Proposta de dissertação do MEI Título: Infering likely data invariants Proponente(s): Carla Ferreira Créditos: 42 ECTS Área científica: Software Construction and Analysis Início preferencial: Qualquer semestre URL: Já estão em curso trabalhos preliminares executados pelo alunos: Breve descrição: Developing distributed applications is challenging, so it is crucial to provide developers supporting techniques, methodologies, and tools. At NOVA LINCS we contributed to research milestones in the verification of distributed applications. These verification tools rely on some sort of formal specification that describes abstractly the application code. However, the fact that the analysis is done at specification level is a major barrier for the acceptance of these tools by developers. Not only developers have to be knowledgeable in two very distinct languages (one for programming and one for specifying), they also have to write and maintain the source code and the specification. This thesis aims to take a first step in decreasing the specification effort of developers. The goal of the thesis is to study and evaluate current approaches for invariant inference (e.g. Daikon) and use them to derive a set of restricted type of invariants (e.g. referential integrity, numeric invariants). Observações: This proposal is part of projects EU LightKone and FC&T Precise.