Enhancing Software Development Efficiency: An Automated Problem Frames to UML Transformation Method *

This article has 0 evaluations Published on
Read the full article Related papers
This article on Sciety

Abstract

Modern software engineering faces a critical challenge in bridging the gap between requirements analysis and system design phases. Traditional methods lack systematic mapping mechanisms, leading to the disconnect from requirements design and inefficient transformation processes. This research addresses this challenge by establishing an automated mapping mechanism between problem frames and a unified modeling language. We propose the problem frames to a unified modeling language method, which combines Jackson’s Problem Frames concept with model-driven engineering principles. The approach consists of three key components: an extended Problem Frames meta-model with enhanced semantic elements and transformation support structures, a graphical modeling platform developed using Eclipse Modeling Framework and Sirius following Meta-Object Facility three-layer architecture, and 24 transformation rules implemented using Atlas Transformation Language technology.Experimental validation on five representative open-source systems demonstrates significant improvements in software development efficiency. The method achieves a transformation accuracy of 93. 59%, representing an improvement 130% over the original approaches. 0 * The PF2UML tool presented in this work has been accepted for demonstration at FSE’25 Tool Demo Track. Practical implementation shows an average time reduction of 78.64% in end-to-end modeling processes, with a theoretical return on investment of 170. 8% and a practical average return on investment of 236.6%. The platform automatically generates four types of UML diagrams: conceptual class diagrams, use case diagrams, service diagrams, and system sequence diagrams. The PF2UML method successfully establishes seamless integration from requirements model-ing to system design, enhancing traceability while reducing manual expertise dependence and significantly improving consistency for complex interdisciplinary system development.

Related articles

Related articles are currently not available for this article.