While advances in computing have empowered users to design and interact with objects in virtual three-dimensional space, little effort has been made to improve or facilitate interaction with the viewpoint. Once we begin to consider this problem, we find that it effectively spans a huge problem domain with many special cases. It touches on many of the fundamental difficulties in 3D interaction: being inside an object vs. being outside, how close is the viewpoint to the object, what is the user looking at and/or is interested in, egocentric vs. exocentric thinking, parallel vs. perspective viewing projections, multiscale and level-of-detail issues, what kind of data is being examined (abstract, incomplete, photoreal, engineering, CAD, entertainment, medical, simulation, etc.), and what is the user task (authoring, inspecting, etc.). Additional technical issues include correct handling of the clipping planes and floating-point precision problems. To help understand and address some of these issues, we have an ongoing research program to improve the state-of-the-art in 3D navigation.