the ship in yards of azimuth or range. You used the "SPOT" knobs to do that maintaining the initial firing point as your 'base' solution. It made the response quicker because the 'computer' didn't have to resolve the entire equation.
Once the people calling for the fire mission were "pleased", all hell broke loose with the comand "FIRE for effect". Which meant don't stop sending them till we tell you too. We had some big brusser loaders in one of our gun mounts that could hand load and fire a total of over 40 (20 form each barrell) 5" rounds a minute. Serious hell on earth.
The computer could actually have two different guns firing different shells at the same time to achieve different objectives over the same target using the knobs.
Sounds like a simple task but in addition to that, the computer was constantly adjusting for the motion of the ship in speed, changing direction of travel and with the gyro, role. Accuracy? Ours was pretty accurate. A 'typical spot' was usually less then 40 yards with targets as far as 10 miles away (with obviously some error from the initial coordinate called in).
"Typically" there were 3 or 4 enlisted men operating the computer either 'hands on' or keeping an eye on the dials and outputs to make sure the computer was solving the problem correctly. Tell tails on some dials indicated whether things were going well or out of whack. The operators would CONFIRM before any firing that the machine was showing a 'good' solution. In addition, an officer would be in the room to over see things. This was a job for a junior officer (ensign or Lt JG) in the weapons department and the enlisted guys knew far more about what was happening.
I suspect Gordon Jones knows a heck of a lot more about it then I do. It wasn't my primary piece of gear or responsibility.