An initial approach towards measuring and characterising software evolution

Proposes a number of metrics to assess the maintainability of code. Dominance relations are used to express the complexity of call relations and an assumption is made that, when more complex relations are present within a source code module, it is harder for an application to be maintained. The paper describes a case study whereby the changes in dominance relations are investigated over versions of the software. Based on the changes to the relations, assumptions are made regarding the maintainability of the software. These assumptions are later verified by investigating change documentation and interviewing the maintainers. The results of this study seem to indicate that the dominance relations do reflect some principles of code maintainability.