Advancing the Understanding of Software Evolution with Change Composition

Abstract

Software evolution is the continuous process of developing, adapting, and maintaining software systems to meet changing requirements. Understanding this process is vital for improving quality, reducing costs, and guiding good practices. A promising approach to understanding evolution is through change composition, which aggregates low-level modifications into highlevel change entities. Currently, refactoring is the most widely adopted form of change composition, but it has limitations. This dissertation advances the understanding of software evolution by firstly investigating three key limitations: a technical gap between refactoring detection granularity and practical application granularity, a human-centric challenge in the awareness of reviewability in interpreting the evolution path through refactoring, and a conceptual limitation of the limited expression ability of refactoring. Secondly, it proposes solutions to address these limitations.

BibTeX

@phdthesis{chenlei-dthesis,
    author = {Lei Chen},
    title = {Advancing the Understanding of Software Evolution with Change Composition},
    school = {東京科学大学 情報理工学院 情報工学系 情報工学コース},
    year = 2025,
}