Empirical evaluation

CodeCity has been used in a controlled experiment to empirically validate the Software Systems as Cities approach against the state of the practice in software development. The results of the experiment, published in the ICSE 2011 proceedings, showed that CodeCity enabled solutions to reverse engineering tasks which were 24% more correct and obtained in 12% less time than when using Eclipse or spreadsheets with metric data.

Replicability

To make the replication of this experiment easier, we provide all the artifacts we designed and used during the experiment.

Groups T1 T2 T3 T4
Tool CodeCity CodeCity Eclipse Eclipse
Analyzed System Azureus Findbugs Azureus Findbugs
Treatment Artifacts T1-codecity.zip
T1-src.zip
T2-codecity.zip
T2-src.zip
T3-metrics.zip
T3-src.zip
Eclipse Galileo
MS Excel
T4-metrics.zip
T4-src.zip
Eclipse Galileo
MS Excel