Understanding software evolution using a combination of software visualization and software metrics

Abstract

Coping with huge amounts of data is one of the major problems in the context of software evolution. Current approaches reduce this complexity by filtering out irrelevant information. In this paper we propose an approach based on a combination of software visualization and software metrics, as software visualization is apt for complexity reduction and metrics introduce the possibility to qualify evolution. We discuss a simple and effective way to visualize the evolution of software systems which helps to recover the evolution of object oriented software systems. In addition we define a vocabulary that qualifies some specific situations that occurs when considering system evolution. RÉSUMÉ. Analyser un très grand volume de données est un des problèmes majeurs lors de la compréhension de l'évolution de logiciels. Les approches existantes réduisent cette complexité en filtrant les informations non pertinentes. Dans cet article nous proposons une approche basée sur la combinaison de métriques et de visualisation, la visualisation permettant une réduction d'information et les métriques permettant une qualification de l'évolution. Ainsi nous présentons une matrice d'évolution : une visualisation simple et efficace qui aide à comprendre l'évolution des applications orientées objets. En plus, nous définissons un vocabulaire permettant de qualifier les situations caractéristiques rencontrées.

Studying the co-evolution of production and test code in open source and industrial developer test processes through repository mining

“…Spectrographs allow to visualize the decay of a particular property over time, e.g., an entity that has just been changed can be colored red, while a component that has been changed a little while longer ago, can be colored orange. Lanza (2001) and Lanza and Ducasse (2002) extend on the previous idea of polymetric views (Lanza and Ducasse 2003) by using a visualization they call the evolution matrix. This visualization depicts time on the X axis, entities of the software system on the Y axis and they use polymetric views, i.e., boxes of different colors and shapes to express properties of the entities over time.…”

On the Relationship between Program Evolution and Fault-Proneness: An Empirical Study

“…Lanza et al [30] presented an evolution matrix to display the evolution of the classes of a program. Each column of the matrix represents a version of the program, while each row represents the different versions of the same class.…”

Dimensions of reengineering environment infrastructures

“…The more models, the more information must be dealt with. Only Moose supports multiple models, mainly for evolution analysis [29].…”

“…The possibility of analyzing multiple models simultaneously is useful in evolution analysis, where multiple versions of the same system need to be analyzed [29,30]. Likewise, it is interesting to analyze parallel branches of similar applications, for instance, to develop a framework by abstracting common assets in these branches.…”