TY - GEN
T1 - Multiscale Empirical Analysis of Software Network Evolution
AU - Gou, Xiaodong
AU - Fan, Long
AU - Zhao, Li
AU - Shao, Qi
AU - Bian, Chong
AU - Yang, Shunkun
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - Understanding the evolution of the complex software system during the updates is useful for a variety of software development and maintenance activities, however, few works have studied the multiscale evolution characteristics of software in different programming languages from the perspective of the network. In this paper, we investigate the evolution of twelve open source software in ten programming languages from three scales: macroscopic, mesoscopic and microscopic. Analysis results show some interesting observations, for example, the in-degree distributions exhibit the coexistence of power-law distribution and lognormal distribution, and may even change to each other in evolution process. In addition, software does not always become more orderly and modular during the update process, but the software in different programming languages has similarities in microstructure. What's more, there are strong correlations between the software's macroscopic, mesoscopic, and microscopic properties and the size of the software, while there are no significant correlations between these three properties. The analysis and discussion in this paper can provide useful insights for software developers to understand the complexity of software.
AB - Understanding the evolution of the complex software system during the updates is useful for a variety of software development and maintenance activities, however, few works have studied the multiscale evolution characteristics of software in different programming languages from the perspective of the network. In this paper, we investigate the evolution of twelve open source software in ten programming languages from three scales: macroscopic, mesoscopic and microscopic. Analysis results show some interesting observations, for example, the in-degree distributions exhibit the coexistence of power-law distribution and lognormal distribution, and may even change to each other in evolution process. In addition, software does not always become more orderly and modular during the update process, but the software in different programming languages has similarities in microstructure. What's more, there are strong correlations between the software's macroscopic, mesoscopic, and microscopic properties and the size of the software, while there are no significant correlations between these three properties. The analysis and discussion in this paper can provide useful insights for software developers to understand the complexity of software.
KW - complex network
KW - multiple programming languages
KW - software evolution
UR - https://www.scopus.com/pages/publications/85140903274
U2 - 10.1109/QRS-C55045.2021.00165
DO - 10.1109/QRS-C55045.2021.00165
M3 - 会议稿件
AN - SCOPUS:85140903274
T3 - Proceedings - 2021 21st International Conference on Software Quality, Reliability and Security Companion, QRS-C 2021
SP - 1109
EP - 1118
BT - Proceedings - 2021 21st International Conference on Software Quality, Reliability and Security Companion, QRS-C 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 21st International Conference on Software Quality, Reliability and Security Companion, QRS-C 2021
Y2 - 6 December 2021 through 10 December 2021
ER -