3.1 going to Alpha. Feature freeze

emo10001

https://github.com/godotengine/godot/issues/21490

-emo

Ace_Dragon

And the bug fixes are starting to stream in at a nice pace :blush:

I hope the many contributors set aside time to fix as many as they can, as Godot could easily stand against Unity and Unreal by not sharing their reputation as buggy engines.

MagicLord

@Ace Dragon said: I hope the many contributors set aside time to fix as many as they can, as Godot could easily stand against Unity and Unreal by not sharing their reputation as buggy engines.

Each engine have bugs, i never seen any with no bugs :D

Let's be realistic Godot has been mainly focused on 2D engines for many years, it is great for 2D. But on 3D , it is good for mobile and small 3D games only, it is just starting in 3D.

I started a Metroid like project in Godot and i'm already very limited because many fundation features doesn't exist in Godot there is no good occlusion system in Godot for bigger 3D games navigation is bad, it needs to be rewritten, native, multithreaded, and get more features like agent and crowd avoidance. armature hierarchy is not available and can't quickly attach objects terrain and grass tools are missing tools to quickly prototype or make game art like Probuilder and Polybrush are needed no good water system (users made ones are not optimized and sometimes not great) while in UE4 i have great one for free * C# is not fully 3D stable and complete to get a faster game and to make it easy to code a complex game.

I have to decrease my expectations in Godot, while this is the kind of Metroid you can do with UE4

When you try to make something bigger and more complex in 3D , Godot is really not the best choice. UE4 and Unity have hundred tools and plugins, they have bigger teams, more releases, and many big games published, and they also have many thousand tutorials and free templates :

So i don' think saying Godot can start challenging the big ones makes sense :D You'll find in 3D there is many things not covered in Godot, one example is Godot cloth is a good addition but without the right tools like a brush tool it's not very usable and stays mainly as an Api instead of an editor tool or plugin.

It will take time before you could consider Godot on top 3 engines to make 3D games, saying it would have enough features and tools to not need to use Unity or UE4 (both also get major new features on each release). But i think Godot or Amory 3D can grow and become the Blender of the 3D engines :) (Why using Maya when Blender does more and better ?)

Ace_Dragon

No doubt Unity and Unreal can make amazing games, but Unreal and Unity devs. have also learned to replace limited built-in systems with plugins as well as author sophisticated workarounds for the many bugs they might encounter. Many major Unreal devs. even maintain their own fork of the engine to fix issues themselves and cherry pick commits. I would also not think you can download Unity and make Cities Skylines or Ori and the Blind Forest. with no money, as many major Unity devs. spend hundreds or even thousands of dollars on plugins to get the tools they want (though you theoretically could, it would just take a very long time to create all of the needed systems from scratch).

It can be done, but working around bugs or having your own engine fork takes a lot of time, and having Godot buck the trend of buggy software would make game development easier and faster for everybody. You said it yourself many months back, Godot should place focus on keeping the bug count to a minimum rather than spend all of the dev. time on the feature race (because in many cases, a rapid expansion of the feature set can mean many bugs and regression that need time to be fixed).

To conclude, "all engines have bugs" should not be an excuse to not work towards as low a bug count as possible. I came to Godot from the Blender Game Engine, and I can tell you it's not easy nor quick to work around engine bugs (to no guarantee your workaround will even function).

MagicLord

I'm not a Godot, UE4 or Amory 3D fanboy, i like those, i use the one that suits what i'm making, so my previous observations was very objective. It was not to discredit Godot , or mean UE4 is great, but show what is the differences with the big ones.

No 3D Engine is the best, even if Godot could challenge UE4 or Unity in some ways, the choice will always be a matter of preferences or some differences that would make it more easy or better to work with one and not another to make some kind of game. For example you could want UE4 for some very advanced features, or you could choose Godot because there is no royalties and it would cover all your needs.

@Ace Dragon said: No doubt Unity and Unreal can make amazing games, but Unreal and Unity devs. have also learned to replace limited built-in systems with plugins as well as author sophisticated workarounds for the many bugs they might encounter. Many major Unreal devs. even maintain their own fork of the engine to fix issues themselves and cherry pick commits. I would also not think you can download Unity and make Cities Skylines or Ori and the Blind Forest. with no money, as many major Unity devs. spend hundreds or even thousands of dollars on plugins to get the tools they want (though you theoretically could, it would just take a very long time to create all of the needed systems from scratch).

Lately they are making big changes : motion matching is coming new terrain system, tools and shaders are coming news post process stack HDR pipeline and shaders high performance ECS and job system new shader visual editor probuilder and polybrush for 2D, lot of 2D new tools

Unity is trying to challenge UE4 with a "no plugins needed to make an awesome game". Ori was lot of custom plugins mainly, buying some plugins is not very expensive when your game target is to become a number one hit during some months.

Plugins are needed, sometimes they can be what make you choose some engine instead of another when the plugin is the main tool to generate or help on your game content. Godot will also need good ones to help people speed up their game creation. Without plugins some Unity games would not have been possible (Gaia, Dungeon architect ...).

You don't need thousand plugins to make your game, many times only 2 or 3 plugins are must needed to speed up your game creation, because they generate your game levels ten times faster.

I'm not sure when Godot will be popular the best plugins will be free, the best Blender plugins are paid ones to allow the author to fix bugs and improve them, without them you would put ten times more work to create your game content ;)

It can be done, but working around bugs or having your own engine fork takes a lot of time, and having Godot buck the trend of buggy software would make game development easier and faster for everybody. You said it yourself many months back, Godot should place focus on keeping the bug count to a minimum rather than spend all of the dev. time on the feature race (because in many cases, a rapid expansion of the feature set can mean many bugs and regression that need time to be fixed).

I agree, Godot needs to stop adding things , make a pause on features and take some months on bug fix, but after that it should really focus on on bringing some big feature missing (navmesh and crowd, terrain, ocean , sky and procedural clouds, ai graphic tools, cinematic editor) and not only small features.

To conclude, "all engines have bugs" should not be an excuse to not work towards as low a bug count as possible. I came to Godot from the Blender Game Engine, and I can tell you it's not easy nor quick to work around engine bugs (to no guarantee your workaround will even function).

With more than 3900 issues, i would not consider Godot better than UE4 or Unity :D Among those bugs, some are blocking making it not possible to use Godot, for example it's not possible to bake navmesh on 3D tiles, it was the fundation for a 3D game i wanted to make while i can do it in UE4.

I think Godot will become the top open source engines for 3D games, like Blender it is the more active with more developers. Armory is awesome also with a full integrated modeler as Blender, but it really needs more developers, with only one active developer i don't find it's enough stable and reliable 3D engine.

Ace_Dragon

You talk about Godot being behind on assets, fortunately though, progress is being made.

Now of those, these aren't hardcoded engine features, but one thing that Godot developers do not want to do is allow the engine to grow to a multi-gigabyte monolith with performance degradation and game size issues due to bloat (though I have proposed an auto-strip system to ameliorate the latter).

In addition, the 3900 issues include feature requests and general discussion on engine topics, the actual issues are ones that contain the 'bug' label.

MagicLord

@Ace Dragon said: You talk about Godot being behind on assets, fortunately though, progress is being made.

Let's be realistic a little, none of those is really complete enough and polished, the same features are native in UE4 with a quality beyond those.

Actually i'm very disappointed about Godot users plugins for 3D, none is great or polished. For example the VR plugins works great, but there was too much unnecessary GDscript, it's also slowing it down while many things should have been native. There are too few plugins for 3D and the most important are not there.

@Ace Dragon said: Now of those, these aren't hardcoded engine features, but one thing that Godot developers do not want to do is allow the engine to grow to a multi-gigabyte monolith with performance degradation and game size issues due to bloat (though I have proposed an auto-strip system to ameliorate the latter).

