1) If your nested comp is Continuously Rasterized, then the camera in its precomp will be neglected. You can still use a camera in the precomp as will allow you to have a better idea where the layers are in 3D space. IOW, the camera is still useful in the precomp for placing your layers in 3D space.
2) In the precomp, mask out the areas of any layer that extend over the composition boundary. This ensures that your nested comp doesn’t carry over any pixels into the ‘final comp’ that are outside of the precomp’s comp boundary
3) Precomps are used to increase creative opportunities, to overcome technical limitations in AE and for grouping a set of layers to simplify a ‘final comp’.
4) You can and should scale your nested comps so that they fit the area that they are supposed to be placed into. You should of course ensure that their 3D position values are correct, relative to the filmstrip. The workflow to make this happen requires (1) placing the nested comp in accurate 3D space relative to the filmstrip (2) scale the nested comp to fit into its required position. When scaling, take note that the Anchor Point acts the center for scale.
You should have sufficient info to get going. If you hit any obstacles, study the issue and try to understand the cause. Then make corrective/alternative measures.
HTH
RoRK
Latest AE Workshop – MoGraph Intensity – Shapes & Text
Intensive mocha & AE Training in Singapore and Other Dangerous Locations
Imagineer Systems (mocha) Certified Instructor
& Adobe After Effects CS6 ACE/ACI