To be fair, it's perfectly fine for 3D, such as traditional "isometric" style 3D. It suffers more with the lack of clever culling in first person style cameras, especially near large objects (it renders the whole object, even if 5% of it is seen).

Absolutely. Vertex culling is a long-standing 3D engine optimization.

No.

Godot 4.0 will include occlusion culling actually, so there's hope.

It's also slowly, becoming an alternative for VR development. They even had hand tracking on the Oculus Quest before Unity and UE

> They even had hand tracking on the Oculus Quest before Unity and UE

facebook is a large financial contributor to the godot project. I think they do some dev work too. I get the impression they really want more VR games to exist, and work well on their device, no matter how they get there.

Yeah, FB/Oculus has quite the interest in Godot. But at the moment there is no official Quest game made with Godot yet AFAIK. So far we made it into AppLab with VRWorkout, but the store is a completely different story

And as mentioned in the article/release, they are serious about maintaining Godot 3 and won't neglect it in favor of 4.

The more comments I read in this thread, the more this sounds like Perl 5 and Perl 6.

It seems as though this is kind of a common paradigm for game engines. Because the tech and/or best practices move so quickly, trying to lock in full backwards compatibility becomes a burden and thus engine devs fall back to "If you started your game on 3.x then it's probably best to stay there. Your next game can start fresh on 4.x"

It'll take some time before you can confidently make that assessment, though. It took ten years for Perl 6 to come out, and then actually it didn't come out as Perl.

Why Pearl instead of the dozens of other successful projects that maintain a stable version while working on a new version?

Because I worked with Perl and got burned a bit so it came to mind. I understand game engines are different and people build against stable versions for years sometimes.