Add schematics RFC

rfcs/-schematics.md

@@ -0,0 +1,181 @@
+# Feature Name: `schematics`
+
+## Summary
+
+The goal of this RFC is to facilitate and standardize two-way conversion and synchronization between ECS worlds.
+The approach described here aims to be more user friendly than just handcoded synchronization systems, but more flexible than a prefab system.
+This design is buitl on top of the [many worlds RFC][many_worlds].
+
+## Motivation
+
+When devolping an editor for bevy, the information visible should not necessarily be in the same format that most runtime systems operate on.
+The runtime representation is often optimized for efficiency and use in those systems, while the information displayed in the editor should be optimized for
+easy understanding and stability agains changes.
+We propose to address this by adding another default world to the ones described in the [many worlds RFC][many_worlds] that we call the "schematic world",
+and which contains the information used in the editor and its scene file formats.
+
+The main problem that arises from this approach is the synchronisation between the main and schematic worlds, so the RFC focuses on this aspect.
+Similar problems, where data needs to kept synchronized between worlds, might also appear in other areas.
+One example for this is the "render world", for which the extract phase could be adapted to use schematics.
+Another example could be a "server world" in a multiplayer game, that only contains information which is shared with other clients and the server.
+Finally the approach chosen here would help in implementing a "prefab" system.
+
+## User-facing explanation
+
+### The `CoreWorld::Schematic`
+
+### The `SchematicQuery`
+
+The main interface added by this RFC is the `SchematicQuery`, which you will usually only use in systems that run on `CoreWorld::Schematic`.
+We call such systems "schematics".
+A `SchematicQuery` works almost the same as a normal `Query`, but when iterating over components will additionaly return `SchematicCommands` for the component queried.
+This can be used to insert and modify components on the corresponding entity on `CoreWorld::Main` as well as spawn new entities there.
+The entities spawned in this way can not be modified by `SchematicCommands` from other systems, but will be remembered for this system similar to `Local` system parameters.
+
+A typical schematic will look like this:
+
+```rust
+#[derive(Component)]
+struct SchematicA(u32, Entity);
+
+#[derive(Component)]
+struct MainA(u32);
+
+#[derive(Component)]
+struct MainAChild(Entity);
+
+/// Translates a `SchematicA` to a `MainA` as well as a child entity that has a `MainAChild`.
+/// The system can contain any other parameter besides the schematic query
+fn schematic_for_a(query: SchematicQuery<A, With<Marker>>, mut some_resource: ResMut<R>) {
+    for (a, commands) in query {
+        some_resource.do_something_with(a);
+        // You can modify components
+        commands.insert_or_update(MainA(a.0));
+        // And spawn other entities.
+        // This will only spawn an entity on the first execution. It will remember the entity
+        // by the label "child" so `child_commands` will operate on the same entity the next
+        // time this system is run.
+        commands.spawn_or_select_child("child", |child_commands| {
+            // Entity references within the schematic world need to be mapped in this way.
+            // (This might not be needed depending on the implementation of many worlds)
+            let entity = child_commands.map_entity(a.1);
+            child_commands.insert_or_update(MainAChild(entity));
+        });
+    }
+}
+```
+
+The `SchematicQuery` will automatically only iterate over components that were changed since the system last ran.
+Other than a normal query, the first generic argument can only be a component, so tuples or `Entity` are not allowed as the first argument.
+
+The main methods of `SchematicCommands` are:
+* `insert_or_update<B: Bundle>(bundle: B)`
+* `spawn_or_select_child<L: SchematicLabel, F: FnOnce(SchematicCommands)>(label: L, F)`
+* `spawn_or_select<L: SchematicLabel, F: FnOnce(SchematicCommands)>(label: L, F)`
+* `despawn_if_exists<L: SchematicLabel>(label: L)`
+* `remove_if_exists<B: Bundle>()`
+
+### The `default_schematic` system
+
+Since with many components you want to just copy the same component from the schematic to the main world, a `default_schematic` is provided.
