Efficient Specification Techniques for Software Visualization

Software Visualization is developing into an important tool for teaching algorithms and debugging complex sequential and parallel programs. However, specifying the relationship between program objects (static ones as code statements or variables and dynamic ones as data structures or object states) and their geometric counterparts turns out to be the major bottleneck in using SV systems [1, 2, 3]. My work aims at the development of a simple, universal, flexible, and modular architecture allowing novice and expert users to efficiently specify visualizations at arbitrary levels of abstraction (abstract data type operations, program statement or variable observation). Thus, it represents a natural extension of the concept of a debugger program, but is also a valuable tool for Intelligent Tutoring Systems. My approach is best described as a combination of visual programming and direct manipulation techniques with the additional possibility of conventional programming. A key feature of the..