I'd like to rewrite the ray-tracer using Haskell. I'm currently taking a graphics course where I'll have to write a ray-tracer in c++, but thats not different enough. Due to the fact that a Haskell raytracer will be useless, this will be most definitely open source.
If you haven't seen it already, this guy has an interesting pure approach to writing a raytracer in haskell:
ReplyDeletehttp://sebastiansylvan.wordpress.com/2009/07/04/ray-tracing-signed-distance-functions/
The basic idea is: every object is represented as a function from a vector to a number, the number being the smallest distance to the nearest object. To combine objects for a scene, you just compute the minimum!
Also, since you like language stuff, you should check out pages 278-280 of this book:
ReplyDeletehttp://www.itu.dk/~sestoft/pebook/pebook.html
which outlines how partial evaluation can be used to speed up a raytracer. I can't find the Mogensen paper, but Anderson did a similar paper:
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.119.227
Dude, wow, I can't believe you actually found my other blog. I haven't checked this in ages.
ReplyDelete