Domain Modeling Made Functional – Scott Wlaschin

Discover how to create a robust domain model using functional programming principles and the type system, avoiding jargon and focusing on concepts and rules, to ensure correctness and maintainability.

Key takeaways

Domain modeling should be done with natural language and words used by experts in the domain.
Avoid using programmer jargon and focus on concepts and rules.
Use types to model constraints, such as email addresses and verified emails.
Represent options as a sum type or discriminated union.
Composition is key to building complex systems, and types are used to build bigger types from smaller ones.
Functional programming can be used to build domain models, and the type system can help catch errors at compile-time.
Modeling business rules as types can help ensure that errors are caught early and that the design is correct.
Domain modeling is about building a shared mental model, and using types and functions can help create a shared understanding.
Avoid null values and verbose code, and use immutable objects to ensure correctness.
Don’t mix up concepts, such as email addresses and phone numbers, and use different types to represent them.
Use control flow types to represent conditional logic.
Represent missing values as a choice type.
Use executable code to model the business logic, rather than using UML or other diagramming tools.
The type system can catch errors that wouldn’t be caught by a natural language description.
Domain modeling is a subjective process, and the goal is to create a model that is understandable and maintainable.

Domain Modeling Made Functional – Scott Wlaschin

More talks