The universal model is a great experiment so far. Unfortunately, it started from the software model and evolved toward an ECS design. Because of that, we have a bit of legacy in the code, which makes it harder to evolve and be efficient with it. Those next few PRs in that realm aim to purify the universal model into a layered ECS design. At its core, it should be Gradleless and deal only with entity-component-system. The layers above would introduce observability over component change, helper for building entities, and dynamic class generation. For example, we should consider component values to be fixed or provided (Gradle provider) because of Gradle. However, coding using the universal model should not rely on the provided value in the components. It should strictly be a way to provide a value + implicit dependencies + source.
This first step chips away the component registry from the model node.
The universal model is a great experiment so far. Unfortunately, it started from the software model and evolved toward an ECS design. Because of that, we have a bit of legacy in the code, which makes it harder to evolve and be efficient with it. Those next few PRs in that realm aim to purify the universal model into a layered ECS design. At its core, it should be Gradleless and deal only with entity-component-system. The layers above would introduce observability over component change, helper for building entities, and dynamic class generation. For example, we should consider component values to be fixed or provided (Gradle provider) because of Gradle. However, coding using the universal model should not rely on the provided value in the components. It should strictly be a way to provide a value + implicit dependencies + source.
This first step chips away the component registry from the model node.