REST Again? Oh, Come On!
summary
The talk argues that real-world REST APIs often violate REST constraints and lists concrete problems: inconsistent use of GET request bodies, unclear PATCH semantics, caching bugs causing stale data, chaotic versioning, misuse of HTTP status codes, and N+1 roundtrips increasing load. It cites tooling and industry choices — Google using gRPC, teams using OpenAPI/Swagger, Yandex encountering caching bugs — and recommends pragmatic alternatives such as gRPC or JSON‑RPC while retaining OpenAPI/Swagger for discoverability instead of enforcing textbook REST.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 6:25What is REST
- 11:43REST constraints
- 18:11HTTP and URI issues
- 24:50N+1 and caching
- 30:21Protocol independence
- 33:38Stateful vs Stateless
- 36:10When to use REST
- 39:42Summary & Q&A
related talks
01h 01mREST - The ignored parts
22mAre REST APIs Still Relevant Today?
42mLet’s Have Some Effective REST!
21mThe REST API Myth
43mTUTORIAL: We've "REST"ed long enough -- let's have an API adventure!
01h 39mLet’s have some Effective REST
51mREST beyond the obvious – API design for ever evolving systems
51mTop 10 Rest API Design Pitfalls