Gandalf3 has done a good job of answering your questions, but I'll share a work-around I developed for the challenges the second question posed to me: I began to use "helper geometry". Helper geometry consists of objects I use as an aid to making a particular model. These objects include place markers, label holders, and modeling tools. Place markers and label holders are self explanatory; an example of a modeling tool would be in the circumstance where I want two objects to be a designated distance apart, say 8.235 blender units, but the distance is not parallel to the x, the y, or the z axis. In this instance I might create a plane mesh that is 8.235 blender units in one dimension, place it so that one edge of the plane is at the first point, and orient the plane in the direction I want to measure the distance, and then locate the other object at the edge opposite the first.
To help me remember the locations of objects in the grid, I'd create an object of suitable size. In the case of an object to remember the location of camera 1, it would probably be a cylinder of very small diameter, named "camera1location". As long as I don't apply the location transformation to helper geometry, I can read the co-ordinates of the item in the properties window, object tab.
To keep my modeling projects orderly, I place all cameras and lamps in layer 1; all helper geometry in layer 2, and do my modeling in layers 3 through 20.