為因應資訊快速發展，軟體開發商需在短時間內完成軟體開發以提升其市場競爭力，這也導致其無法在完整的系統分析下投入開發。因此，軟體開發者可使用軟體反向工程 (SRE) 工具令開發完的軟體快速產出系統設計文件。然而，若倉促間實作的程式碼品質不佳，會進而影響SRE產出之UML圖的品質。為解決此問題，本研究針對SRE產出之設計文件與程式碼品質開發一個系統。本研究針對軟體的複雜度品質進行探討，並應用程式碼氣味與反面模式以發展複雜度的Rule-based偵測模式，並結合程式碼層級與設計層級的指標綜合分析系統之複雜度品質。最後，根據系統品質評核的結果提出重構建議。此外本研究將運用本體論以建立品質評核的知識庫，並且實作一套 Web-based的軟體設計品質評核系統，再透過一個專案來展示系統，並用五個測試案例驗證該系統的功能與效益。;In response to the rapid development of information, software developers need to complete software within a shorter time to enhance their competitiveness. As a result, the software development team cannot go through a complete system analysis process before implementing the software. Therefore, software developers can use reverse engineering (SRE) tools to quickly produce system design files for the developed software. However, if the hastily implemented code brings to poor/bad quality, it will consequently affect the quality of UML diagrams produced by SRE. To solve this problem, this study develops a system for the quality of design documents produced by SRE and the code of the project. This study focuses on the complexity quality of the software, and applies the code smells and anti-patterns to develop the rule-based detection for complexity, and also combines the code-level and design-level metrics to comprehensively analyze the system’s complexity quality. Finally, refactoring suggestions are made based on the results of the quality assessment for the system. In addition, this study uses ontology to build a knowledge base for quality assessment and implements a Web-based software design quality evaluation system. Furthermore, this study demonstrates the system through a project and uses five test cases to verify the system’s accumulative performance and benefits.
