System requirements and goals

When starting to build DynSys3D, we wanted to develop a workbench for several researches working collaboratively in the field of three-dimensional dynamical systems. Therefore several requirements and goals influenced the design of DynSys3D:

• Extendibility: It should be very easy to extend the system by new components, e.g., new visualization techniques. Similarly it should be as simple as possible to apply already implemented visualization techniques to new dynamical systems.
• It should be possible to easily compare different techniques, e.g., the behavior of different numerical integration methods. Components of the system, which obviously allow alternative solutions, should be exchangeable as well.
• The system should especially support rapid prototyping in a collaborative environment, i.e., the structure of DynSys3D should allow multiple users to develop additional system components in parallel, even if they depend on each others work.
• Interactivity: Real time visualization is quite difficult to achieve in most situations. Within DynSys3D the geometric representation is controlled by a parameter which influences the geometric level of detail to allow an interactive examination of the results.
• The system design should induce some symmetry guidelines to help developers to implement modules, which are intuitive to combine and use. Questions as where to use the AVS data-flow mechanism or which parameters should be handed over to the user, should be answered prior to individual decisions during the development.
• The system should provide control mechanisms to allow developers to selectively offer only those parameters to the user, which he is primarily interested at in his current investigation. It is very important to adjust number and representation of module parameters to provide a sensible user interface. Too many parameters as well as missing ones hamper the usability of a module.
• It should be possible for developers to concentrate specifically on the visualization part of the investigation. This requirement especially targets related problems as, e.g., how to find critical points or cycles of a dynamical system. General solutions to these tasks were not in the scope of this project.

mailto:helwig@cg.tuwien.ac.at.