Hi all!
I am creating a procedural game (procedural/infinite terrain). No problem so far in single player. The terrain is chunk based, only a few chunks around the player are displayed of course, when getting far away from a chunk, it is getting destroyed and a new one is created.
The problems I have are limitation with network views. At least to my knowledge.
All child object of a terrain chunk (trees, rocks, enemies,…) and the chunk itself are destroyed when the chunk is far away from player.
But: if another player is still on that chunk, he will send networkViews update to all players. But as others do not have those object any-more in their game, errors pop-up as those networkViewsIDs were removed from their game instance.
Thus I wonder how I could manage networking this type of game.
-
Should I completely stop using network views? But then how to communicate objects/players/enemies updates?
-
Should I use only one master network view for all network updates? Using rather complex RPCs from that network view to update all object/player/enemy manually. In that case I could maybe filter incoming RPCs to check if the update received concerns an object present on ‘my’ game instance. Or even find a way to send from the server updates only to players that will need it.
I know that this question might be rather complex, and of course I am not asking for a code to do it, just a reflection on how you guys would try to solve this case.
Hoping my English was good enough and that I was clear!
Thank you all for your help