Actually, it is a collection of images for different zoom level.For example, the lowest zoom level uses one image and the next one uses two, and then four (from

Actually, it is a collection of images for different zoom level.For example, the lowest zoom level uses one image and the next one uses two, and then four (from what I remembered). You can find this out in the google map API. I imagine vector calculation is just unnecessarily complicated when simply storing the images works well enough. In that case, panning and zooming seems easier to implement, as you are probably just moving the images around and reloading different ones when you zoom. jquery comes to mind for these tasks. Maybe it will take some work to make it run as smoothly as Google does it, maybe not.