[Haskell-cafe] Computer Graphics and Haskell - Radiosity Methods

[Maarten Hazewinkel]

> 2010/3/1 Hector Guilarte <hectorg87 at gmail.com>:
>> Hello cafe,
>> While I was studying for my computer graphics test I have tomorrow I
>> realized that maybe some of the major problems I've read so far about
>> Radiosity Rendering Algorithms may be reduced significantly if it was
>> implemented in Haskell and taking advantage of the lazy evaluation so that
>> only what can be seen from the viewer's perspective point of view is
>> calculated, and the rest of the scene just remains as thunks waiting for
>> them to be calculated in case they are needed.

The way radiosity works, those invisible parts of the scene can still
add illumination to the visible parts.
So the first time you query the radiosity data for any part of the scene,
you'll end up forcing the calculation of the entire radiosity solution.

That's basically the difference between plain raytracing, which is lazy in
that way and works backwards from the viewer's perspective, and radiosity.


Maarten