I am told that is possible to locate lights that the simulator should know about simply by referencing named 3D model objects. (I have never reached a stage in my own add-on attempts yet that I would have started bothering with lights, sadly.)
In general, it would be extremely cool if the simulator in the future would grab more information about locations and dimensions of various things from the 3D model. Either by using fixed names for objects in the 3D model, or, preferrably, by giving the names in a configuration file.
This would mean that in some cases you wouldn’t have to keep the same information in multiple places. Or at least that you could locate something in a much more natural way, by placing it in the 3D model, seeing at the same time how it relates to other parts of the aircraft, than by estimating numeric coordinates and typing those into a file or the Aircraft Editor.
A very obvious example is the cameras in cameras.cfg. It is fairly silly to have to tell the camera location in a cfg file when you could simply place an empty object in the 3D model for each camera, at the desired location and the pointing in the desired direction. The cfg file would then just specify the name of that empty object.
Also, for instance, for fuel tanks, you could just create a 3d object in your modeling software for each fuel tank, making sure it matches (to some reasonable degree of accuracy) that in the real aircraft, and the sim could then calculate its volume, and could also automatically calculate the contribution of each fuel tank to the aircraft’s centre of gravity and moments of inertia. The objects in question would not need to be rendered, you could mark them as using a special kind of material. Just like the occluder objects.
Ideally, you could even tell the simulator about the aerodynamic surfaces like this. You would add simplified shapes representing each wing, canard, fin, tailplane, movable control surface, etc, as many as the aircraft has, to your 3D model and the sim would work out the rest. The sim would check from the 3D objects what the area, sweep angle, incidence, anhedral, maybe even twist, of each wing is.
But yeah, that would probably require quite strict rules on how such aerodynamic objects in the 3D model need to be constructed. Probably just as simple boxes from which only general shape, incidence and anhedral could be deduced, and you would still need to enter a potential twist of a wing. And the rotation angles of the moving control surfaces. And again, the objects would not be the ones actually rendered, you want to use more details for them, obviously.