Domain-Driven Design (DDD) With F# and EventStore

The post describes a methodology for implementing Domain-Driven Design with F# and EventStore. It argues that functional programming and DDD share the same aim of making domain knowledge explicit, and it uses event sourcing to show that alignment.

Reading notes

#

• The example is a spike in early progress and uses F# and EventStore to explore a more idiomatic functional approach.
• The author contrasts functional and object-oriented paradigms and says F# changed his view of language value.
• DDD and functional programming are presented as sharing the goal of formalizing domain knowledge for automation and integration.
• The sample is based on SimpleCQRS and related F# implementations, but tries to move further away from object-oriented habits.
• The use of classes is limited to serialization and EventStore integration, and the choice is justified as a way to avoid technology for its own sake.
• The InventoryItem module contains state, commands, events, assertions, and command execution in a single F# source file.
• Discriminated unions are used so that all commands and events are explicitly handled with static verification.
• The identity value is kept outside the domain module because the domain logic does not need it.
• The Aggregate module defines an abstraction for aggregates and a persistent command handler.
• Persisted events are applied with fold.
• The IntegrationTests module wires the aggregate module to EventStore and serialization.
• The EventStore module is specific to InventoryItem in this version, but the author says it could be generalized by generalizing the serializer.
• The summary claims the F# version is more concise, more encapsulated, more declarative, more statically verified, and more readable.
• The author says explicit factoring by functions replaces class hierarchies and keeps the informational core at the center with infrastructure around it.
• Future work mentioned includes projections, read models, a service layer, more complex domains, domain services, complex invariants, and types for domain state.
• The author also wants to test the approach outside event sourcing and explore computation expressions for persistent workflows.