Some things need to be native for performance and because it's part of most 3D games, spline modeling like the one in UE4, tiled terrain and tools, vertex paint, cinematic editor and many others. Why 2D features additions are all native and 3D important features (terrain, vertex paint ...)should be plugins ? Should not 2D features become plugins also for people making 3D games that does not need any 2D Tools ? About plugins the problem is none is great and complete or enough quality. Many water systems for Godot , none is really good , none is optimized, none contains advanced features like foam and underwater effects. Or grass and terrain plugin too much basic. *This is the main issue with plugins developed by users, they can abandon the plugin or decide to leave for many reasons. This is really bad when the feature would be a fundation feature. Plugins are great, but until those plugins become officially supported by Godot developers it will not be good. * I think Godot needs some approval and notation for plugins. And separate repositories : = one for everyone's extra plugins with no obligation to maintain the plugin when Godot has a new release with significant changes, the users can leave developing the plugin any time = another repository for Godot approved plugins that would become official and supported by Godot developers.

* Godot could unfortunately become like Unity a good base 3D engine, but lacking as native common 3D features for most 3D games (terrain, shader editor, vertex paint, cinematic tool, spline tool, object paint brush ...). This is the difference between Unity and UE4 that contains lot of usefull tools and plugins become officially supported by Epic without needing to buy on the store before it's something very needed.

Otherwise the quality expected on main features and support won't be there (tile terrain and tools, spline tool, cinematic editor, visual shader editor ...). Or until Godot would have like Unity a paid Asset Store that would keep plugins developers motivated because they can only win money when their plugin is top quality and support.

Schuster

Exactly as @MagicLord said.....water system.....nothing ... unusable navigation/obstacle systems...unusable.....terrain system....unusable... ....a lot of work on these plugins no doubt...but are practically unsuccessful... unattached and slow-not optimized-over scripted. ....what a softbody I have when I do not have quick water system.

Shin-NiL

I only want to see Godot 3 usable for 2D mobile development. The lack of OpenGL ES 2 and the screen capture bug left me stuck on 2.1. I hope all these issues will be fixed on 3.1, so I can finally discover the new features ;)

MagicLord

@Schuster said: Exactly as @MagicLord said.....water system.....nothing ... unusable navigation/obstacle systems...unusable.....terrain system....unusable... ....a lot of work on these plugins no doubt...but are practically unsuccessful... unattached and slow-not optimized-over scripted. ....what a softbody I have when I do not have quick water system.

I agree Godot should not throw half made features , it's better to get complete features usable with the right tools instead of barely usable Api only. For example :smile: = navigation needs a full re write and update to standards, more features ( crowd avoidance, agent) and be native and optimized with multi threading = Cloth Api needs the right brush tools = VR needs to be optimized, become native exposing only high functions and some polishing = We need a full tile terrain system with most common terrain tools because users small plugins are barely usable, incomplete, not native and not polished.

A vertex paint tool for 3D and a quick brush object placement tools will also be needed integrated in the editor like UE4 because those are among the fundation tools used to make almost any 3D project. Godot should focus on usability instead of throwing not polished and incomplete features.

Where is the benefit to announce GLFT 2 support long time ago when it was crashing and not able to import animations ? Why announcing cloth when there is no tools and it's barely usable ? VR support was also announced too early, when it would need to be native, optimized and get more work ? I don't think it's a benefit to announce lot of things when they are not complete and not polished, it leads user deception most of the time.

I think for 3D we don't want early access features that are disappointing, instead Godot should really change the way new features get in. Take all necessary time to deliver polished,complete, optimized features with the needed editor tools instead. This way some features could take more months to be available, but this is really better that way.

Schuster

.....but it's also true that developers do not have a lot of money available, and the competition with the best engines (Unreal,Unity) is still in the beginning (I mean 3D) ...implement modern 3D features is not easy.....it's as it is...I mean improvement...polishing and optimizations will follow after 3.1 stable. I love Godot engine ( it is a unique...ingeniously designed engine ) and Godot is really already now ready to create an interesting smaller 3D game or some creative content........I trying to create such 3d projects.........but for a something bigger like AAA titles still have to wait....wait for this polished systems.

Ace_Dragon

Here's the thing.

Those here expecting a carbon copy of Unity to come about have come to the wrong engine. Godot is not about copying the competition, it's about creating a unique approach to game development that aims to be better. Godot is aiming to be an engine that is fully-featured yet lightweight in terms of needed disk space and processing power.

