RESTRuler: Towards Automatically Identifying Violations of RESTful Design Rules in Web APIs (2402.13710v1)

Abstract: RESTful APIs based on HTTP are one of the most important ways to make data and functionality available to applications and software services. However, the quality of the API design strongly impacts API understandability and usability, and many rules have been specified for this. While we have evidence for the effectiveness of many design rules, it is still difficult for practitioners to identify rule violations in their design. We therefore present RESTRuler, a Java-based open-source tool that uses static analysis to detect design rule violations in OpenAPI descriptions. The current prototype supports 14 rules that go beyond simple syntactic checks and partly rely on natural language processing. The modular architecture also makes it easy to implement new rules. To evaluate RESTRuler, we conducted a benchmark with over 2,300 public OpenAPI descriptions and asked 7 API experts to construct 111 complicated rule violations. For robustness, RESTRuler successfully analyzed 99% of the used real-world OpenAPI definitions, with some failing due to excessive size. For performance efficiency, the tool performed well for the majority of files and could analyze 84% in less than 23 seconds with low CPU and RAM usage. Lastly, for effectiveness, RESTRuler achieved a precision of 91% (ranging from 60% to 100% per rule) and recall of 68% (ranging from 46% to 100%). Based on these variations between rule implementations, we identified several opportunities for improvements. While RESTRuler is still a research prototype, the evaluation suggests that the tool is quite robust to errors, resource-efficient for most APIs, and shows good precision and decent recall. Practitioners can use it to improve the quality of their API design.