Replication package with data used in the study: "The effect of code smells on the relationship between design patterns and defects. An empirical study"

This is a replication package with data used in a study by T. Alkhaeir and B. Walter "The effect of code smells on the relationship between design patterns and defects. An empirical study"

This dataset contains the following folders:

• "Analyzed systems" folder:

  • For each subject system (Ant-1.7, JEdit-4.2, Lucene-2.4, Camel-1.6, Log4j-1.2, Xalan-2.7, Poi-3.0, Ivy-2.0, Xerces-2.0, Velocity-1.6), we identify the following datasets: SDP, nSDP, SnDP, and nSnDP. Each dataset is represented by a separate csv file.

  • Those csv files include raw data about every class in every release. Each file includes columns which represent:

    • System:  The analyzed system

    • className: A fully qualified class name

    • Pattern: if the class is part of any pattern the cell contains the name of the pattern, and “null” otherwise

    • Smell: if the class is part of affected by any smell  the cell contains the name of the smell, and “null”  otherwise

    • Bugs: Number of defects reported inside the class (extracted from the PROMISE dataset)

• A "detailed analysis" folder:

  • For each pattern, we report all the classes which participate in it in all the analyzed systems. The csv files inside this folder follow the same structure as the other csv files reported above