The MR Toolkit Peer Package is an extension to the MR Toolkit that allows multiple independent MR Toolkit applications to communicate with one another across the Internet. The master process of an MR Toolkit application can transmit device data to other remote applications, and receive device data from remote applications. Application-speci c data can also be shared between independent applications. Nominally, a n y n umber ofpeersmay communicate together in order to run a multi-processing application, and peers can join or leave the collaborative application at any time. This paper will introduce the peer package, and will explain the theory of its operation. The last part of the paper will discuss our experience with a demo program we h a ve written called multi-player handball that uses the peer package. I n troductionOne of the main research thrusts in the study of virtual environments is to examine how many people can work together on a common virtual application. The basic idea is to allow people who are geographically dispersed to collaborate on some task in one common virtual space. Software that enables this kind of interaction must support both adequate single-person VR style interaction, and must support communications to other computers. The presence or absence of other users in the simulation should have no e ect on the visual presentation, and should not induce undue lags or delays in the interaction.The MR Toolkit Peer Package is an extension to the MR Toolkit 7, 6 ] that allows multiple independent MR Toolkit applications to communicate with one another across the Internet. The peer package allows the master process of an MR Toolkit application to transmit device data to other remote processes, and to receive device data from remote processes. Application-speci c data can also be shared between independent M R Toolkit applications. Nominally, a n y n umb e r o f p e e r s m a y c o m m unicate together in order to run a multiprocessing application, and peers can join or leave the collaborative application at any time. Thus, the peer package allows each user to have a smooth interaction with the virtual environment, while updating each remote peer with the local device data. Previous WorkThe most famous example of a multi-user virtual environment is SIMNET 1], which i s a n i n teractive network system for real-time person-in-the-loop battle engagement simulation and war-gaming. SIMNET uses wide-area ethernets and satellite links to join hundreds of independent s i m ulated components such a s

Despite rapid growth in workstation and networking technologies, the workstation environment continues to pose challenging problems to shared processing. In this paper, we present a computational model and system for the generation of distributed applications in such an environment. The well-known RPC model is modified by a novel concept known as template attachment. A computation consists of a network of sequential procedures which have been encapsulated in templates. A small selection of templates is available from which a distributed application with the desired communication behavior can be rapidly built. The system generates all the required low-level code for correct synchronization, communication, and scheduling. This results in a system that is easy to use and flexible, and can provide a programmer with the desired amount of control in using idle processing power over a network of workstations. The practical feasibility of the model has been demonstrated by implementing it for Unid-based workstation environments.

The UofA* User Interface Management System (U&IS) generates graphical user interfaces based on a high-level description of semantic commands supported by the application. A main part of the UIMS, called Chisel, generates the presentation component of interfaces. Chisel selects interaction techniques, determines their attributes, and places them on the screen of the display device. While doing so it is capable of considering device properties, end user's preferences, and interface designer's guidelines. The aim of this paper is to discuss in detail the design and implementation of Chisel.The aim of this paper is to discuss the design and implementation of Chisel. An overview of the complete UIMS is presented in [Singh 89a] with greater detail being available in[Singh 89b]. Chisel is responsible for producing presentation components for graphical user interfaces. Chisel s~lccts intcraction techniques, determines their attribute values, and places them on the screen of the display device. While doing this, it is capable of considering user's prcfercnces, interface designer's guidelines, and characteristics of the display device. The input of Chisel consists of the dialogue rcquiremcnts (automatically produced by Diction), a description of the display device (selected from library of device descriptions provided by the UIMS), and end user's preferences (optional).Based on these inputs Chisel selects interaction techniques and "glues" them together to form a complete presentation component.

No abstract

A high-level UIMS which automatically generates the lexical and syntactic design of graphical user interfaces is presented. The interfaces generated by the UIMS can easily and rapidly be refined by the designer by using highly interactive and graphical facilities. The UIMS accepts a high-level description of the semantic commands supported by the application, a description of the implementation device, and optionally, the end user's preferences. Based on these inputs the UIMS generates graphical user interfaces in which the commands are selected from menus and command arguments are provided through interaction with graphical interaction techniques.