Improving API Documentation Comprehensibility via Continuous Optimization and Multilingual SDK
share
Optimizing and maintaining up-to-date API documentation is a challenging problem for evolving OpenAPIs. In this poster, we propose a data-driven continuous optimization solution and multilingual SDK generation scheme to improve the comprehensibility of API documentation. We compute the correlation between API integrity and API trial success rate. Based on this, we partition the API to ensure that each API has a correct optimization direction. Then, we propose a fine-grained(i.e., parameter level) continuous optimization solution to annotate problems in API documents in real-time. Based on the above resolutions, we can provide theoretical analysis and support for the optimization and management of API documents. Finally, we explore the crucial challenges of OpenAPIs and introduce a tailored solution, TeaDSL, a multi-language SDK solution for all OpenAPI gateways. TeaDSL is a domain-specific language that expresses OpenAPI gateways, generating SDKs, code samples, and test cases. The experiments evaluated on the online system show that this work's approach significantly improves the user experience of learning OpenAPIs.
READ FULL TEXT