Implementing Geometric Algorithms Robustly

This note is meant as a sequel to Steven Fortune's note on ‘Robustness issues in geometric algorithms’ [For96] in these proceedings. We revisit some of the issues raised by him, such as the consistency between the combinatorial and numerical data in geometric algorithms, and then we elaborate on a number of additional topics, including issues in proving correct geometric algorithms meant to be executed with imprecise primitives, and in the rounding of geometric structures so that all their features are exactly representable.