VRML 2.0 is a scene description language which is human readable. There is currently no ratified API associated with VRML. The language has the following capabilities:
- Built in geometric primitives including face sets and solids
- Lighting, material, texure, movie control
- Spatialised sound
- Absolute time for animations
- The concept of an avatar to allow collision detection
- Hyperlinking, viewpoints and navigation methods
- The ability to extend the language through prototyping
- Event handling to detect when geometry is interacted with in some way
- Routing of events to allow one object to affect another
- Scripting in various languages which allows logic in the world
VRML files describe a scene graph structure which is parsed by a browser. The browser may be a plug-in for a web browser or a helper application.The scene graph is composed of nodes which fall into two categories, grouping nodes and others. These nodes may have routes between them, above and beyond the scene graph hierarchy, which define the possible interactions of one node with another. Nodes have fields which define what actual values the node has associated with it.
Within this scene graph certain special nodes, known as bindable, have stacks associated with them which allow them to be pushed and popped to control certain world parameters.
Confused? Don't be, all this is explained in the concepts sections. The bottom line is this, it has never been easier to write 3D graphics.