Multi-user Object-Oriented Environments

Abstract

Network-based computer environments emulating selected aspects of the physical world and allowing interaction among their human users first appeared in the late 1970s. These environments implemented a networked version of a role-playing fantasy game known as Dungeons and Dragons, hence their acronym MUD (Multi-user Dungeons). When it became obvious that many MUD users use the environment mainly to meet virtually with other users rather than to play games, new forms of MUDs and other types of emulated universes were developed and the term Virtual Environment (VE) was introduced to refer to all of them.

Since their inception, VEs evolved in several interesting directions, including environments with new types of user interfaces, for example emulating visual aspects of physical reality via VRML and other means [8], and object-oriented environments known as MOOs (MUD Object-Oriented) allowing their users to extend the environment and communicate with it via a programming language [22]. As virtual environments became better known and more sophisticated, their user populations reached hundreds of thousands. At present, VEs are used for recreation, education [16], particularly distance education, and in work, particularly for collaboration among physically separated team members [6, 26, 36, 15, 45]. There is also a large and growing body of research on existing and potential VE uses, and several meetings similar to this workshop are held every year to address both the usage and the technical aspects of VEs. Collaborative work is, of course, an area of major interest to software developers and we will thus address this issue further.

Collaboration is emerging as a major application of computer networks and the use of MOO-based virtual environments based for this purpose is gaining interest. This is because of the growing pervasiveness of computer networks, and because MOOs perfectly satisfy collaboration needs. They are persistent (thus capable of maintaining history), allow user proxies and software agents to inhabit and create separate scopes for private and group communication in both synchronous and asynchronous forms, let them posses and hold on to objects and tools, and enable navigation and porting of tools and objects from one internal scope to another. Most importantly, MOOs are extendible without limits both in terms of instantiation of existing types of entities and creation of new ones.

Because of their extendibility and persistence, MOOs can support such CSCW features as knowledge management, role-based control of access to tools and control of their usage policy, objects, or places, workflow definition, separation of projects and teams from one another, history, and others. The fact that MOOs are based on a very natural metaphor and are easy and fun to use is critical for their usability.

Besides the fact that virtual environments offer a powerful paradigm for collaboration,VEs also have a great technical interest as well. TraditionalMOOs are client-server Internet-based applications. The typical server holds the universe and implements an interpreter of MOO commands coming from the client. Commands often use a special-purpose language developed just for this purpose. Until very recently, the client was typically a Telnet-based textual interface used to enter commands interpreted on the server, and to display messages from the server. Recently, new approaches addressed the implementation of more user-friendly clients, mainly aimed at replacing the primitive text-based user interface with HTML or Java applets.