Ok, this is going to be a little hard to explain.
This kind of shot is called stereoscopic shot.
Basically, each one of the images was obtained dislodging the camera by a couple of inches.
The result you get is two almost identical images.
The thing is, each one of them corresponds almost exactly to the image that each eye receives when looking to the same scene.
To visualize the 3d effect, you must cross your eyes (it's a bit hard, but it becomes easier with practice) until the two images merge into a central one. Try to still your eyes, and they will focus on this image.
The result is quite astonishing
Good luck!
they refuse to do what you say
Take a look at my 3D pics to see what I mean [link]
While making something meant for parrarel view, you have to keep a composition in which a very limited amount of objects would be invisible in one of the photos. And it's practicaly impossible to do while photographing a city.
To picture what I mean, look at this amazing example I found while looking on 3dphoto.net forums:
[link]
See? The objects don't hide, they are showed from different perspective. And because of that, your mind can process it much more seamlessly.
Shit happens when you think about three things at the time