That's absolutely right but I was talking about wide sceneries and make  groups or parts of them invisible  if not viewed by the camera; managing the visibility function in the sequencer tab you can virtually create animations of vast lands puzzle-like (i.e. Aerial views), without constrain C to work in vain and as consequence more slowly. So if you create many little ecosystems you can theoretically merge, mount and dismount  together sequentially