OpenGL is a 3D graphics API intended for use in interactive applications. It has been designed to provide maximum access to hardware graphics capabilities, no matter at what level such capabilities are available. This efficiency stems from a flexible interface that provides direct control over fundamental operations. OpenGL does not enforce a particular method of describing 3D objects and how they should appear, but instead provides the basic means by which those objects, no matter how described, may be rendered. Because OpenGL imposes minimum structure on 3D rendering, it provides an excellent base on which to build libraries for handling structured geometric objects, no matter what the particular structures may be.
The goals of high performance, feature orthogonality, interoperability, implementability on a variety of systems, and extensibility have driven the design of OpenGL's API. We have shown the effects of these and other considerations on the presentation of rendering operations in OpenGL. The result has been a straightforward API with few special cases that should be easy to use in a variety of applications.
Future work on OpenGL is likely to center on improving implementations through optimization, and extending the API to handle new techniques and capabilities provided by graphics hardware. Likely candidates for inclusion are image processing operators, new texture mapping capabilities, and other basic geometric primitives such as spheres and cylinders. We believe that the care taken in the design of the OpenGL API will make these as well as other extensions simple, and will result in OpenGL's remaining a useful 3D graphics API for many years to come.