Domain Experiments
summary
Argues for running rapid experiments directly on the domain model by combining example mapping, mob programming (3–4 engineers sweet spot) and business-oriented TDD. Demonstrates a live code‑storming cycle against a crowdfunding domain: starts with tests for “adult vs young vs criminal” borrowers, discovers KYC should gate the system, and settles a rule that projects under $10k auto‑start while higher ones require manual verification. Claims this finds boundary mistakes in days/weeks rather than months and yields throwaway discovery code then a refined design phase.
AI-written summary · about
chapters (21)
- 0:00Why software fails
- 1:20Learning too slowly
- 2:25Prototypes vs experiments
- 3:50Introducing domain experiments
- 5:13Knowledge crunching first
- 6:16Code storming phases
- 7:42Model decision records
- 8:23Design level storming limits
- 10:12Why invite business people
- 10:52Prerequisites for experiments
- 12:39Demo setup crowd sulcury
- 13:42Big picture session
- 14:48Example mapping process
- 16:59Code storming in action
- 21:58Iterating on the model
- 27:03First conceptual breakthrough
- 30:37Final model iterations
- 33:09Shifting left the learning
- 33:52Gloom Haven analogy
- 37:08Code as source of truth
- 38:12Q&A
related talks
49mDomain Experiments
51mDomain - Driven Transformation
01h 00mDomain-Driven Architecture at Scale
45mDon't Overdose on Domain-Driven Design
41mDomain Re discovery Patterns for Legacy Code
47mDomain-Driven Design: A Complete Example
38mWhat I talk about when I talk about Domain Driven design
55mDomain-Driven Design: back to the roots