Semantic Driven Modeling in distributed systems
summary
Proposes using semantic metadata plus Roslyn C# code-generation to generate JSON converters that map evolving public models at the HTTP boundary so consumers never need recompilation or restart; supports event-sourced archives and IoT device variants by converting old JSON slices to the current model. Implementation generates single, tree-aware converters, benchmarks show ~6.1 µs for mapped serialization vs 9.7 µs native and reflection-based approaches are about 1000x slower, metadata loads at runtime.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 2:34Agenda
- 5:51Scenarios
- 7:32Updating a public model
- 9:19Upgrading to version 2
- 11:09IoT devices
- 12:39Multiple reading models
- 15:51Multiple reading scenarios
- 17:39Mappings
- 19:49Semantic Analysis
- 22:18Domain Specific Language
- 25:02ubiquitous language
- 27:07writing down your model
- 27:32domain elements
- 32:32negotiate versioning
- 37:58request different version
- 39:45order controller
- 46:58graphql
- 48:23conversion
- 54:58feedback
related talks
25mDomain-specific model differencing in visual concrete syntax
34mUsing CRDTs Beyond Text Editors
28mWhat You See Is What It Does: A Structural Pattern for Legible Software
43mC4 models as code
05mModel Driven Software: Putting the Cart Before the Egg
14mAutomating the Synthesis of Recommender Systems for Modelling Languages
45mUsing modern .NET technologies to decompose a monolith into micro-services
47mDomain driven redesign