Haskell Ray Tracer, Part II

Part II - Geometry

Constructing new shapes

This is the chapter where typeclasses really start to shine for us, just like how classes and methods work really well in OOP languages. If you are interested, you can check this very nice paper on ad-hoc polymorphism. In short, typeclasses are useful when we want certain functions to behave differently depending on the type of the data they have to operate with. I believe that the motivation is quite natural in our case: how a ray scatters will depend both on the geometry of the object and on its composition, i.e. on its material. A light ray which collides with a sphere will have a different angle of incidence when hitting a plane. Similarly, the scattering rules for reflective materials will be quite different from the scattering rules for refractive and diffusive objects. Using typeclasses will aid us in the implementation of a natural solution with very little boilerplate. Basically, we can use a small set of functions to handle the optical path of the ray and the coloring of each pixel, with very little work when adding different materials and geometries. Hopefully this will be less abstract in the next few sections.

The remaining geometry notes are still being drafted.