A Geometric Framework for Computer Graphics

that the output is a raster, or pixel-based, device and that the onus
for shadow computation lies on the framework user rather than
on its designer. These assumptions resulted in a wealth of simple
graphics algorithms and libraries, but one problem in graphics,
shadow computation, tenaciously resists attacks based on raster
approaches.

The talk will describe a framework for geometric computing in graphics
that makes it possible to address shadow computation from point,
linear, and area light sources. These light sources are handled
properly, in the sense that the framework is aware of the change in
shadows that result from the insertion of each light source.

In addition to the versions of these problems in space, a description
for the planar case will also be fully given. The versions in the
plane both provide valuable intuition before tackling the problems in
space and are useful for many restricted graphics environments.