Visualization for eResearch: past, present and future

Abstract

Visualization has been a part of computing for a long time, however with the growth in data produced by researchers and the computing resources available visualization capability has not developed in a way that provides researchers with the ability to include visualization as part of their standard analysis of these very large datasets. To address this issue and to understand how visualization systems must adapt to meet the new needs the eResearch brings we examine the past, present and begin to look into the future at visualization systems and architectures, to aid in understanding how visualization may be used or wish to be used a scenario involving a variety of Earth Science researchers working at a variety of locations to collect data and conduct analysis is presented. Reviewing traditional visualization systems, in particular, Modular Visualization Environments and Visualization Toolkits, to understand the heritage of visualization systems and the challenges that researchers have identified face. We look at current visualization systems that begin to take advantage of grid computing technologies, including those that modify traditional systems, those that a new architectures and those that have been developed in a bespoke manner for particular eResearch projects. Whilst these current visualization systems address some of the challenges of visualization for eResearch several challenges still exist and we examine ways in which these systems need to develop into the future to meet these challenges relating to use of multiple datasets, display devices, variation in bandwidth availability, the need for interaction and the role that predictive rendering can play in this, the need for new and revised algorithms, a focus on the end to end performance of visualization pipelines and the ability to integrate in to a researchers workflow rather than be an additional activity.