The first Beta of Godot Engine 3.2 has been released

Blender and Godot are very good examples of how much open source software can achieve.
Both are the reason why I migrated to Linux last year.
I use Kubuntu since Artful Aardvark and loving it.

Blender and Godot are very good examples of how much open source software can achieve.
Both are the reason why I migrated to Linux last year.
I use Kubuntu since Artful Aardvark and loving it.

Blender - yes. Godot is still far from usable state for anything serious. But I really hope it will be ready in next 2-3 years.
These two are really not comparable. You need to try to use both of them to understand why. Blender is absolutely fantastic. Godot is maybe on the road to begin something.

Godot is in state wanted by its developers, according to their vision. Visions change very slowly, so I guess 2-3 years is too optimistic. However even now if your C++ skills and experience are good enough you can write quite decent games using it, but prepare to code a lot first. The problem is mainly a conflict of interests between minimalist public which considers any additional feature as personal insult and feature hungry public. Currently the interests lean to minimalism even at cost of lack of functions. That is not going to change any time soon, so to have some decent FOSS engine for your 3D game one have to either compose one from existing separate libraries (OGRE, Bullet, etc.) which will take human-years to complete (including asset pipeline establishment), so requires large team efforts, or take Godot + external libraries and have something coded within months which is much less effort. All the complete solutions are not FOSS, so Godot is compromise.

Sometimes its really strange how very small and less-than complex contributions that would add absolutely essential features to Godot seem to get stuck in the approval-queue for ages. Like for example https://github.com/godotengine/godot/pull/23864 that would allow adding properties to tiles and is needed for pretty much any 2D RPG or RTS imaginable. I was really hoping to see it added to 3.2, but I guess not... :S

I get why they feel they need to become a more serious option for 3D games, but personally I wish they'd get the 2D side right first, because I think that's what most current Godot users are developing.

Other than that, glad to see that 3.2 is coming along nicely, anyway! :)


Last edited by Kimyrielle on 6 November 2019 at 5:38 pm UTC

Obligatory "I've been waiting on this" joke.

Godot is in state wanted by its developers, according to their vision. Visions change very slowly, so I guess 2-3 years is too optimistic. However even now if your C++ skills and experience are good enough you can write quite decent games using it, but prepare to code a lot first. The problem is mainly a conflict of interests between minimalist public which considers any additional feature as personal insult and feature hungry public. Currently the interests lean to minimalism even at cost of lack of functions. That is not going to change any time soon, so to have some decent FOSS engine for your 3D game one have to either compose one from existing separate libraries (OGRE, Bullet, etc.) which will take human-years to complete (including asset pipeline establishment), so requires large team efforts, or take Godot + external libraries and have something coded within months which is much less effort. All the complete solutions are not FOSS, so Godot is compromise.

I have to agree somewhat.
Love to see the engine progress, but I've often been questioning the priorities and the lack of clarity of what is going on at any current time, what the plans are for the next month/week, etc.

Is it really necessary to work on a Vulkan renderer while OpenGL+DirectX works everywhere and other features are way more incomplete than the rendering?
It's not the lack of graphical bling! that keeps people from using Godot, it is the lack of features. When asking "can Godot do this?" and the answer is no, well then most people won't pick Godot if Unity or others can do it.

Don't get me wrong, I'm not questioning what the unpaid contributors work on, they'll work on what motivates them and that's fine. But shouldn't the paid ones be more mindful of priorities?

I gave up waiting on the 3D navigation, for example, and started working on my own GDNative module to implement recast/detour properly for my needs. It won't be in a way that suits everyone (no editor integration, for example and only takes MeshInstance), but that's just what happens when people develop for their own needs.
It is a few weeks of work (spread over one day per week that I have for it) that would have been nice if I could have spent them on something else.


Last edited by TheSHEEEP on 6 November 2019 at 6:56 pm UTC

I gave up waiting on the 3D navigation, for example, and started working on my own GDNative module to implement recast/detour properly for my needs. It won't be in a way that suits everyone (no editor integration, for example and only takes MeshInstance), but that's just what happens when people develop for their own needs.
It is a few weeks of work (spread over one day per week that I have for it) that would have been nice if I could have spent them on something else.

Seen my old PR and navigation branch? might be useful in your implementation.

Also I need to say that if Armory engine or Torque3D or any other FOSS 3D engine on horizon will get same set or more in features I will be more than happy to move or at least try, but with currernt state of affairs Godot seems only option for 3D game.

I've been playing with game engines recently and Godot is far from usable yet really.
But what I have discovered is Urho3D, where I have rapidly made progress with some ragdoll physics using the bullet library.
Urho3D really is a great and polished library but not that well known (yet?).
I suppose its downsides are it is limited to OpenGL over vulkan but that isn't such a big issue.
It also seems to have a large and active community of developers.


Last edited by nattydread on 7 November 2019 at 4:43 pm UTC

Urho3D is very limited with what you can do (i.e. small number of bones per mesh, very inflexible asset pipeline, ragdoll only works with example meshes, IK is as limited as one in Godot, no scene editor, no asset pipeline for anything beyond a few static meshes unless you use older Autodesk tools. Blender export will require much more manual work than with Godot or Torque3D. There is much less features than with Godot and you can take Urho now only if you're for minimalism (in which case I suggest composing your own engine using OGRE and Bullet, which will produce even better thing for your game). The engine architecture in Urho is very good though and the engine is easy to understand, so can be used as starting point if your requirements are small and C++ skills are strong enough. However due to lack of effort and visibility you will encounter architecture bugs and shortcomings nobody will fix for you, so if you're not ready to maintain your engine fork for your game with heavy changes, I'd suggest only use Urho as example of how engine can be composed and use other tools. Currently I think that it might be a good project to port T3D to mobile and fix its renderer and get fully-featured engine than play with Godot or Urho3D. Creating own engine from FOSS libs is another option.

The big problem I see with Urho3D is that its development more or less stopped when the main developer moved on to other things 1-2 years ago.
There are still one or two people doing semi-regular pushes to the code, so I wouldn't call it dead at all, but that is not really comparable to the work being put into Godot on a daily basis.

Though, yes, for C++ folks, it might be easier to get into Urho3D than Godot initially. Still, I think you'll end up with a dead end there sooner or later and in contrast to Godot, won't have much of a community to help or other ways to manually deal with it.

I've been playing with game engines recently and Godot is far from usable yet really.

What was it that was missing for you?
I know what was missing for me and decided to just roll that part on my own, but the Godot community generally wants to know reasons why people do not want to / can not use Godot to see where the shortcomings are.

Hi Slapin, TheSHEEEP

Thanks for your comments, its interesting to hear another viewpoint.
You said that ragdoll only works with example meshes, but isn't that true of every open source game engine?
I haven't seen any code that really does this in a flexible way.
I did start composing my own code using OpenSceneGraph and Bullet but I felt like I was reinventing the wheel! I might go back down this route, because ultimately I want to switch between the graphics engine and the physics engine and I really would like it to use vulkan in the end.
My requirements are small and C++ skills are strong so Urho is suitable for now :)
It was quite a while ago that I tried Godot, I might revisit it soon though as I know it is rapidly changing.


Last edited by nattydread on 8 November 2019 at 3:00 pm UTC