RT Journal Article
A1 Ana Vranković
A1 Tihana Galinac Grbac
A1 Željka Car

PB iet
T1 Software structure evolution and relation to subgraph defectiveness
JN IET Software
VO 13
IS 5
SP 355
OP 367
AB Network analysis has been successfully applied in software engineering to understand structural effects in the software. System software is represented as a network graph, and network metrics are used to analyse system quality. This study is motivated by a previous study, which represents the software structure as three-node subgraphs and empirically identifies that software structure continuously evolves over system releases. Here, the authors extend the previous study to analyse the relation of structural evolution and the defectiveness of subgraphs in the software network graph. This study investigates the behaviour of subgraph defects through software evolution and their impact on system defectiveness. Statistical methods were used to study subgraph defectiveness across versions of the systems and across subgraph types. The authors conclude that software versions have similar behaviours in terms of average subgraph type defectiveness and subgraph frequency distributions. However, different subgraph types have different defectiveness distributions. Based on these conclusions, the authors motivate the use of subgraph-based software representation in defect predictions and software modelling. These promising findings contribute to the further development of the software engineering discipline and help software developers and quality management in terms of better modelling and focusing their testing efforts within the code structure represented by subgraphs.
K1 average subgraph type defectiveness
K1 software evolution
K1 software engineering discipline
K1 different defectiveness distributions
K1 software modelling
K1 subgraph frequency distributions
K1 defect predictions
K1 three-node subgraphs
K1 software versions
K1 software developers
K1 network analysis
K1 system defectiveness
K1 analyse system quality
K1 network metrics
K1 software network graph
K1 structural evolution
K1 structural effects
K1 software structure evolution
K1 subgraph-based software representation
K1 subgraph defects
K1 code structure
K1 system releases
K1 system software
DO https://doi.org/10.1049/iet-sen.2018.5060
UL https://digital-library.theiet.org/;jsessionid=w7kltjq9f61a.x-iet-live-01content/journals/10.1049/iet-sen.2018.5060
LA English
SN 1751-8806
YR 2019
OL EN