This is really cool, and thanks for putting the time in to create working demos. I haven't done any game programming in Rust yet, but I will definitely reach for this if I do.
I wonder where one can find a physics engine that can stack boxes without them flexing, wobbling and falling in a very unrealistic way [1]. It seems very common.
Well most physics engine, including nphysics, bullet, box2d, are able to achieve much more realistic simulations: just increase the number of iterations for the constraints solvers, and reduce the timesteps length. Of course this will no longer be real-time but will be suitable for generating animations. In the end, it always boils down to a compromise between performances and realism unfortunately.
I don’t want to sound too grumpy and all; but it’s probably a fifth n-body project in wasm I see here on Hacker News. On the other hand the big elephant in the room is the UI on the web.
Why is noone doing it? I’ve only seen a single half-baked UI toolkit for wasm+canvas (or however you call the combination of 3d and wasm).
I understand there are ways to build interfaces on the web, but it seems a good UI toolkit is something a lot of startups would find incredibly exciting in the long run.
Because people build these projects in their spare time.
Just like you have observed there's no perceived business value to managers yet in WebAssembly etc. for companies yet that aren't directly involved with 3D (like AutoCAD Web, game engines, ...) or other visual rendering (like PDFKit), so most of the things that are being created are just cognitively stimulating for the developers creating them, and that's not necessarily creating UIs.
Building things like this physics engine is like playing Lego for adults.
When you create UI, people always have opinions and you suddenly need to justify why things work differently than what they think is how things should work. That's exhausting.
