The Windows-Icons-Menus-Pointer (WIMP) interface paradigm dominates modern computing systems. Yet these interaction techniques were originally developed for machines that are now 10, 15, or nearly 20 years old. Human-computer interaction currently faces the challenge of getting past this "WIMP plateau" and introducing new techniques which take advantage of the capabilities of today's computing systems and which more effectively match human capabilities. Two-handed spatial interaction techniques form one possible candidate for the post-WIMP interface in application areas such as scientific visualization, computer aided design, and medical applications. The literature offers many examples of point design, offering only a description of the thing (what the artifact is) and not the process. But point design only provides a hit-or-miss coverage of the design space and does not tie the multiplicity of efforts into a common understanding of fundamental issues. To get past the WIMP plateau, we need to understand the nature of human-computer interaction as well as the underlying human capabilities.