Collaborative Visualization

By

Jason Wood

The University of Leeds, School of Computer Studies

February 1998

Submitted in accordance with the requirements for the degree of Doctor of Philosophy

 

The candidate confirms that the work submitted is his own and that appropriate credit has been given where reference has been made to the work of others.

Abstract

As researchers begin to tackle larger and more complex problems, so the need to collaborate within multi-disciplinary teams increases. These teams are likely to be geographically dispersed so incurring high cost, both financial and personal, when required to meet face-to-face. Recent advances in high speed networking, emerging standards for distributed information and increased computing power at the desktop offer potential for a solution to the problem of collaborating in a distributed environment. Existing tools support video and audio conferencing, but for collaboration over scientific data visualization there is no off-the-shelf solution.

Current visualization systems are designed around a single user model, making it awkward for large research teams to collectively analyse large data sets. This thesis shows how the popular dataflow approach to visualization can be extended to allow multiple users to collaborate - each running their own visualization pipeline but with the opportunity to connect in data, generated by a colleague, to any point in the pipeline. Thus collaborative visualizations are programmed in exactly the same plug-and-play style as is now customary for single-user mode. The work describes a system architecture, based on a new extended reference model, that can act as a basis for the collaborative extension of any dataflow visualization system. These ideas are demonstrated through a particular implementation in terms of IRIS Explorer. It also examines how this architecture can be used, in conjunction with the application building features of this class of system, to create tailored collaborative turnkey visualization applications.

While the work outlined above extends single-user visualization to real-time collaborative visualization across distance, support is also needed for collaborators who wish to work across time. This thesis examines the potential offered by the World Wide Web to support such collaboration. It begins by investigating the potential for single user visualization across the Web, and proposes an architecture and implementation for server-side visualization. This concept is demonstrated by an online real-time air quality data visualization system. This single-user architecture is then extended to support asynchronous collaboration. It offers collaborators the ability to visualize data, storing selected work along with relevant comments and for collaborators to review and extend this work at a later time. This forms the basis of an asynchronous collaborative visualization environment.

 

Acknowledgements

I would like to thank my supervisors Ken Brodlie and Helen Wright for their guidance and support throughout my time at the University of Leeds. I would especially like to thank them for their patience and encouragement while I have been writing this thesis. It has not been easy for any of us.

I would like to acknowledge NAG Ltd for their technical support on a number of occasions, and in particular Jeremy Walton and Alan Gay. I would also like to thank a number of people from the University who have helped in many ways including Justin Ware, Alison Tomlinson and Joel Smith. This work has been conducted within the VWE lab and has benefited from many discussions held with its members, especially Neil Hunter, Richard Drew, Mike Swaby and Dave Morris.

I would also like to thank Lisa for an enormous amount of support, patience and encouragement during the difficult periods of this work. Lastly, I would like to thank my parents for always supporting the choices I have made and believing that I could accomplish what I have set out to do.

 

Publications

The work in this thesis has been presented at the following conferences:

 

Contents

Chapter 1 - Introduction

Chapter 2 - Visualization Systems Review

Chapter 3 - Considerations for Collaborative Visualization Systems

Chapter 4 - Collaborative Visualization Using The COVISA Toolkit

Chapter 5 - Visualization Over The World Wide Web

Chapter 6 - Using The Web For Collaborative Visualization

Chapter 7 - Conclusions and Future Work

References

 

Table of Figures 

 Figure 1.1: Classification of working domains
Figure 2.1: Based on original paper [Haber90b]
Figure 2.2: Screen sharing architecture
Figure 2.3: Broadcast architecture
Figure 2.4: Synchronised model
Figure 2.5: Architecture of tempus fugit
Figure 2.6: Architecture for collaborative tempus fugit
Figure 2.7: Grave et al’s collaborative IRIS Explorer architecture
Figure 2.8: Based on Wierse93
Figure 3.1: Dataflow reference models
Figure 3.2: Osland visualization reference model
Figure 3.3: Dataflow pipeline reference model
Figure 3.4: Collaborative visualization pipeline
Figure 3.5: Selectively sharing an application - user A acting as master
Figure 3.6: CSpray - user B controlling spray can initiated by user A
Figure 3.7: Extended reference model for collaborative visualization
Figure 3.8: User A and user B view the same data in different ways
Figure 3.9: Users A and B synchronise their views
Figure 3.10: Users A and B collaborate on mapping stage (public data)
Figure 3.11: Users A and B collaborate on mapping stage but the filtered data remains private to A
Figure 4.1: Architecture for second prototype
Figure 4.2: Multi-user architecture
Figure 4.3: Local server module
Figure 4.4: Advisor module
Figure 4.5: Share module showing both disconnected and connected states
Figure 4.6: Each doctor studies their own data
Figure 4.7:Both doctors view all data
Figure 4.8: Using a shared pointer to indicate points in the scene
Figure 4.9: Sharing an alternative visualization of the data
Figure 4.10: Plugging together collaborative COVISA applications
Figure 4.11: Local server module in application mode showing both disconnected and connected states
Figure 4.12: COVISA used in on-the-fly mode
Figure 4.13: A singe pipeline with added sharing modules
Figure 4.14: Complete shared turnkey system
Figure 5.1: Extended dataflow pipeline reference model
Figure 5.2: Serving images of visualization results
Figure 5.3: Serving 3D scenes of visualization results
Figure 5.4: Client side visualization using a helper application
Figure 5.5: Client side visualization using Java
Figure 5.6: Server side visualization
Figure 5.7: Server side visualization - architecture
Figure 5.8: Current data display method (provisional data).
Figure 5.9: Web interface for air quality visualizer
Figure 5.10: Histogram view of ozone from three sites in London (prov data)
Figure 5.11: Surface view of ozone from three sites in London (prov data)
Figure 6.1: Collaborative eyechem
Figure 6.2: Extended web visualization architecture
Figure 6.3: Detailed architecture of history server
Figure 6.4: Thread numbering example
Figure 6.5: Extended page
Figure 6.6: History tree interface
Figure 7.1: Typical components of a single-user visualization application
Figure 7.2: Client/server, distributed and co-operative processing
Figure 7.3: Web based visualization
Figure 7.4: Collaboration via screen sharing
Figure 7.5: Collaboration by co-operative processing
Figure 7.6: Asymmetric collaboration