Privacy Distillation @ Best RESG Research 2014
Some interesting questions were discussed following my presentation, including:
- How does the distillation process cope with the overall mobile software eco-system? At the moment we have only considered the peer-to-peer information flows between the end users of the mobile application. However, it should be possible to use the Privacy Facets Framework to consider the information, information flows and actors in the overall mobile software eco-system. Of course some extensions will be required, for example to capture factors such the legal and regulatory aspects of privacy associated with the places in which in the information, information flows and actors operate.
- Can distillation also be used for other requirements of the application? We don't advocate using distillation for eliciting functional requirements because there are already well established methods for this. Instead we see distillation as a complementary technique that can augment functional requirements elicitation by focussing on 'more tacit' types of requirements like privacy.
- Does focussing on individual privacy sensitive contexts hide some of the 'bigger picture' of the system, based on the interactions between different contexts? This is a valid concern and is not currently addressed in the distillation process. One way to deal with this problem might be to look at ways of using the information flow modelling to relate different privacy sensitive contexts to each other. These relationships could be highlighted using a visualisation tool to help the analyst to explore the interactions and keep track of the 'big picture'.
- What types of privacy requirements were discovered in real-world applications? We used the distillation process to analyse two types of mobile applications, Mobile Facebook and a mobile location sharing app. When we carried out the empirical study of Mobile Facebook users in 2008/2009, Facebook did not have any functionality to group friends into lists and specify privacy preferences using these lists. Our analysis highlighted the need for such functionality in order to address users' privacy concerns.