24 . 3 Visualization by Linking Machines : Web - based Visualization

The Internet has changed the face of computing. The use of e-mail and desk-top video-conferencing has made collaboration between people very easy and very natural. Application-sharing technology, such as provided by Microsoft's NetMeeting, allows a single application to be accessed by a group of people at different locations. The Internet likewise enables collaboration between computers-the Distributed.net activity [18] harnesses spare computing power on the Internet to tackle mathematical tasks that otherwise could hardly be contemplated. In this chapter we look at the opportunities offered by the Internet for the visualization community, and the volume visualization community particularly. The volume visualization task is often compute-intensive and complex-how can the Internet help by distributing the computational load? This we shall term: Visualization by Linking Machines. Again volume visualization requires judgement and skill, particularly in the segmentation stage-how can the Internet help by distributing the human processing between a number of people? This we call: Visualization by Linking Humans. The field of 'Distributed Cooperative Visualization' is reviewed in detail in the Eurographics State of the Art Report [3]; our aim in this chapter is to give a more concise treatment, and to look more specifically at volume graphics applications. We begin by looking at visualization by linking machines. There are two main approaches. The first, which pre-dates the Web, splits the task into a number of sub-processes, and places the computationally intensive parts on a super-computer. This is covered in section 24.2. The second approach, described in section 24.3, uses Web technology to distribute processing between a client front-end and a server back-end. This is an increasingly popular approach, and we shall try to compare and contrast the different styles of Web-based visualization. We then go on to look at visualization by linking humans. This is much less developed. In section 24.4, we describe some general approaches to collaborative applications, and explain why these general approaches may not be suitable for visualization. In section 24.5, we turn specifically to visualization. Although the popular Modular Visualization Environments (MVEs)-such as IRIS Explorer [6], AVS [12] and IBM Data Explorer [1]-were designed from the outset to support distributed processing, they were all based on a single-user model. If you want to collaborate, you cluster around the same workstation. It is only recently that we have