Code and Change History Analysis for Information System Evolution: A Survey of Approaches

Abstract

Long-term system development gradually obscures architectural structure and makes evolution harder to interpret at the system level. This paper reviews four main directions of system evolution analysis: static code analysis, change history analysis, metrics-driven analysis, and AI-based analysis. It shows that these directions provide valuable but partial support: static analysis recovers structural views, change history reveals evolutionary dependencies, metrics support assessment, and AI improves prediction and recommendation tasks. However, existing studies usually address these aspects separately and therefore do not provide a unified view of their effect on system evolution. This paper argues that future progress depends on integrated and explainable approaches that combine structural, historical, semantic, and quantitative evidence. Such approaches can better identify responsibility areas, evolution-sensitive zones, and broader system-level patterns needed for more coherent long-term development. Authors propose an integrated system-level approach that combines structural, historical, semantic, and quantitative evidence into a unified representation of evolutionary cohesion. The approach includes domain structuring, role interpretation, and evolutionary risk assessment, producing interpretable decision-support outputs.