Please note: This PhD defence will take place online.
Justin Tracey, PhD candidate
David R. Cheriton School of Computer Science
Supervisor: Professor Ian Goldberg
As daily life becomes increasingly subject to surveillance economies and surveillance states, so do privacy enhancing technologies (PETs) become increasingly important to those who wish to live unmediated by these mass data collection practices. While improvements to the design and implementation of PETs have allowed more people than ever to take advantage of these tools, the research and development practices surrounding PETs require the work of domain experts—with their own biases, values, motivations, and awareness—doing their best to accommodate the needs and desires of users. As a result, an inevitable gap forms between what is built by those who have the skills and resources to develop these technologies, and what is needed by those who can only use what is available.
In this thesis, we examine techniques for lowering barriers to entry for research and development contributions to PETs, so that users who wish to make such contributions are more readily able to do so. Towards this end, we use as a concrete example the Tor project, and how it interfaces with current research and development practices to demonstrate three methods of lowering such barriers: (i) a set of tools and techniques for conducting statistically sound Tor experiments, (ii) an analysis of the viability of Tor as a means of providing simple metadata protection in messaging apps, and (iii) an investigation into the effect on new contributors when porting a codebase written in a programming language without memory safety to the memory-safe Rust language, as Tor is doing. We find that, using these methods, the barriers to entry can be reduced, but that considerable future work still remains in this vein.