ANDROID PROJECT MIGRATION METHODOLOGY BASED ON MVVM PROJECT REVIEW TO JETPACK COMPOSE AND MVI ARCHITECTURE

Abstract

The article provides a comprehensive analysis of the transformation of architectural approaches in Android development in the context of the transition from imperative UI models to declarative frameworks, in particular Jetpack Compose. The relevance of the study is due to the increasing complexity of mobile applications, increased requirements for state management, predictability of interface behavior and scalability of architectural solutions, which reveals the limitations of traditional MVC, MVP and MVVM patterns. The purpose of the study is to scientifically substantiate the formation and experimental verification of the methodology for migrating Android projects from the MVVM architecture with XML interfaces to the declarative Jetpack Compose model in combination with the Model–View–Intent (MVI) architecture, as well as to assess its impact on the structural complexity, performance, stability of state management and testability of applications. The methodological basis of the work is a systematic and comparative analysis of architectural patterns, a critical review of peer-reviewed publications from 2020 to 2025, architectural modeling of the MVI cycle, incremental migration of UI components from XML to Compose, as well as experimental testing of the proposed methodology on the example of a real Android module. The study shows that the combination of Jetpack Compose and MVI forms a new paradigm "UI-as- State", within which the interface is defined as a deterministic function of a single immutable state. It is established that the transition to unidirectional data flow and centralized state management based on StateFlow and coroutines significantly reduces architectural complexity, eliminates desynchronization of UI states, and minimizes side effects. Experimental results showed a reduction in the number of files and lines of code, a decrease in UI defects, faster rendering and increased interface stability, as well as a significant improvement in testability due to the use of pure reducer functions.