... I couldn't find anything on it on the internet though it's nothing over the edge.
Forward raytracing follows the natural model of light travel. The rays a cast from the light rather than from the screen. Since the resulting image is overlapping magnified versions of objects you need a 'human eye' device to get a focused image. The objects need to be built from spheres.
The reason why people don't use this method of rt is because most rays get scatered in all direction as in real life and only a small fraction falls on the capturing/exposure area.
However if you want to have multiple viewports (stereo vision, multiple users etc) on the scene this proves to be a good thing since you don't have to repeat the raytracing all over again for each viewport (each viewport does have to have it's own eye device though)
I will branch off a tread to log my work on the raytracer (to keep me motivated).