On the role of contectual 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 exercisein the area of evidence-based trust models for ubiquitous computingenvironments has given us an impulse to take a closer look at thedefinition(s) of privacy in the Common Criteria, which we thentranscribed in a more formal manner. This lead us to a further reviewof unlinkability, and revision of another semi-formal model allowingfor expression of anonymity and unlinkability -- the Freiburg PrivacyDiamond. We propose new means of describing (obviously only observable)characteristics of a system to reflect the role of contexts forprofiling -- and linking -- users with actions in a system. We believethis approach should allow for evaluating privacy in large data sets.