Broco
November 22, 2020, 9:08pm
1887
Yeah. You might be interested in a post I wrote earlier about my theory for why emulators have rollback more often (or better) than games running direct on the OS:
This seems to be a pattern with rollback netcode implementations. An emulator supporting rewind gives you a very clean surface area, mathematically, to perform rollback in a principled way, whereas if you try to do it in the game engine itself you run into a world of pain as various quirks in the engine get in the way at every turn.
Another thing that this reminds me of is the DkS1 PC version got a hack within days of release giving it quite good support for higher resolutions, whereas the official version had a fixed resolution after From spending like a year porting. It’s because the hacker did it as a surface-level translator of the DirectX command stream, which was not only the only feasible way for a hacker to do it, but also happened to be the easy way to do it. The funniest thing about this one is that there was nothing preventing From from doing the same thing, but they just didn’t have enough experience at porting to realize it.
My one game industry job circa 2006 was in a Bluepoint-style (except lower budget levels) specialized porting contractor that turned out acceptable-quality PC ports from undocumented codebase dumps with a 4-month development cycle, and our secret sauce was likewise a collection of surface-level generic wrappers for input and graphics that could be applied to any game.
7 Likes