+The implementation of this is just:
+
+```rust
+fn default_schematic<A: Clone>(query: SchematicQuery<A>) {
+    for (a, commands) in query {
+        commands.insert_or_update(a.clone());
+    }
+}
+```
+
+### Adding schematics to your app
+
+`Schematic`s can be added to schedules like any other system
+```rust
+schedule.add_system(schematic_for_a);
+```
+
+Usually you will want to syncronize between the app's `SchematicWorld` and `CoreWorld::Main`.
+In this case you can use
+
+```rust
+app.add_schematic(default_schematic::<A>);
+```
+
+## Implementation strategy
+
+* `SchematicQuery<C, F>` is basically `(Query<C, (F, Changed<C>)>, Local<HashMap<Entity, SchematicData>>, AppCommands)` where `SchematicData` is something like
+  ```rust
+      struct SchematicData {
+          corresponding_entities: HashMap<SchematicLabelId, Entity>
+      }
+  ```
+* `SchematicCommands` is something like
+  ```rust
+      struct SchematicCommands<'a> {
+          app_commands: &'a mut AppCommands,
+          schematic_data: &'a mut SchematicData,
+          current_label: Option<SchematicLabelId>,
+      }
+  ```
+
+## Drawbacks
+
+* Adds another data model that programmers need to think about.
+* The design may not be very performant and use a lot of extra memory
+* Need to add additional code to pretty much any component, even if it is just `app.add_schematic(default_schematic::<C>)`
+* It is not possible to check invariants, e.g. synchronization should be idempotent, schmatics shouldn't touch same component, ...
+
+## Rationale and alternatives
+
+This design
+* Integrates well into ECS architecture
+* Is neutral with regard to usage
+* Can use existing code to achive parallel execution
+* Can live alongside synchronization systems that don't use `SchematicQuery`
+
+The problem could also be solved by "prefabs", i.e. scenes that might expose certain fields of the components they contain.
+But this would be a lot more restrictive the "schematics" described here.
+What might make more sense is to implement prefabs atop schematics.
+
+This design does not allow the schematic world to behave sensibly for use in the editor by itself.
+It would need something like [archetype invariants](https://github.com/bevyengine/bevy/issues/1481) additionally.
+
+## Prior art
+
+*TODO: Compare with Unity's system of converting game objects to ECS components*
+
+See https://github.com/bevyengine/bevy/issues/3877
+
+Discuss prior art, both the good and the bad, in relation to this proposal.
+This can include:
+
+- Does this feature exist in other libraries and what experiences have their community had?
+- Papers: Are there any published papers or great posts that discuss this?
+
+This section is intended to encourage you as an author to think about the lessons from other tools and provide readers of your RFC with a fuller picture.
+
+Note that while precedent set by other engines is some motivation, it does not on its own motivate an RFC.
+
+## Unresolved questions
+
+- What parts of the design do you expect to resolve through the RFC process before this gets merged?
+- What parts of the design do you expect to resolve through the implementation of this feature before the feature PR is merged?
+- What related issues do you consider out of scope for this RFC that could be addressed in the future independently of the solution that comes out of this RFC?
+
+## \[Optional\] Future possibilities
+
+Think about what the natural extension and evolution of your proposal would
+be and how it would affect Bevy as a whole in a holistic way.
+Try to use this section as a tool to more fully consider other possible
+interactions with the engine in your proposal.
+
+This is also a good place to "dump ideas", if they are out of scope for the
+RFC you are writing but otherwise related.
+
+Note that having something written down in the future-possibilities section
+is not a reason to accept the current or a future RFC; such notes should be
+in the section on motivation or rationale in this or subsequent RFCs.
+If a feature or change has no direct value on its own, expand your RFC to include the first valuable feature that would build on it.
+
+[many_worlds]: https://github.com/bevyengine/rfcs/pull/43