Design Specification

	This is the design specification of Pane, a globally illumination 
raytracer.

	Pane will be written in C++ using the object-oriented paradigm. 
It will also use Open Inventor, which uses Open GL. I intend to use STL 
for data structures and Henrik Jensen's photon map class.

Flow of data

     Scene file (.iv file) + command line options (using getopt) -> pane
     pane reads scene into internal format
         (using an octree for fast ray-triangle intersection)
     pane does any lighting
         (using photon map)
     pane outputs the scene to an iv file (using OIV), or an image
         (using internal algorithm)

Code Reuse Policy

	I will reuse as much code as I want. This is necessary to meet 
the quick pace of the project. I am confident that enough code will be 
mine to satisfy authorship of the program. Of course, reused code will be 
cited. To be clear, Josh Grant has already written a program similar to 
pane and implemented many of the parts necessary already, and has made
his source available. I intend to use his source as a reference and for
key parts, such as OIV callbacks for triangle acquisition.

Phase 1	 

	"By the first day the user should be able to read in a scene file 
and create an unlit image of the scene."

	I will use OIV to read in a scene file, and register a callback 
to store the triangles into an internal data structure. This data 
structure will just be a class with a linear list of triangles (and 
properties).

	Next I will hard code a camera and raytrace the scene. For each
pixel I will do ray-triangle intersection with the scene and if there is 
a collision, use the angle of incidence as the shading factor. If I am 
ambitious I will also include the color. Then I will use this as the 
color of the pixel. After I find a color for each pixel, I will write out 
the image using the PPM format.

	The justification for using incident angle as the shading factor 
is that ??????? figure out later.

   	Pane will be used in the following fashion:
	
	pane -i scene.iv -o image.ppm


Phase 2

	I will replace the linear data structure with an octree and use 
recursive raytracing, with shadow tests (see if light sources - emissive 
triangles - are around).

Phase 3

	I will shoot photons from the light into the scene and store them 
in a photon map. Then I will balance the photon map and use it for 
ray-tracing using the algorithm in Jensen's book.

Phase 4

	I will use the photon map generated lighting information to 
modify colors in the scene, and write the new scene out to an iv file.
Also I will figure out how to have an intelligent and configurable 
camera.

Further Phases

	Transparency? Image output format options (jpeg?). Lens 
distortion? Texture mapping?