report.txt

Report--Ray Tracing in One Weekend, Haskell Adaptation

My implementation for this project reached the "Medium" stage of my proposal.
I implemented ray tracing for Diffuse Materials, Dielectrics, and Metals. I also included a command-line interface to generate particular scenes.
My code is structured and based on the work in the book Ray Tracing in One Weekend, whose code was originally in C++.
I found it enjoyable and, at times, quite challenging to interpret certain C++ paradigms into a Haskell context.
I would say the most notable example of this was in shared pointers, where instead I used an existential type to wrap different potential datatypes.
I also found it interesting how I had to thread StdGens across different functions, always being careful to return a new one if I applied a random function within a random function.
That is to say, I had to store and return the StdGen returned by random functions called within random functions because otherwise, the value returned would be stale.

My code is organized such that the file names more or less correspond to what aspect of the implementation they pertain to.
A notable exception to this is Utils, where at least one major typeclass (Material) is defined--this decision was made to avoid circular dependencies.
The code in RayTracing.hs relies on all of the other files to create a raytraced PPM image.
Thus, to try the project, run RayTracing.hs with --scene <demo, glass-demo, or metals>.