On the Evaluation of Software Maintainability Using Automatic Test Case Generation

Abstract

The measurement of external software attributes and the analysis of how those attributes have evolved through the software's releases are challenging activities. This is particularly evident when we discuss the maintainability of Object-Oriented (OO) systems which, due to their specific characteristics, hide information that cannot be gathered through static analysis. As maintainability can be defined as the "speed and ease with which a program can be corrected or changed", we believe that test data are reflective of the changes performed during maintenance. Moreover, empirical observations allow to speculate about the relationships between maintainability and the behaviour of the software when executed by test cases (e.g., coverage values) and the test cases' characteristics (e.g., generation time). Our aim is to complement the state-of-the art by proposing a new approach for understanding and characterizing the maintainability of OO systems, which makes use of test data and of the information gathered from the tests' execution.

Scented since the beginning: On the diffuseness of test smells in automatically generated test code

“…We found a medium correlation for Eager Test (τ = −0.31) and a small one (τ = −0.24) for Assertion Roulette. This result seems to support what has been previously shown by Nogueira et al [96] on the ability of inheritance metrics to provide information on the quality of test cases.…”