A method for structuring functional and technical domains of aninformation system based on a project change history

Abstract

Relevance: Information systems continuously evolve through commits and iterative modifications, which gradually reduce architectural transparency and blur responsibility boundaries between components. In long-lived systems, static architectural documentation often becomes outdated, while repository change history preserves objective evidence about how artifacts evolve together and how evolutionary dependencies emerge over time. Aim of the article: The paper aims to develop a method for the automatic structuring of the functional and technical domains of an information system, based on a project change history and evolutionary coupling signals extracted from repository data. Tasks. The research tasks include formalizing the change history model, defining evolutionary cohesion indicators, constructing an integrated evolutionary cohesion graph, developing a graph-based domain grouping procedure, differentiating recovered domains into functional and technical responsibility areas, and defining the output representation and validation strategy. Methods: The proposed research combines repository mining, evolutionary coupling analysis, graph-based clustering, semantic similarity analysis, temporal cohesion analysis, and dependency-aware modeling. The approach integrates co-change frequency, temporal proximity, semantic similarity, and structural dependency signals into a unified evolutionary cohesion measure represented as a weighted graph. Scientific novelty: The scientific novelty of the research reflected in the paper lies in recovering domain structures not from static architecture or documentation, but from observable evolutionary behavior recorded in repository history. The proposed method introduces an integrated evolutionary cohesion measure and an explainable differentiation mechanism that separate recovered domains into functional and technical responsibility areas based on ownership dispersion, reuse index, and cross-domain connectivity indicators. Practical significance: The proposed method improves architectural transparency and supports maintenance planning, refactoring, change impact analysis, technical debt management, and responsibility reasoning in evolving information systems. The resulting domain map and interaction graph can assist developers in identifying highly coupled areas and understanding evolutionary dependencies between subsystems. Results. The paper presents a complete pipeline for transforming repository evolution data into an interpretable domain structure model. The method constructs an evolutionary cohesion graph, identifies candidate domains through graph-based grouping, and classifies them into functional and technical areas using explainable indicators. An illustrative example demonstrates how stable co-evolution patterns form interpretable domain areas even without explicit architectural documentation. Conclusions. The proposed method enables recovering interpretable functional and technical domain structures from a project change history by combining co-change, semantic, dependency-aware, and temporal signals. The resulting domain decomposition supports architectural understanding and maintenance-oriented reasoning while reducing dependence on incomplete or outdated architectural documentation. Full empirical validation and further enhancement of classification and weighting strategies remain directions for future research.