Hacker News new | past | comments | ask | show | jobs | submit login

I love godot, but:

* gdscript is not python, it's lacking several small python things that makes python awesome: list/dict comprehension, .items(), sets, although gdscript is quite enough and still quite a good language for game dev as it does just everything well, especially for natives types etc.

* I tried gdnative and I still need to try gdextension, but so far writing C++ for godot seems quite hairy and the programmer must use godot types and "new()" everywhere, which is not really easy to deal with, so godot is not really a C++ engine, C++ is only usable if you want to optimize something a bit tedious, which is fine, but important to remember, and it confirms the whole "use a scripting language 99% of the time, and use C++ only if scripting is not fast enough"

* I would still want to inspect the real potential performance of gdscript compared to other languages, and I don't know if mono/.net is faster. It seems like a very good language especially because it integrate perfectly with how the environment is designed, but I'm still curious if they plan to improve it.

* I'm planning to make a multiplayer FPS game, but I don't think godot provides network prediction, and I wish it did.

* I'm a bit worried about opengl support in the future, it's not entirely done yet, and I'm afraid it's too difficult to properly implement because new API like vulkan or DX12 are so radically different.

Overall godot is the greatest thing that ever happened to open source and game development, game development is one area where open source has trouble making its teeth, and the lightweight design of godot makes it just such a big relief when you compare it to monsters like unreal and unity.




> gdscript is not python, it's lacking several small python things that makes python awesome: list/dict comprehension, .items(), sets, although gdscript is quite enough and still quite a good language for game dev as it does just everything well, especially for natives types etc.

This is my biggest issue with Godot, although not that GDScript isn't python, its that it isn't a "real" language (or I guess general purpose might be a better word). In fact its probably the only thing I miss from Unity.

To be fair I'm not sure I'd pick Python myself, as much as I love the language I'd probably pick something with a bit more focus on performance. Go might be a good choice, I dont know.

It's not even that GDScript is bad, in fact without comparing it to other languages its pretty great, but you just lose so much when compared to general purpose languages. I know theres bindings, but having done the same with Raylib the support for non-official languages is pretty much non-existent. Its pretty bad for Mono, which is an official binding


The official C# bindings are .NET, not Mono. They were using Mono in older versions, though.


There are official bindings to C#, the official docs have C# examples in them, and it can use the latest and greatest .NET versions, not just mono.


The docs have c# examples, but the community by and large uses gdscript, it's the "default" language. Again, it's the same with Raylib, bindings have documentation but it's difficult to find anything using that language outside the docs.


Yes, .NET is faster. Though even AAA games often use slow scripting languages for game logic. Many games, the game logic just isn't a big CPU hog so it doesn't matter.


What does network prediction from a game framework look like? You describe what properties of objects should be interpolated and the engine handles it for you?


Yes, I guess, properties like projectile/player velocity? Network prediction is a bit tedious to do because it's a lot of fine tuning.

It's only useful for latency sensitive games like FPS games or any games that involves simple physics with player versus player situations.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: