Specifying privacy requirements with goal-oriented modeling languages
Mariana Peixoto, Carla Silva · 2018
Context: Privacy of personal data is a growing concern regarding users of software systems. In this sense, the literature reports that in order to avoid privacy breaches, there must be systematic approaches to specify privacy requirements from the early activities of software development. Objective: Motivated by this situation, this paper presents a framework of privacy modeling capabilities that must be addressed by requirements modeling languages to better support privacy specification. The capabilities will be used to compare three goal-oriented modeling languages (i*, NFR-Framework and Secure-Tropos). Method: The framework was created with basis on a conceptual foundation and a conceptual model of privacy built from an analysis of a standard, a regulation, guidelines and other bibliographical sources related to privacy. A health care example is used to illustrate how the framework can be used to compare the chosen modeling languages. Results: Fourteen privacy modeling capabilities were defined in the framework and it was observed that the analyzed modeling languages do not fully support them. Conclusions: The proposed framework contributes towards the consolidation of a privacy conceptual foundation that can be used to evaluate modeling languages for privacy in Requirements Engineering. The comparison performed by using this framework indicates Secure-Tropos as the most complete language to model privacy among the analyzed goal-oriented modeling languages.