An approach for synchronizing UML models and narrative text in literate modeling

Abstract

A major challenge in adopting UML in industrial environments is the lack of accessibility and comprehensibility of some diagram types by non-technical stakeholders. Literate Modeling improves comprehension of these diagrams by adding narrative text, but lacks good tool support for synchronizing model and text. This paper presents an approach for keeping model and text synchronized by effectively combining state-of-the-art natural language processing technology with OCL model querying. Thereby, consistency of element names in the UML model with their counterparts in the text is achieved by using text annotations to provide the semantic link. At a structural level, we propose an algorithm that checks element relationships in the UML model using a set of validation constraints when particular sentence characteristics are detected. An analysis of the runtime complexity shows the feasibility of including the proposed solution in one of today's CASE tools. © 2012 Springer-Verlag.