Also, please note that due to the open source nature, features that become commonly used among contributors will likely see patches made to expand them. Also, there is more than one way to approach any given game mechanic even though it might take more work (so I'm not sure if it's productive to say a game project can't continue because a certain feature isn't at full AAA quality). If you require the whole gamut of features at AAA quality to even proceed on a project, then Unreal is really your only option. It places a huge load of CPU and disk resources and is often overkill for smaller games (not to mention the rate bugs are introduced), but it gives every possible feature that could possibly be wanted.

I'm not arguing against the improvement of the current feature set (there are many improvements that can be made across the board), but true AAA quality isn't going to come overnight and like with Blender, various areas need to be prioritized when faced with thousands of todo items and limited resources.

MagicLord

@Ace Dragon said: Here's the thing.

Those here expecting a carbon copy of Unity to come about have come to the wrong engine. Godot is not about copying the competition, it's about creating a unique approach to game development that aims to be better. Godot is aiming to be an engine that is fully-featured yet lightweight in terms of needed disk space and processing power.

There is no unique approach when you need a tile terrain system and tools, the goal is to get complete tools and features, enough polished instead of some work in progress unfinished feature.

Also, please note that due to the open source nature, features that become commonly used among contributors will likely see patches made to expand them. Also, there is more than one way to approach any given game mechanic even though it might take more work (so I'm not sure if it's productive to say a game project can't continue because a certain feature isn't at full AAA quality). If you require the whole gamut of features at AAA quality to even proceed on a project, then Unreal is really your only option. It places a huge load of CPU and disk resources and is often overkill for smaller games (not to mention the rate bugs are introduced), but it gives every possible feature that could possibly be wanted.

Why do you bring AA quality in the discussion ? I don't consider Unity for AAA games , but for indie games instead, so Godot unlike Unity lack fundation features and tools. For example the cloth system is incomplete without a brush tool, i consider the brush as fundation not as some extra features or AAA feature.

Indeed you won't get all UE4 terrain tools, but Godot should get the fundations for terrain and enough polished.

What do you think about users plugins ? When they can stop working on them and leave, what users can do in that case ? For example the grass and terrain plugin for 3.2, no one tried to migrate it to Godot 3, too complex and complicated. Should Godot select some plugins and specify as officially supported by Godot (like the terrain and grass) ? I think is a more important subject.

I'm not arguing against the improvement of the current feature set (there are many improvements that can be made across the board), but true AAA quality isn't going to come overnight and like with Blender, various areas need to be prioritized when faced with thousands of todo items and limited resources.

Why do you talk about AAA quality ? This is not the subject. The whole discussion is about delivering polished features instead of not working like Glft2 , barely usable like cloth system, or not optimized like VR, not working like navigation system, lacking features like terrain and tools fundations or like a good occlusion system.

Godot should allow to make the same indie games (non AAA ones) as Unity, and as easy with the right features and tools. This is not about quality but polished features and tools.

This is not about Godot becoming a copy of another 3D engine, this is impossible, each is different, but each 3D engines copies other when it comes to good ideas. Unity copied UE4 object selected with an outline in the editor for example, or they are copying UE4 graphic presets in some way. There is many good things in Unity or UE4 that would make 3D game creation in Godot faster or more easy, for example Unity scene view Gizmo is very intuitive and allow to change views in one click. This is not about become an exact copy of another 3D engine.

Ace_Dragon

I'm confident the features will improve over time. Our physics developer even has a pull request for a fix to the cloth system. https://github.com/godotengine/godot/pull/21599

Also, some features require a certain method to really get working like the navmesh. Examples of this can be found with the suite of Godot demo projects that show how to get a large number of features to work.

Furthermore, no matter which engine is in use, a developer will often have to end up choosing whether or not to let their project be abandoned because an engine feature isn't going all the way in terms of implementing a mechanic. This is where the decision needs to be made to create a system that makes up for the limitation and keep the project going.

MagicLord

@Ace Dragon said: I'm confident the features will improve over time. Our physics developer even has a pull request for a fix to the cloth system. https://github.com/godotengine/godot/pull/21599

I'm also confident Godot will bring what is missing and should start polishing features. Just saying it will take time to bring the right tools and fundation features, instead of too much fast development of features and saying it's available and completed.

Also, some features require a certain method to really get working like the navmesh. Examples of this can be found with the suite of Godot demo projects that show how to get a large number of features to work.

