How to drill down to ReST APIs: Resource harvesting with a pattern tool

REST has become a popular architectural style among service providers. It is considered as an easy way to design and consume Web services. REST can be realized as using HTTP PUT, POST, and GET operations. However, the focus on the implementation technique often leads to ignoring the original REST constraints and definitions proposed by R. Fielding. Thus, this way of thinking might result in misuse of REST. In addition, less emphasis is put on designing good REST APIs, which indeed is not a trivial task. In this paper, we propose a questionnaire-based method, motivated by speech-act theory, to harvest the essential API concepts and their relationships from the functional service requirements. We present our pattern-based implementation of the method. We define a reusable REST API pattern, which can be applied in different contexts to produce an API model according to the REST design principles. The main benefit of the questionaire-based method is on shifting the focus from an operation-based to a resource-oriented mindset. The paper includes an empirical evaluation.