Visual approach for change impact analysis: a controlled experiment

In the context of Software Maintenance, when a source code element must be changed, there is the need to identify if other elements will be affected by the change, in order to keep the code consistent. This identification is performed during the activity of change impact analysis. Aiming to support maintainers during this activity, software visualization tools allow a visual exploration of source code elements. In this paper, we present a study aimed at evaluating the support provided to change impact analysis by visual representations of Java program elements and their associations. To this end, we conducted a controlled experiment involving 24 undergraduate students, comparing the visual support approach and an ad hoc approach, where only the source code is analyzed to estimate impact change. Results showed that the effectiveness obtained by using the visual approach is significantly superior. This is an indication that visual support should be considered to change impact analysis aiming at reducing software maintenance costs.