Abstract | ||
---|---|---|
Modularity flaws can hamper the reuse and maintainability of code or even indicating the architecture degradation of a system. Therefore, researchers have increasingly investigated new mechanisms to assist the detection of these anomalies. Strategies for detection these flaws usually ignore information about the software change history. However, recent studies report that these strategies have been considered counter-productive. This article proposes and evaluates the use of detection strategies consisting of metrics that consider historic properties of evolving source code. It also proposes tool support for history-sensitive detection of modularity flaws. The strategies are evaluated in terms of precision and recall to detect three classic modularity flaws over 16 versions of two systems. Several observations were made, including: (i) exploiting information about the code evolution can contribute to effective detection of modularity flaws; and (ii) in both systems, history-sensitive strategies presented results superior to conventional strategies. |
Year | DOI | Venue |
---|---|---|
2010 | 10.1109/SBCARS.2010.11 | SBCARS |
Keywords | Field | DocType |
modularity flaw,evolving code,code evolution,detection strategy,detecting modularity flaws,source code,effective detection,history-sensitive detection,architecture degradation,conventional strategy,history-sensitive strategy,classic modularity flaw,history,software maintenance,xml,anomaly detection,measurement,surgery,computer architecture,dsl,source coding,empirical software engineering | Data mining,Anomaly detection,Software engineering,Source code,Precision and recall,Software,Empirical process (process control model),Software maintenance,Engineering,Modularity,Maintainability | Conference |
Citations | PageRank | References |
2 | 0.39 | 10 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Leandra Mara da Silva | 1 | 2 | 0.39 |
Francisco Dantas | 2 | 168 | 9.56 |
Gustavo Honorato | 3 | 7 | 0.83 |
Alessandro Garcia | 4 | 2231 | 143.70 |
Carlos Lucena | 5 | 589 | 41.51 |