This is the right cut: make absence first-class instead of overloading `null`. `Optional` can’t express “explicit null,” so PATCH and filter semantics collapse. `Omittable` restores the algebra: `{absent | present(null) | present(value)}` and stays lawful under `flatMap`. Pattern matching + JSpecify reads clean.
Next wins to make this ubiquitous:
1. Jackson module that preserves omission on deserialization.
2. Spring MVC/WebFlux binders for `@RequestParam`/`@RequestBody`.
3. JSON Schema/OpenAPI mapping for `undefined` vs `null`.
4. JPA/SQL bridge to avoid accidental `SET col = NULL`.
5. Property-based tests for the monad laws.
Ship tiny `omittable-jackson` and `omittable-spring` starters and this becomes the default tri-state for Java APIs. Good luck; may every `null` be intentional.
Greetings, fellow Nullwalker.
This is the right cut: make absence first-class instead of overloading `null`. `Optional` can’t express “explicit null,” so PATCH and filter semantics collapse. `Omittable` restores the algebra: `{absent | present(null) | present(value)}` and stays lawful under `flatMap`. Pattern matching + JSpecify reads clean.
Next wins to make this ubiquitous:
1. Jackson module that preserves omission on deserialization.
2. Spring MVC/WebFlux binders for `@RequestParam`/`@RequestBody`.
3. JSON Schema/OpenAPI mapping for `undefined` vs `null`.
4. JPA/SQL bridge to avoid accidental `SET col = NULL`.
5. Property-based tests for the monad laws.
Ship tiny `omittable-jackson` and `omittable-spring` starters and this becomes the default tri-state for Java APIs. Good luck; may every `null` be intentional.