In order to display three dimensional models on two dimensional drawings, sheets, or a display device, it is necessary to determine what geometry is visible and then to display the visible geometry by projecting it on a two dimensional surface. There are two distinct approaches to determine the visibility. The first one is the raster or pixel oriented approach that uses a depth buffer, commonly referred to as z-buffer to determine the currently visible geometry at each pixel on the screen. The raster display is quite fast, particularly when implemented in specialized graphics hardware, and can be used to dynamically display the 3D geometry in real time. This approach is also referred to as Dynamic display.

Although the raster display technique is extremely efficient, there are limitations to it. The primary limitation is the accuracy; a raster display determines the visibility at a fixed and limited resolution that may not be sufficient for annotation and documentation purposes. The more accurate alternative to the raster display is explicitly determining the visibility of geometry edges. These edges can also be displayed at a higher resolution or plotted with pen orientated devices. However, this approach is far less efficient, as it involves making comparisons between the geometry that can be dimensioned. Since the calculation of explicit visible edges for large complex models can take significant amount of time, this process must be performed at discrete times initiated by the user, rather than being generated dynamically as the model changes.