Identifying Personal Data Using Relational Database Design Principles

The European Union (EU) directive on personal data and resulting data protection legislation of EU member states require from data controllers, a notification of their activities to the appropriate supervisory authority. Included in this notification is also a description of the data or categories of data which are processed. Legislation in some EU member states (e.g. Slovenia) require that not only a description but also a concrete list of personal data attributes needs to be included in this notification. In such cases it is sometimes difficult to ascertain in concreto whether some collected attribute represents personal data (and should therefore be included in the list of attributes) or whether it is a non-personal attribute. Similarly, under the EU directive data subjects have various rights, including the right to access their data, and data controllers are sometimes faced with the problem of determining whether various data items constitute personal data. Further, the impending case in the European Court of Human Rights arising out of the decision of the UK case of Durant v Financial Services Authority (which narrowed the scope of personal data) has added some uncertainty to the interpretation of the EU directive. In view of the legal uncertainty regarding what constitutes personal data, this paper examines whether relational database design principles can be applied to identifying personal data. Using this approach, the paper explores various parallels between personal data identification and principles of relational database design. The paper thus makes a novel contribution to the ongoing uncertainty in data protection law. The paper also discusses the wider issue of applying computing/scientific principles to interpreting the law.