Etna brings rigorous testing to property-based evaluation across multiple languages

Property-based testing has become a powerful technique in functional programming communities, with numerous frameworks and tools now available. Yet developers often struggle to understand which approaches deliver the best results, and the field lacks systematic comparisons to guide these decisions.

The challenge#

The landscape of property-based testing (PBT) tools is crowded, offering different strategies for generating test inputs and validating code behavior. While researchers have developed many innovative techniques, the academic literature offers little guidance on how these methods compare in practice or which ones suit different use cases. This leaves both newcomers and experienced practitioners making choices largely based on intuition or anecdotal experience.

Etna's approach#

Researchers have developed Etna, a systematic platform designed to bridge this gap through empirical evaluation. The system integrates several widely-used PBT frameworks alongside a collection of testing workloads drawn from real-world research examples. Its modular design allows teams to add new frameworks and test suites without extensive effort, while the platform handles the technical details of performance measurement and comparison.

Testing across languages#

To demonstrate its utility, the team used Etna to conduct experiments comparing popular property-based testing strategies implemented in five different programming languages: Rocq, Haskell, OCaml, Racket, and Rust. This cross-language approach reveals both language-specific insights and universal patterns in how different testing strategies perform.

The work addresses a clear need in the development community: turning the abundance of PBT innovation into actionable knowledge about best practices and engineering tradeoffs.