I have been developing a game for a while and seem to constantly change the design. I realized I am unskilled at art and don’t really want to spend too much time on it. So I want to make a roguelike.
I want to make a whole world, except, the player controls a character and other characters have their own agenda in the world. Say I have a randomly generated but finite world, maybe 1000 animals and 500 humanoids all acting on their own. So maybe the computations relevant to you (your screen/where you are in the world) and the characters take priority over the many that exist in the world. Then, the characters in the world update as you are making decisions between turns.
if (you are providing input)
Update your character and others nearby
else
Update other characters until input is provided/ Interrupt this if it is.
This is overly simplistic, I understand that but I do believe I would be able to implement this. Are there any caveats with my plan? I seem to think too much before starting, maybe that is my problem.
You’re focusing on the wrong problem: Making this much content will take way too much time (and generating it procedurally won’t make it take less).
Focus on finishing a project. And for finishing a project, you need to make it as simple and small as possible – it will always take much more time than you think. If you don’t keep the scope small, you’ll never finish. Finished projects are much more valuable than your ideas and abandoned projects and you can then grow your project size as you complete them.
Game jams are a great way to learn how to keep projects small and to iteratively learn, instead of getting stuck redesigning the same project forever.
I would agree to keep it small and get your basic game running first. Then you can scale up and IF you encounter performance problems you can face them directly. Profile and figure out where the performance is lost. Simulating nearby characters with higher precision/more often is one way to do it. Or you can split all characters into groups and simulate each in a different frame. But optimizing before you know the problem is not advisable. But yes it can’t hurt to think if your idea will SOMEHOW work.
Thank you for the feedback. I appreciate the concern but I’ve made about three, which may not be a lot to most but want to make something more large and complex. I made all this AI and want to do something with it, learn memory management and optimization.
But let’s say I take your advice, because I do see value in it. If I make a demo with only animals with simple behaviors (Hyenas, lions, and kudu), I could make some demo where I still practice working with loading in chunks of world data, with a simple world (2D nature world generated with Perlin noise, no Z level). Then if this works well with the animals, I can scale this up like Olipool said. And thanks for your feedback by the way Olipool.