Hacker News Reader

Comment by HammadB | Hacker News Reader

I'd love to hear your thoughts on why you don't recommend that. Think I could learn.

There was a time when Unity was by far the best tech platform for building your game on. Then they took on a ton of VC money, grew very fast, and went after all forms of mobile, VR, and desktop development, at which point they morphed from the best platform for game development to the worst platform for enterprise development.

Speaking as a 10 year customer of Unity, I can’t wait for Three.js, gltf, and a couple other components to get just a bit more solid so I can drop my dependence on Unity. Three or four years ago it would have been inconceivable for me to say something like that, as I was a long term hardcore fanboy.

It’s just not the same company or solid development experience it once was.

Speaking of Three.js, have you heard of PlayCanvas (https://playcanvas.com)? It's pretty much a Unity-like IDE based on Three.js, and allows a whole bunch of cool features, like live collaboration.

Here's an example game made in it: Tanx, https://tanx.io. And here's a lightmap-heavy scene for eye candy: https://playcanv.as/p/txPePQvy/

It looks good, but on my Macbook Air I get really bad framerates (<10 FPS). I'm not sure if that's because of this tool, or WebGL itself, but I'm able to play Unity games like Cities: Skylines without any issues.

I see, I also look forward to Three.js / WebGL stack to become a suitable replacement. These are exciting times.

For that WebGL needs to be able to offer the same feature set as native 3D APIs, which will never happen due to security and the new direction of graphics programming.

It only a suitable replacement for those not requiring more than what OpenGL ES 3.0 offers.

> For that WebGL needs to be able to offer the same feature set as native 3D APIs, which will never happen due to security and the new direction of graphics programming.

I work with many of the people involved in the Web GPU effort, and I'm sure they'd be surprised to hear that they aren't actually working on what they're working on!

This is just silly.

Silly is pretending WebGL will ever match 1:1 the capabilities of native graphics APIs.

It is absolutely a goal to match Vulkan/D3D12/Metal with the new Web GPU efforts.

If you're talking about matching every single feature in every API ever, of course no API can do that; no "native" API does that either. There are always gaps around the edges where some API has features others don't (e.g. conservative rasterization support is spotty). But general feature parity with the new generation of APIs is the target. I'm going to the W3C Web GPU meeting next week to talk about just that…

I like WebGL, just don't like having it lacking features like geometry and compute shaders, or having to manually white list the GPUs on devices where native applications execute without problems.

Looking forward to the outcome of W3C Web GPU meeting.

That is only worthwile if your games don't require anything more than OpenGL ES 2.0, or eventually OpenGL ES 3.0.

i'm primarily a programmer not designer nor game dev - so a priori an ide that store program state in the ui is weird (e.g. fiddling with booleans by either checking boxes or setting procedurally). but that's fine - you just get used to your state being spaghetti. the real problem with unity is it's closed source and all of its goofy bugs and breakneck release cycle make for really nerve wracking times. maybe the real thing is you shouldn't build your software business around a closed source vendor.

Reply on news.ycombinator.com