A survey on code coverage as a stopping criterion for unit testing
Ben H. Smith, Laurie A. Williams · NCSU Libraries Repository (North Carolina State University Libraries) · 2008
The evidence regarding code coverage as a predictor of software quality is conflicting and inconclusive.However, an estimate of the software testing practices of the majority of professionals can help researchers know how code coverage is being used-or whether it is being used at all.The purpose of this report is to present the results of an online survey we conducted to estimate the percentage of software developers who use code coverage as a stopping criterion for unit testing.We find that a majority of participants 1) perform automated unit testing; and 2) use code coverage, though not always as a stopping criterion.Those people who do not use code coverage, do not find it useful or provide some other reason.Finally, in place of code coverage, we find that most participants stop testing when they have "tested the most important parts of the code".