The Group-centric Secure Information Sharing (g-SIS) family of models has been proposed for modeling environments in which group dynamics dictate information-sharing policies and practices. This is in contrast to traditional, dissemination-centric sharing models, which focus on attaching policies to resources that limit their flow from producer to consumer. The creators of g-SIS speculate that it may not be strictly more expressive than dissemination-centric models, but that it nevertheless has pragmatic efficiency advantages in group-centric scenarios [12]. In this paper, we formally and systematically test these characteristics of an access control system's suitability for a scenario - expressiveness and cost - to evaluate the capabilities of dissemination-centric systems within group-centric workloads. We show that several common dissemination-centric systems lack the expressiveness to meet all security guarantees while implementing the wide range of behavior that is characteristic of the g-SIS models, except via impractical, convoluted encodings. Further, even more efficient implementations (admissible under relaxed security requirements) suffer from high storage and computational overheads. These observations support the practical and theoretical significance of the g-SIS models, and provide insight into techniques for evaluating and comparing access control systems in terms of both expressiveness and cost.

The rationale for a software system captures the designers' and developers' intent behind the decisions made during its development. This information has many potential uses but is typically not captured explicitly. This paper describes an initial investigation into the use of text mining and parsing techniques for identifying rationale from existing documents. Initial results indicate that the use of linguistic features results in better precision but significantly lower recall than using text mining.