PhD Proposal: Designing Effective Property-Based Testing Frameworks

Property-Based Testing (PBT) is a widely adopted random testing methodology that combines user-defined logical propositions with automated test case generation. Over the past two decades, PBT has gained significant traction; today, virtually all major programming languages host multiple competing PBT frameworks, each innovating in interface design and underlying techniques.This proposal presents new quantitative and qualitative methods for comparing and designing Property-Based Testing frameworks optimized for effective bug discovery.
First, I introduce Etna, a multi-language benchmarking and evaluation platform for PBT that enables rapid experimentation and systematic comparison of the bug-finding capabilities of different PBT tools.Second, I present Deferred Binding Abstract Syntax (DBAS), a novel representation for properties that facilitates greater flexibility in PBT framework design.Building on these contributions, I propose the design of a domain-specific PBT framework for database testing, incorporating human-in-the-loop mechanisms to address domain-specific challenges more effectively.
Together, these efforts aim to advance the theory and practice of Property-Based Testing by providing more principled, flexible, and domain-aware testing methodologies.