Nikhil Patnaik

Nikhil Patnaik, Andrew C. Dwyer, Joseph Hallett et al.

Producing secure software is challenging. The poor usability of security APIs makes this even harder. Many recommendations have been proposed to support developers by improving the usability of cryptography libraries and APIs; rooted in wider best practice guidance in software engineering and API design. In this SLR, we systematize knowledge regarding these recommendations. We identify and analyze 65 papers spanning 45 years, offering a total of 883 recommendations.We undertake a thematic analysis to identify 7 core ways to improve usability of APIs. We find that most of the recommendations focus on helping API developers to construct and structure their code and make it more usable and easier for programmers to understand. There is less focus, however, on documentation, writing requirements, code quality assessment and the impact of organizational software development practices. By tracing and analyzing paper ancestry, we map how this knowledge becomes validated and translated over time.We find evidence that less than a quarter of all API usability recommendations are empirically validated, and that recommendations specific to usable security APIs lag even further behind in this regard.

Joseph Hallett, Nikhil Patnaik, Benjamin Shreeve et al.

Does the act of writing a specification (how the code should behave) for a piece of security sensitive code lead to developers producing more secure code? We asked 138 developers to write a snippet of code to store a password: Half of them were asked to write down a specification of how the code should behave before writing the program, the other half were asked to write the code but without being prompted to write a specification first. We find that explicitly prompting developers to write a specification has a small positive effect on the security of password storage approaches implemented. However, developers often fail to store passwords securely, despite claiming to be confident and knowledgeable in their approaches, and despite considering an appropriate range of threats. We find a need for developer-centered usable mechanisms for telling developers how to store passwords: lists of what they must do are not working.