JBCT

Structure Java backend code using monadic patterns by adopting four return kinds—T, Option<T>, Result<T>, and Promise<T>—with typed failures. This ensures predictable, testable data flow by composing logic through six structural patterns like Leaf, Sequencer, and Fork-Join.

The parse-don't-validate pattern for Java value objects constructs valid states at creation time, preventing invalid instances from existing. This approach replaces runtime validation checks with guaranteed valid domain models within structured use cases.

Organize a Java project into vertical slices with domain and shared value objects, defining clear adapter boundaries. Use case architecture requires validated inputs, nested steps, and growing-context data to ensure clear, predictable data flow.

Automate Java use case generation and compliance checks using the jbct-coder and jbct-reviewer subagents. The jbct-coder generates complete use cases, while jbct-reviewer performs quality checks and refactoring against deterministic backend patterns.

The six structural patterns for composing Java backend logic are Leaf, Sequencer, Fork-Join, Condition, Iteration, and Aspects. These patterns define how nested steps interact within use cases to maintain predictable data flow and typed failures.

Use Result and Option types instead of exceptions when writing predictable Java backend code that requires typed failures. This monadic approach makes invalid states unrepresentable and ensures testable use case logic without exception-driven control flow.