Modeling stories for conceptual model validation

• Autor(a) principal: Braga, Bernardo Ferreira Bastos
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/9867
• topic: Ciência da Computação; Arte de contar histórias; Ontologia; Sistemas de recuperação da informação; 004

Conceptual modeling is a challenging activity and assessing the quality of conceptual models is key to ensure that they may be used effectively as a basis for understanding, agreement and construction of information systems. A model can be assessed for different types of model quality and in this work we focus on the accuracy of an ontology-based conceptual model in characterizing the conceptualization it is supposed to represent. Validating the accuracy of a model involves understanding the admissible worlds states implied by the model and comparing that to the world states deemed admissible in the domain conceptualization. Previous efforts towards ontology-based conceptual model validation have created a model simulator that allows modelers to be confronted with the consequences of their modeling decisions. The model simulator generates sequences of snapshots of model instances, revealing the dynamics of object creation, change and destruction. Even though these efforts contribute to model assessment, they can be hard to understand and use and this work improves the approach using a mix of informal and formal storytelling. Stories have always been used as means of communicating complex affairs and we argue that they may be used effectively to assess models and reveal modeling decisions. This dissertation proposes an approach to assess conceptual models by creating narratives about a subject domain. These narratives exemplify how concepts of the conceptual model are employed in context. To use them in the existing model simulator, the natural language narratives are formalized as abstract stories using a specification language we define. These abstract stories are then used to guide the model simulation, generating instance diagrams. The natural language narrative is used to provide an intuitive understanding of the meaning of concepts. Comparing Natural Language Narratives to object diagrams that show the instantiation of the formal model allows one to understand how concepts are formalized. Contrasting these guided simulations with the intended conceptualization is the basis for model assessment in this approach.