Interaction and Behavior in Web - Based Shared Virtual Environments

VRML has already become the standard for the exchange of static 3D scenes on the Web. However, several approaches were recently made to extend it by behaviors and interactions in order to support interactive 3D applications on the Internet. Most of these approaches including the current VRML 2.0 draft specification apply to single, isolated users, browsing through virtual worlds. But already in the near future, large numbers of users will share a single virtual space and participate at virtual events. To support this kind of applications, a more flexible approach is required. In this paper we will present our approach to support interactions of multiple users as well as behaviors in shared virtual environments , based on VRML by using a flexible and distribut-able event and message passing model. 1. Introduction The first step to support virtual reality on the Internet was the introduction of VRML in 1995. VRML 1.0 [1] is a Web-based static description language for 3D scenes, several approaches exist to extend it by behaviors and multimedia features. These approaches lead to the new VRML 2.0 specification. However, even this new standard will be closely related to the World Wide Web, since most VRML files will still be accessed from Web pages. Additionally, VRML files are transferred by a HTTPD server using the HTTP protocol, similar to ordinary Web pages. In the future, local virtual environments will not be sufficient for the growing Internet community. New application areas spanning over all areas of the individual's life will be incorporated into virtual spaces distributed world wide over the Inter-net. To provide an adequate interface for this kind of applications , support for distributed shared virtual environments has to be realized. The obvious choice for a platform-independent solution is to enhance existing standards such as VRML by adding appropriate new features. In this paper we will present our approach to realize distributed virtual worlds, based on our flexible behavior and message passing mechanism. This paper is based on the The Power of Dynamic Worlds proposal [4] for VRML 2.0, which was officially submitted to the VAG (VRML architecture group) along with five other contributions. In the first section of this paper we will compare three of these proposals: Moving Worlds (the current VRML 2.0 draft specifi