Empirical evaluation of software maintainability based on a manually validated refactoring dataset

Abstract

Context: Refactoring is a technique for improving the internal structure of software systems. It has a solid theoretical background while being used in development practice also. However, we lack empirical research results on the real effect of code refactoring and its application. Objective: This paper presents a manually validated subset of a previously published dataset containing the refactorings extracted by the RefFinder tool, code metrics, and maintainability of 7 open-source systems. We found that RefFinder had around 27% overall average precision on the subject systems, thus our manually validated subset has substantial added value. Using the dataset, we studied several aspects of the refactored and non-refactored source code elements (classes and methods), like the differences in their maintainability and source code metrics. Method: We divided the source code elements into a group containing the refactored elements and a group with non-refactored elements. We analyzed the elements' characteristics in these groups using correlation analysis, Mann-Whitney U test and effect size measures. Results: Source code elements subjected to refactorings had significantly lower maintainability than elements not affected by refactorings. Moreover, refactored elements had significantly higher size related metrics, complexity, and coupling.

An Empirical Study on Refactoring-Inducing Pull Requests

“…We formulated eight hypotheses on the differences/similarities between refactoring-inducing and non-refactoring-inducing PRs, discussed as follows. This is an expected result in light of the findings from Hegedüs et al, since refactored code has significantly higher size-related metrics [40]. We speculate that reviewing larger code churn may potentially promote refactorings.…”

Multi‐criteria decision‐making approach for modular enterprise resource planning sorting problems

“…Software maintainability analyses should be led at the levels of component, subsystem and whole system to directly and simultaneously enhance its functionality, reliability and efficiency. Maintainability is considered as a crucial prerequisite for lifelong evolution of systems on the basis of efficient updating of their related software (Vogel-Heuser & Ocker, 2018), and is derived from particular source code metrics as indicators (Hegedűs et al, 2018). As observed by Saraiva et al (2015), a considerable number of product metrics has been proposed for software maintainability assessment in the literature, what poses a decision-making challenge for practitioners and researchers with relation to gain a consolidated overview about available and practically beneficial software metrics (Bouwers et al, 2014).…”

Harnessing deep learning algorithms to predict software refactoring

“…In this paper, we used a public empirical dataset containing refactoring data for 7 open-source systems (antlr4, junit, mapdb, mcMMO, mct, oryx, titan) [32]. The experimental dataset used in our study is freely and publicly available at the PROMISE Repository.…”