Abstract: Direct interactive manipulation of 3D objects is a highly desirable but not yet fully realized feature that can make the use of a modeling system more intuitive and convenient. Two kinds of manipulation may be identified: positioning and sculpting. Positioning refers to the placement of objects in relation to each other within a common scene. Sculpting refers to the arbitrary deformation of object shapes. Ideally, we would like to be able to move objects around easily as if they were held in our hands and to reshape them freely as if they were made of clay. However, realizing these goals of 3D editing on a computer is very difficult, especially with an ordinary display monitor. A fundamental problem is to provide sufficient depth and shape cues for the user to perceive the rendered objects and editing cursor in such a way that positioning and sculpting operations can be performed with reasonable flexibility and accuracy. Traditional CAD packages avoid true 3D editing by restricting the controlled movement of the editing cursor to two-dimensional. An arbitrary 3D location or displacement has to be specified in consecutive steps using separately projected views of the geometry. It is conceivably more natural and productive to work in a single display window. However, the challenge is to include in the display meaningful and distinguishable visual cues beyond such standard features as hidden surface removal, surface lighting, and perspective projection.