I want to try and explain the reasoning behind this.

The 2D image maps are mapped to the "inner wall", so to speak, of a 3D sphere. So, rather than looking at a 3D object, you are "inside" the object, looking outward towards that inner skin. While all of the imagery is based on high-res 2D images, the underlying technology does indeed use a 3D engine.

So, rather than thinking of simply zooming into an out of a single image, you're really zooming into and out of images that have all been stitched together into a 3D environment of sorts, with breakthrough image mapping and projection, to create the environment without any distortion.

From Jonathan Fay's own words, WorldWide Telescope was based on his work from 3D terra-server visualizer engines. You aren't simply panning and zooming into one giant flat image.

I hope that helps. This was a tough one that, again, we knew would upset some people, but we had to go with what we believed to be the purest result with the data at hand.