It is useful in understanding the artistic nature of the medium of interactive virtual worlds to first understand the key technical means that it relies on and how these have reached a stage of development to support a new medium and the new expressive language of that medium.
The fundamental basis of the medium is a rendering infrastructure, where rendering is used to mean 'realizing' the declarative description of a 'score' consisting of a composition of objects and streams from various media types such as geometry, images, vector graphics and video, sound, music and speech, and possibly others as well. In addition there must be a means of navigating a camera viewpoint in the world, the 'world' being the sum of all possible experiences on the virtual stage specified by the score and consisting of light and animation and sound. The main technical requirement of the medium is that there be sufficient rendering capability present relative to the specified complexity of the world to permit a rate of state update able to give a convincing illusion of 3D, continuous motion and instantaneous response. The rapid development of better and better hardware, especially for polygon and pixel rendering, means that more and more complex worlds can be designed which maintain an adequate rendering rate.
The other fundamental technical component of the medium is a programmatic infrastructure in which to express intentions of will, reactions, occurrences and behaviors, and changes of view and context. Given the sophistication of computer languages and the increasing power of computer processors, this component provides the means to compose and utilize interactivity and behavior and intelligent control to any degree needed, for example, to create simulations of living things and 'believable' characters and narratives. This aspect of the medium differentiates it from either cinema or simple computer animation.
Finally, the venue for this medium is most naturally the emerging Web based on high-speed high-bandwidth networks. The speed and capacity of these networks are vital in two ways. First, the network can be used as a means of accessing an essentially boundless repository of media assets and production data and materials, which allows worlds to be able to give an illusion of boundlessness in space and in time since they can be dynamically extended as much as required. Raw materials for rendering can be shifted in and out of various network caches so as to be directly at hand when needed. If the speed and bandwidth are high it is easy to provide a continuously changing subset of scenes and media events from a large superset which makes up the full contents of the world and its narrative. Second, networks provide the means to support multi-user 'shared' worlds. The high speed and bandwidth contribute to make message delay minimal, which is necessary to maintain the sense of 'believably instantaneous' shared presence.
Thus the development of fast graphic renderers and computer processors, and the emergence of high-speed networks, have provided the technical means to support a new medium and to enable the creation of the first generation of pieces which use the new unique language of the medium.