AGOCG logo
Graphics Multimedia VR Visualisation Contents
Training Reports Workshops Briefings Index
Also available as an Acrobat File Back Next

Editorial

Abstract

Event History Data

Vis. Methods

Vis. Tools

Comparing Pencils

Further development

Difficulties

New tools

Acknowledgements

References


Case Studies Index

Visualisation of historical events using Lexis pencils

6. Difficulties encountered

We started this project by using AVS as our visualisation engine. This has a modular structure, and it was initially thought that programming the idea of the Lexis pencil display using the interactive module builder would be a relatively simple task. However, it soon became apparent that AVS had been designed for specific real world applications such as computational fluid dynamics and medical image processing. It became clear to us that AVS can still be used for our application, but a complex C program needed to be written both to read the event history data and to process it, providing such facilities as case selection, individual rotation of pencils and assignment of variables to faces which the standard AVS modules could not provide. This was converted into an AVS module and is available for downloading from the AVS repository and our own web pages.

Another disappointment was the ease of use of the AVS interface for panning and zooming. We used AVS on a Hewlett Packard workstation, using an X-terminal emulator running on a Pentium PC as the display. While speed of response was occasionally poor, a greater surprise was the paucity of user control for manipulating the viewpoint around the 3-D pencil world. Even with experience of the software, it was hard to control the viewpoint. In addition, it proved to be impossible to `bookmark' a viewpoint once a suitable view had been reached. Most VRML viewers provide far better control , with options such as `fly to', `walk' and `examine' which provide users with finer navigational control.
Finally, we have considered carefully the problem of dissemination. The software allows Postscript(TM) bitmaps of any view of the Lexis pencil world to be produced, and these can be published. However, it is often important to allow other users the ability to view the dataset themselves, and to identify their own special features of the data. We therefore provide the facility to write VRML (Virtual Reality Markup Language) files which define the 3-D Lexis pencil display. A difficulty with this was that at that time, surprisingly, no suitable VRML translator was available in AVS, and we therefore needed to write our own translator.

VRML viewers can now be purchased cheaply or downloaded free of charge on the World Wide Web to run on any platform. Other users should be able to easily load a VRML Lexis pencil file into their VRML viewer and fly through the 3-D Lexis pencil world on their own desktop computer.

However, it appears that the complexity of the 3-D Lexis pencil world defeats many VRML viewers currently available. Under Windows 3.1, for example, Cosmoplayer and Live3D simply crash the browser, presumably because of lack of memory. VRML viewers appear to be better behaved on UNIX and Windows NT systems, although even a popular VRML browser such as VRweb fails to display labelling text in the Windows NT version. A robust, fully specified VRML viewer has yet to be found, but Cosmoplayer appears to us to give the best results.

Table 1 summaries what we see as the successes and failures of our approach.

SuccessesFailures
AVS software module written with full user control to read and process any event history dataset. Visual complexity means that there are problems in examining more than 200 individual histories.
New substantive and statistical insights obtained into both datasets through the Lexis pencil display. AVS complicated to use and substantial training needed.
Dissemination of ideas has been comprehensive, with published papers, international conference presentations and a web site. Psychological issues of perception have not been addressed.
Pencil displays fit well into existing statistical literature.
Dissemination of displays has been partially successful, with output provided as a static Postscript(TM) file or as VRML code.
Table 1: Evalutation of successes and failures of the case studies

Graphics     Multimedia      Virtual Environments      Visualisation      Contents