No REST for the Wicked. Building a GraphQL
summary
CargoSense replaces bespoke REST endpoints with GraphQL, implementing a full spec in Elixir as Absinthe on Hex and running it nine months in production. The demo shows exact-response typing, schema introspection, compile-time schema checks, and planned query compilation plus complexity analysis to limit cost. Real-time uses Phoenix channels for subscriptions and chunked responses, with experimental defer and stream directives to send core data fast and stream the rest.
AI-written summary · about
chapters (15)
- 0:00Introduction to Absinthe
- 1:02The Need for Composability
- 2:25Limitations of REST
- 5:15Introduction to GraphQL
- 6:58Data Validation Issues
- 9:02GraphQL Implementation
- 10:54Scalars and Type Systems
- 13:07Introspection and Tools
- 16:06Advanced Caching
- 17:17Experimental Features
- 19:41Absinthe Architecture
- 22:31Execution Model Changes
- 24:22Future Absinthe Roadmap
- 27:08Real-time Capabilities
- 29:38Q&A Session
related talks
45mA shot of Absinthe: from zero to GraphQL in 40 minutes
23mAugmenting a Legacy REST API with GraphQL
46mGraphQL 101 - Stop RESTing, Start Querying
59mGoodbye REST; Hello GraphQL
33mGraphQL Will Do To REST What JSON Did To XML
36mGraphQL Will Do To REST What JSON Did To XML
36mGraphQL on Rails - Creating Living APIs for a Data-hungry World
29mFrom REST to GraphQL
