once again thinking about how people say stuff like "make your own engine if you never wanna ship a game"

in one week working part time i created a 3D tilemap engine with chunking, texture atlasing, camera rotation controls and mouse picking targets without gpu readback

meanwhile we've been working on a content update for kitsune tails since august. most of that time was spent on game design, art, maps, and playtesting, not even any new code until february, and it's still not done

making a custom engine just for the game you're making is not the time consuming part. making the actual game *is*

custom engine: 2-3 months

game: 4 years

someone please help me balance this my custom engine is killing my studio

"oh but eni come on the engine for kitsune tails took way more than 2-3 months"

yes and if you think i could've taken a shortcut using out of the box physics engines or colliders or any of that nonsense and still wound up with a game that felt as good and authentic as kitsune tails did you are extremely mistaken

anecdote time! on a now dead project ages ago i had heightmapped terrain the player could traverse. so i wanted them to be able to slip down inclines that were too steep

i wrestled with an out of the box physics library for *months*. i eventually recompiled it to use 64-bit floats instead of 32-bit to avoid jitter far from the origin. i even managed to kludge together a solution for a world that could wrap around on itself, which meant physics wouldn't work at the edges of the world

but no matter how hard i tried i couldn't get the "slip off an incline of terrain this steep" to work reliably the way i needed it to

i bit the bullet and wrote my own collision solver and crude physics system. it took days, and not only solved the slipping issue but also made the physics more predictable, improved performance by removing stuff i didn't need, used 64-bit floats from the start, and fixed the issue at the edges of the world because it was designed from the start for a wrap-around world

@eniko one of the things that makes me sad about the "don't make your own engine" popular advice is how much it cuts down on the possibility space, as it is often like saying "don't make games that don't fit within the limitations of off the shelf engines"

@aeva yeah I agree. People act like generalized engines like Unity or Unreal don't impose their own limitations but I don't feel like that's true. Sure you *could* do things in a completely bespoke way but at some point you wind up fighting the way the engine wants to do things so much its just way worse than going from scratch so like, nobody would ever actually do that

@eniko @aeva I have this exact same problem with the culture of monolithic frameworks in programming in general...

Follow

@eniko @aeva (And most especially with the widespread - and also wrong - belief that monolithic frameworks are the only way to develop anything efficiently, or work in a team, or reuse other people's work, or or or)

· · Web · 0 · 0 · 4
Sign in to participate in the conversation
Pixietown

Small server part of the pixie.town infrastructure. Registration is closed.