Not translated yet — showing English.
Who Sees What, and Which Machine Runs It
Two questions that sound like one. Who sees what is about a client: which slice of the room's state reaches which player. Which machine runs it is about a host: which process owns an entity, and which one owns it next. The word that runs them together is replication — in a game engine it usually names the first, and here it names the second.
| You mean | Read |
|---|---|
| which client receives which state, and how much of it | Visibility, with Data and Prediction |
| which machine owns the entity, and what happens when it dies | What Survives Losing a Host |
They are declared in two different places
Neither is configured at runtime, and they do not share a declaration.
| Declared on | Which names | |
|---|---|---|
| who sees what | the aspect — Data, Visibility | the visibility predicate, the object cap and its order, which neighbouring areas are visible, and the delivery mode |
| which machine runs it | the room type — Rooms | the authority mode, how far an external authority is trusted about an outcome, and the behaviour when the host drops |
The two also differ in what happens if you say nothing. An aspect with no visibility rule of its own is delivered in the shared packet, which is the default and is right for a small room. A room type that names no authority mode is refused — there is no default, because nothing can pick between our simulation and an external one on your behalf.