Ceiling hosted lights, location assumptions by Revit

[gordonp147484]

What governs the location of a ceiling hosted light before the user discretely locates it? We are having some problems where a wall will get moved by a foot, and a light will fly off by 6'. There is a move afoot here to abandon ceiling hosted everything because of this, but that feels like throwing the baby out with the bath water. I am wondering if there is some hidden relationships there, ala Sketch Dimensions, that we just can't see? I would rather understand what Revit is trying to do, and if Revit is wrong 75% of the time then call it a bug and see it fixed. But if Revit is right 75% of the time, and we are throwing that away to address the other 25%, I am not sure it is a good move.
Some comment from the Factory would be very welcome here. Something along the lines of "We intend the behavior to be this and that, and imagine the workflow would look like such and such."

Thanks,
Gordon

[patricks]

hmm I don't think I've experienced the kind of behavior you're describing.

I just started a new project, drew 4 walls, placed a ceiling in it (automatic sketch) and then placed some lights in the ceiling. Then I moved each of the 4 walls around, and while the edges of the ceiling stretched with the wall, the grid placement itself did not move at all, and because of that, neither did the lights move.

I know that if you move the grid in the ceiling, lights will move with it. Wall movement should not affect the grid or the lights. The only thing that did affect my lights was if I pulled one wall in to make the ceiling so small that some of the lights ended up floating in space. The lights still did not move, but I got an error saying the light instances would be deleted.

[sfaust]

I have had instances where I edited the sketch of a ceiling and deleted lines associated with a wall and redrew new ones, and that made the lights go all wacky, but I've never had lights move just by moving a wall (unless of course it was constrained to the wall).