On the role of contextual information for privacy attacks and classification

Many papers and articles attempt to define or even quantify privacy, typically with a major focus on anonymity. A related research exercise in the area of evidence-based trust models for ubiquitous computing environments has given us an impulse to take a closer look at the definition(s) of privacy in the Common Criteria, which we then transcribed in a bit more formal manner. This lead us to a further review of unlinkability, and revision of another semi-formal model allowing for expression of anonymity and unlinkability – the Freiburg Privacy Diamond. We propose new means of describing (obviously only observable) characteristics of a system to reflect the role of contexts for profiling – and linking – users with actions in a system. We believe this approach should allow for evaluating privacy in large data sets.