We know that eventually, a realistic delay between sending orders to a probe and seeing the results will be implemented, at least on the highest difficulty. The problem here is that, with current methods of controlling the craft, performing a task like "positioning the drill on the head on the rover arm on that rock" will take about one year. Moving the arm too little means waiting a quarter of an hour to see your mistake, and waiting another quarter to see if you got it right this time. Moving too much means waiting a quarter of an hour to watch your precious rover perform a drum solo on a Martian rock.

We would need a better system than this.

What if, instead of sending instructions one at a time, we devise a number of steps for the rover to take, then send those instructions, and watch the results.

Here's how it works. We have a computer terminal at mission control. We click into it, we see a list of current missions. Let's pick the rover, from the example above. We see a simulation, a 3-D map of the rover and the immediate surroundings, including the rock we wish to penetrate. The rover is in PRESENT STATE, identical to the one the real rover is in right now. We take control of the rover, and we drive it up to the rock (Remember, this is a simulation, so there is no delay.). Once the rover is next the rock, we "save" it's new state, FUTURE STATE 1. Then, we move the arm, save it as FUTURE STATE 2. You get the idea, move the head, save, start the drill, save. We have just created a series of movements for the real rover to perform. The instructions are beamed at the speed of light to the red planet. We go off and wait for 15 minutes, check on other probes, play Gravon, do donuts in the Mars yard, whatever. Then we are alerted, the feedback from the rover is nearly here. We check up on the rover. We see the rover moving from PRESENT STATE to FUTURE STATE 1, on it's own. It doesn't need to do the same route we did in the simulation, it just needs to match the position and heading we saved in the sim. Then it moves the arm to FUTURE STATE 2, the head to FUTURE STATE 3, and it begins drilling. Hail Probe!

Why is this idea a good idea? Because we just turned a hit-and-miss, dangerous work of several hours into a super precise operation of 15 minutes. It's also realistic. These rovers are gosh-darn ROBOTS, not remote-control cars. They are capable of at least semi-autonomous behaviour.

Thank you for reading this long and dull suggestion. Continue being awesome!