Focal Length + Framing = Perspective
If the Focal Length is the variable, and Framing is the constant, then Perspective changes. Your link is a good example of this.
I remember recently looking at a PDF that I think was made by Canon showing the same phenomenon with a woman standing next to a bridge in front of some buildings (I think). Can't remember where I found the link though....
Sigma has something like that on their website. "Focal Length Finder" or something to that end. Same shot, but showing the framing of it at any focal length you chose. In their example the shots would be taken from the same position... so as your focal length gets shorter you obviously take more of the scene in the framing, as your focal length gets longer you get less in.
The example in the link was quite good, but the movement of the subject and/or photographer as he got to the widest FL's showed extra distortion.
The same example would likely be best taken with a mannequin head set up on a stand at the same height as the camera's position on a tripod (to coincide with the center of the mannequin head to be the center of the lens- both laterally and longitudinally). So, as your move the focal plane (via moving the camera on the tripod, but not raising or lowering it) to compensate for the change in focal length (via zooming or changing lenses) in order to achieve a comparable framing (as defined by placement in the frame of the facial features... i.e. the eyes should be in the same place throughout the series) you still have the mannequin head's face aligned at a 90degree angle to the center of the lens. If you notice the 24mm example the forehead begins to look much larger than the chin because the camera is no longer at the same angle to the face as it was at the other FL's.



