Case description:
The user sees a 2D-map that has limited size. He can drag this map. He can also zoom in/zoom out, like in Google Maps (the mouse pointer remains on the same coordinate, both in map/virtual world and on the user's screen).
So, when the user drags the camera to one of borders and zooms out - the visible world shrinks, and there appears a visible gap between the map's boundaries and the view's boundaries. (Imagine that Earth is plain, it's placed on a turtle, and Google implements Google Maps.)
What's the optimal way to handle zooming at the map's boundaries?