Test Smell Research

Provided below are supplements to the original research on test smells that was performed using tsDetect.

Experiment Data

The data generated by our tools have been made available on our website. In-addition to the test smells exhibited by the projects, the dataset also includes details on the project's commits, releases and issues. Also made available are the unit test files used in the precision and recall experiment of our study.
Click to continue reading...

Tool Correctness

tsDetect is able to correctly detect test smells with a precision score ranging from 85% to 100% and a recall score from 90% to 100% with an average F-score of 96.5%
Click to continue reading...

Tool Comparison

As part of our study, we also performed a comparative exercise of tsDetects ability of smell detection against state-of-art detection strategies.
Click to continue reading...

Test Suite Characteristics

Do specific test smells have an impact on test suite characteristics?

To further understand the degree to which test smells can impact software maintenance activities of an Android app, we studied the impact of each test smell type on specific characteristics of the apps test suite.
Click to continue reading...

Test Smell Refactoring

The goal of this work is to understand the relationship between refactoring changes and their effect on test smells. We empirically explore the impact and relationship between refactoring operations and test smells in 250 open-source Android apps.

Click to continue reading...