Yeah I do have a level of detail system I will try to bring over. I had to rewrite the engine to make it more WebAssembly friendly

Thanks! In addition to zstd you might be interested in the blosc library. Had trouble using it in Cloudflare workers though so I did something similar for my specific case.

Thanks for the ideas! Yeah, would like to improve on that aspect, the players should at least be able to tell where other players are facing and to be able to nod yes or no. I was stuck for a bit with some WebAssembly multithreading quirks for a bit but have mostly figured it out.

Thanks! I'm currently using WebGL2 to target more devices. In the future I plan to have an optional WebGPU renderer as well.

The physics is simulated using multithreaded WebAssembly and then the data is transferred to the GPU to render.

@@GrantKot awesome thank you! Also cool, I for some reason did not realize assembly supported multithreading

Yeah, if you're talking about combining the particle-based physics with this, that's going to be very soon, probably within this month. And then I also want to set up a WebGPU based renderer. I think things are coming together in a good way and I'm very optimistic with this now being cross-platform on the web.

@@GrantKot Oh my god man you are literally creating the engine I have always dreamed of for my dream fantasy sandbox MMO. Amazing work!

Yes, I went on a major compatibility binge this summer and all of my tech builds via WebAssembly and is cross platform including mobile. Demos are on my site

Hi, nice to meet you! Yeah, raytracing but just the simplest form, just primary hit, haven't added other lighting for it yet.

I have a link to my itch io page in the description for this demo (first one there) and some other demos I've shown on my channel.

nice dude

Thanks! This is TypeScript and WebGL2. Once the physics system is there it will be a bit of C++ -> WebAssembly.

Personal opinion (because I had the idea to do something similar to what you did which is why I am thrilled to see it): I like editing stuff with high detail, because you can create things very well that way. So I would have made it block-by-block placement but I guess this works fine as a demo. You stole my idea but at least your a good programmer in contrast to me ;)

The editing experience will be improved soon :) I'll be working on adding things like grids and the ability to customize their block tiles. Mostly here is just load testing

Thanks! Actually, I wonder if this will be useful to do cross-worker physics!

Yeah, I think I did anticipate this design might not work out but I wanted to try it out to better understand what having a central D1 database would be like. While this was a fail, I learned about the speed characteristics of D1 and KV and now have a better architecture planned.

especially with voxel engines it almost makes more sense to use video streaming than streaming the voxel data, considering the unbelievable amount of data trafic needed

@@GrantKot All the respect for tackling it! It's good to push yourself. I'm not a genius, though, so I just stay in my lane and relax. (Excuses! I'm simply lazy!)

@@dottedboxguy Here's a crazy idea: Could you extract voxel data from an image to rerender it with the cloent/viewer's settings? Like if it were a simplified render with no lighting only voxel position and color data, using player coordinants and camera angle, could you fit a grid to the image and quickly extract the visible blocks and then fully render them? Obviously this doesn't seem practical for an actual anything because the server would have to rerender for every player's view at 60fps or more, but would yhe comcept work? For fun?

Actually, if you only render at, like, 15fps and low resolution (480p?) and had the client store the possible voxels and update them as voxel came into better view and render from that it would probably be viable for a test. Still a bad idea, but workable, I think. To cover up errors on far away blocks you could have fog. Definitely a funny idea.