Bad Physics: Errant dispersion images are normal

I mean "normal" in both the common, regular sense and in the perpendicular, physics sense.

Attempts to replicate the classic prism image from Pink Floyd's Dark Side of the Moon typically fall victim to the error illustrated by the image above. The primary error is to show the incident ray and the refracted ray as normal (at right angles) to the air-glass boundary. It just seems like the rays should do that.

Add to that the fact that the prism has been replaced by a tetrahedron. And why are there divisions in the dispersed color bands? Absorption spectra? What's that glass tetrahedron made of?

Here's the Storm Thorgerson original: