Next Previous Up Top Index Index

1. Rationale for a WWW-based visualization service

1. 1. Demand for general availability of visualization techniques

There is a demand to make visualization techniques available for general use. This demand is driven by various clients. A low-cost or even no-cost visualization service on the WWW would drastically lower the entry barrier for the use of visualization. In particular, for users who rarely need visualization the costs or overhead created by network usage may be far less than what is needed to buy and maintain an in-house system.

Frequent users of a given visualization system may still have a strong desire for additional visualization service. Since the scope of methods available in a given system is limited, additional methods may be used remotely and the results may be combined with those results obtained locally. This may apply for highly specialized methods which typically cannot be found in commercial software packages.

Error estimation is another major reason to look for additional visualization methods. In today's software implementation, the various sources of errors are not very well treated. Detailed information about the implemented algorithms is often not available. Therefore, it is good practice to visualize a particular feature of interest in a set of data by using a variety of different and independent algorithms in order to verify the result and obtain an error estimate. In some case this can be done using a single general purpose visualization tool. There are cases, however, where the use of more than one system provides a significant advantage. Access to alternative systems over the WWW would come in handy. A good example would be to compare various algorithms for particle tracing.

The visualization research has good reasons to make new algorithms available to the public. Usually, visualization research does not directly lead to a commercial product designed to make profits. In most cases, advanced methods are first created and used in the scientific community. These developments are soon available to the public through published literature. Many obstacles prevent these algorithms from being rapidly spread and incorporated into a large number of systems as system development is in the hands of few organizations with scarce resources. On the other hand, application of algorithms is a key issue to future developments and improvement. Wide spread usage of visualization algorithms triggers cooperation and new ideas in the visualization community. Today this cooperation is slowed down by hardware and software constraints in the various research organizations.

Comparative visualization has attracted significant attention recently [10]. This often requires the combination of data from different sources and different visualization algorithms for a single image. While one part of the data is usually familiar one along with the methods that treat this part of the data, the other part of the data may only be used once for the sole purpose of comparison. For example, a researcher numerically simulating flows tries to match his result with some flow experiment done elsewhere. In such a case the appropriate methods to visualize the experimental data may not be readily available to him. Experience shows that direct cooperation between persons in different organizations will probably be required to reach the goal. However, such cooperation is subject to time constraints. Once researchers, who supply data also supply visualization methods on the network, these obstacles will vanish. The likelihood for actual use of research results would increase.