An empirical analysis of open source software defects data through software reliability growth models

Abstract

The purpose of this study is to analyze the reliability growth of Open Source Software (OSS) using Software Reliability Growth Models (SRGM). This study uses defects data of twenty five different releases of five OSS projects. For each release of the selected projects two types of datasets have been created; datasets developed with respect to defect creation date (created date DS) and datasets developed with respect to defect updated date (updated date DS). These defects datasets are modelled by eight SRGMs; Musa Okumoto, Inflection S-Shaped, Goel Okumoto, Delayed S-Shaped, Logistic, Gompertz, Yamada Exponential, and Generalized Goel Model. These models are chosen due to their widespread use in the literature. The SRGMs are fitted to both types of defects datasets of each project and the their fitting and prediction capabilities are analysed in order to study the OSS reliability growth with respect to defects creation and defects updating time because defect analysis can be used as a constructive reliability predictor. Results show that SRGMs fitting capabilities and prediction qualities directly increase when defects creation date is used for developing OSS defect datasets to characterize the reliability growth of OSS. Hence OSS reliability growth can be characterized with SRGM in a better way if the defect creation date is taken instead of defects updating (fixing) date while developing OSS defects datasets in their reliability modelling.