Highlighting model elements to improve OCL comprehension

• Autor(a) principal: Sales, Maria Pedro dos Santos
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10071/22061
• topic: OCL; UML; Highlighting model elements; OCL comprehension; Destaque de elementos do modelo; Compreensão do OCL

Models, metamodels, and model transformations play a central role in Model-Driven Development (MDD). Object Constraint Language (OCL) was initially proposed as part of the Unified Modeling Language (UML) standard to add the precision and validation capabilities lacking in its diagrams, and to express well-formedness rules in its metamodel. OCL has several other applications, such as defining design metrics, code-generation templates, or validation rules for model transformations, required in MDD. Learning OCL as part of a UML course at the university would seem natural but is still the exception rather than the rule. We believe that this is mainly due to a widespread perception that OCL is hard to learn, as gleaned from claims made in the literature. Based on data gathered over the past school years from numerous undergraduate students of di↵erent Software Engineering courses, we analyzed how learning design by contract clauses with UML+OCL compares with several other Software Engineering Body Of Knowledge (SWEBOK) topics. The outcome of the learning process was collected in a rigorous setup, supported by an e-learning platform. We performed inferential statistics on that data to support our conclusions and identify the relevant explanatory variables for students’ success/failure. The obtained findings lead us to extend an existing OCL tool with two novel features: one is aimed at OCL apprentices and goes straight to the heart of the matter by allowing to visualize how OCL expressions traverse UML class diagrams; the other is intended for researchers and allows to compute OCL complexity metrics, making it possible to replicate a research study like the one we are presenting.