Virtual software in reality

Abstract

Software visualisation is an important weapon in the program comprehension armoury. It is a technique that can, when designed and used effectively, aid in understanding existing program code. It can achieve this by displaying information in new and different forms, which may make obvious something missed in reading the code. It can also be used to present many aspects of the data at once. Software, despite many software engineering advances in requirements, design and implementation techniques, continues to be complex and large and if anything seems to be growing in these respects. This means that techniques that failed to aid comprehension and maintenance are certainly not going to be able to deal with the current software. Therefore this area requires research to be able to suggest solutions to deal with the information overload that is sure to occur. There are several issues that this thesis addresses; all of them related to the creation of software visualisation systems that are capable of being used and useful well into the next generation of software systems. The scale and complexity of software are pressing issues, as is the associated information overload problem that this brings. In an attempt to address this problem the following are considered to be important: abstractions, representations, mappings, metaphors, and visualisations. These areas are interrelated and the first four enable the final one, visualisations. These problems are not the only ones that face software visualisation systems. There are many that are based on the general theory of the applicability of the technique to such tasks as program comprehension, rather than the detail of how a particular code fragment is shown. These problems are also related to the enabling technology of three- dimensional visualisations; virtual reality. In summary the areas of interest are: automation, evolution, scalability, navigation and interaction, correlation, and visual complexity. This thesis provides an exploration of these identified areas in the context of software visualisation. Relationships that describe, and distinguish between, existing and future software visualisations are presented, with examples based on recent software visualisation research. Two real world metaphors (and their associated mappings and representations) are defined for the purpose of visualising software as an aid to program comprehension. These metaphors also provide a vehicle for the exploration of the areas identified above. Finally, an evaluation of the visualisations is presented using a framework developed for the comparative evaluation of three-dimensional, comprehension oriented, software visualisations. This thesis has shown the viability of using three-dimensional software visualisations. The important issues of automation, evolution, scalability, and navigation have been presented and discussed, and their relationship to real world metaphors examined. This has been done in conjunction with an investigation into the use of such real world metaphors for software visualisation. The thesis as a whole has provided an important examination of many of the issues related to these types of visualisation in the context of software and is therefore a valuable basis for future work in this area.