Your API is a UI
summary
Claim: treat an API as a user interface and design for the humans and machines that consume it. Use documented user stories and consistent style guides, test with unit and functional suites and record and replay production interactions so regressions are visible. Push metrics and logging via statsd into graphite and alerting, version in the URL, and generate IODocs from annotated docblocks as part of the build. DataSift examples include an ops wall with colored alerts and automated docs that actually reduce support load.
AI-written summary · about
chapters (24)
- 0:00Intro
- 2:20What is a UI?
- 5:10What does your API do?
- 6:25What can users do?
- 9:23API Hierarchy of Needs
- 9:53Usability
- 10:44Consistency
- 12:58Patterns
- 13:57Information Hierarchy
- 15:29Abstraction
- 17:12Simplicity
- 17:51Extensibility
- 20:40Testing
- 22:51Metrics
- 24:17Logging
- 26:08Versioning
- 28:43Feedback
- 30:30Proficiency
- 31:57Living Documentation
- 34:23Support / Community
- 36:12Creativity
- 37:56On Pragmatism
- 39:57Your API is You!
- 40:30Questions?
related talks
27mUser focused API Design
23mDesigning APIs Both Humans and AI Love
59mHandy Tools for Designing Great Web APIs
38mZen And The Art Of Automanually Creating API Documentation - An Inquiry Into Process
35mPrinciples of Awesome APIs and How to Build Them
24mDesigning the APIs for an Internal Set of Services
35mBuilding Your API for Longevity
42mREST API design good practices