software frameworks

This study addresses the proliferation of functional redundancy in service-oriented architectures caused by heterogeneous clients, which undermines system evolvability and maintainability. To mitigate this issue, the authors propose a novel reference architecture that synergistically integrates metadata-driven mechanisms with pattern languages. By leveraging metadata management and a plugin-based design, the approach effectively constrains service redundancy while enhancing reuse capabilities. The work innovatively combines metadata mechanisms and pattern languages in architectural construction and validates its efficacy through a triangulated evaluation method incorporating scenario-based assessment and real-world case studies. Empirical results demonstrate that the majority of system changes during evolution require no code modifications—only configuration adjustments or the addition of pluggable components—thereby significantly improving architectural stability and reuse efficiency.

Software architecture suffers from ambiguous abstraction concepts and inadequate tool support. Method: This work systematically reconstructs the seminal 1995 architectural model and proposes, for the first time, a practice-grounded conceptual framework for architectural abstraction—elevating component composition relationships to system-level abstractions that are formally modelable and verifiable. It integrates architectural description language (ADL) design, abstract modeling, prototype tool development, and diachronic historical analysis. Contribution/Results: The study establishes software architecture as an independent concern with rigorous theoretical foundations. Its outcomes catalyzed a surge in ADL research, laid the groundwork for model-based systems engineering (MBSE), and continue to inform the design of cloud-native, microservice, and AI-driven architectures. The framework significantly enhances the expressiveness, formal verifiability, and engineering applicability of architectural abstractions.

Current software architecture design suffers from insufficient human-AI collaboration, opaque LLM-based tools, and rigid, inflexible workflows. Method: This paper proposes an architect-centered semi-automated architecture generation approach that deeply integrates large language models (LLMs) into four core activities—domain modeling, use case specification, architectural decision-making, and evaluation—supported by interpretable and intervenable human-in-the-loop interaction mechanisms to preserve architect agency and control throughout the process. Contribution/Results: We introduce the first LLM-augmented collaboration paradigm spanning the full architecture lifecycle, unifying domain modeling theory, structured decision frameworks, and a multi-dimensional evaluation metric system. Empirical evaluation demonstrates significant improvements: average architect effort reduced by 42.3% in preliminary experiments, alongside enhanced decision consistency and architectural evolvability.

This work addresses the challenge of maintaining up-to-date architectural documentation in microservice systems, which is exacerbated by polyglot implementations, multiple repositories, and rapid independent evolution. Existing static refactoring approaches are often limited to single-repository settings or homogeneous technology stacks. To overcome these limitations, we propose a distributed static architecture reconstruction framework that supports multi-language and multi-repository environments. The framework employs pluggable extractor modules for language-specific analysis and introduces mechanisms for cross-repository data propagation and fusion, enabling seamless interoperability with existing static analysis tools. To the best of our knowledge, this is the first framework to enable distributed, collaborative architecture reconstruction, significantly enhancing the scalability and usability of automated documentation generation and maintenance in complex microservice ecosystems.

Existing software architecture frameworks inadequately model machine learning (ML) systems, as they overlook the needs of emerging stakeholders—such as data scientists and data engineers—and lack expressive support for ML-specific characteristics, including component uncertainty, heterogeneity, and collaborative behavior. Method: Through an empirical study involving interviews and surveys with 61 domain experts from 25 organizations across 10 countries, we systematically identified ML-relevant stakeholders and their concerns for the first time. Contribution/Results: We propose novel, ML-adapted architectural viewpoints and views, extending traditional frameworks to enable unified modeling of both ML and non-ML components. This yields the *ML-Enhanced Systems Architecture Framework Extension Guide*, which has been preliminarily adopted in industry for intelligent system architecture governance. Our work bridges a critical theoretical and practical gap in stakeholder modeling and viewpoint systematization for ML system architecture design.

AI-assisted development tools enable rapid prototyping of services but often lack awareness of architectural constraints, infrastructure dependencies, and organizational standards required in production environments. Consequently, generated artifacts may exhibit brittle behavior and limited deployability. We propose a retrieval-augmented scaffolding approach that combines platform-based code generation with agentic clarification loops to expose and resolve architectural constraint ambiguities. By combining template retrieval with structured interaction, the method embeds production-relevant considerations during service scaffolding. Evaluation indicates improved architectural consistency and deployability compared to general-purpose AI code generation workflows, suggesting that constraint-aware retrieval is essential for aligning AI-assisted service development with production software engineering practices.

This work addresses the common neglect of software design principles in existing automated code generation approaches, which often results in mobile applications with poor architectural quality. To overcome this limitation, the authors propose a novel method that integrates software product line engineering with variability modeling of design patterns. For the first time, the Universal Variability Language (UVL) is employed to explicitly capture structural and behavioral variations of design patterns, enabling their integration as configurable assets within the code generation pipeline. Leveraging UVL models, the Jinja templating engine, and Swift-based code synthesis, the proposed system supports the customizable, automated generation of design patterns such as Singleton and Strategy. This approach not only preserves architectural integrity but also significantly enhances application maintainability and reusability.

This study addresses the fragmentation of code and documentation generated by generative AI across the software development lifecycle, a consequence of the absence of a unified multi-layer architectural framework. To bridge this gap, the work proposes the first integrated architectural metamodel that spans business, system, and development layers, providing large language models (LLMs) with structured architectural context to enable a closed-loop transformation among code, documentation, and code. This metamodel establishes a semantic interface between human developers and LLMs, substantially improving the accuracy, stability, and reproducibility of AI-generated artifacts. Experimental results demonstrate that the approach consistently yields high-quality code and documentation while enhancing architectural consistency in AI-driven development, thereby laying a scalable foundation for intelligent software development lifecycle (SDLC) tools.

This study addresses the frequent performance degradation of coding agents following scaffold updates—a phenomenon often misattributed by developers to the underlying large language model (LLM) rather than the scaffold itself. By holding the LLM constant and systematically varying only the scaffold, the authors conduct a longitudinal evaluation across 35 consecutive versions of Qwen Code CLI. Combining the SWE-bench Verified benchmark, open-source version tracking, and pull-request-level qualitative analysis, they establish a causal link between high-frequency scaffold iterations and agent performance fluctuations. The findings reveal that mainstream scaffolds are updated more than twice daily on average, with certain revisions significantly reducing task success rates. The study further identifies specific pull requests that directly impair performance, offering the first empirical guidance for scaffold engineering in LLM-based coding agents.

This study addresses the lack of systematic understanding regarding the application domains, maintenance characteristics, and effective design practices of GitHub template repositories. Conducting the first large-scale empirical investigation, the work integrates data mining, statistical analysis, code quality assessment tools—detecting code smells, vulnerabilities, and security hotspots—and an LLM-as-a-judge classification approach to systematically uncover domain distributions, language-specific quality variations, and maintenance patterns. The findings reveal web development as the dominant application domain, with high-quality templates consistently adhering to software engineering best practices and offering comprehensive documentation. Through qualitative evaluation, the study distills actionable design guidelines and identifies common pitfalls, providing practical guidance for developers creating or using template repositories.