I know how navmesh works and used it, it's not native, path calculation must be native and multi threaded, agent component is missing, ability to combine physics is missing, tools to create offlink navigation directly ion the editor.

You can find the issue on Github, there is even a dowloadable example. https://github.com/godotengine/godot/issues/17885

I prefer to talk with people that really tried a 3D feature and tried to code a game with it in Godot, instead of people that just assume things.

Furthermore, no matter which engine is in use, a developer will often have to end up choosing whether or not to let their project be abandoned because an engine feature isn't going all the way in terms of implementing a mechanic. This is where the decision needs to be made to create a system that makes up for the limitation and keep the project going.

I use UE4, there is no problem with that, while i prototype in Godot a game ( and indie game not a AAA quality game ;) ), something beyond a tower defense or arcade game so i can give feedback about what is missing. The discussion is not about asking fundation features or fix features issues, there is Github for that.

You could start a 3D project in Godot, let's say an indie survival game, a game with a big environment looking good enough with vegetation, good navigation for AI, and good physics, without writing any plugin and only writing gameplay code, then we will be able to talk together about 3D fundation features.

Ace_Dragon

I'm working on 2 different 3D game projects right now and I haven't had any notable showstoppers pop up. Sure, the render engine needs quite a few fixes, but it's not something that actually stopped me due to some areas having more than 1 approach.

I also use the well-established GDscript language (where the vast majority of functions work the way they are supposed to). I know your preference is C# which is still very much a WIP, but one of the beauties of the Godot paradigm is allowing people choice in what approach they use for game development (which is something that hopefully should become less of a burden because the last year has seen a massive uptick in patch contributors).

MagicLord

I'm reducing the scope of the Metroid game i'm making, it will be something very simple, because it will take time until Godot gets complete on fundation features.

@Ace Dragon said: I'm working on 2 different 3D game projects right now and I haven't had any notable showstoppers pop up. Sure, the render engine needs quite a few fixes, but it's not something that actually stopped me due to some areas having more than 1 approach.

I also use the well-established GDscript language (where the vast majority of functions work the way they are supposed to). I know your preference is C# which is still very much a WIP, but one of the beauties of the Godot paradigm is allowing people choice in what approach they use for game development (which is something that hopefully should become less of a burden because the last year has seen a massive uptick in patch contributors).

What are you assuming again ? All my Github examples or game i'm making are using GDscript :o ( until C# is fully stable and complete )

So your games must use very limited gameplay and no navigation, no terrain, simple level collisions, and you must not have used lot of features if you don't found any showstopper to put on Godot GitHub.

Really, when it's about features i prefer to talk with people that try more complex games, something bigger than arcade mobile games, and people that really tried features in their games able to say if they work good or not, not just assumptions.

Ace_Dragon

@MagicLord said:

So your games must use very limited gameplay and no navigation, no terrain, simple level collisions, and you must not have used lot of features if you don't found any showstopper to put on Godot GitHub.

Really, when it's about features i prefer to talk with people that try more complex games, something bigger than arcade mobile games, and people that really tried features in their games able to say if they work good or not, not just assumptions.

So who's making the assumption in this case? My games being supposedly very simple and limited because I should otherwise be running into blockers and talking about how Godot is broken and unusable?

One of my games does have terrain (with splatting even), courtesy of the shader API and a mesh modeled in Blender. I also have reasonably complex level geometry where collisions work well because of the Bullet physics engine. I am also using the Better Collada exporter which does a reasonable job in importing assets in a working state. My games have scripts with many hundreds of lines of code across up to 2 dozen files with reasonable performance (noting techniques such as segmenting scripts into functions and invoking a "passing game" using the call_group function).

The point to make is that your experience with Godot does not speak for everyone, and neither does my experience. There are things I have done in Godot which actually turned out to be a bit trickier than expected, but I didn't stop work in favor of posting on how the engine failed me.

MagicLord

@Ace Dragon said: So who's making the assumption in this case?

You made many already.

One of my games does have terrain (with splatting even), courtesy of the shader API and a mesh modeled in Blender.

This is not a terrain system, but a mesh. Texture splatting is not a feature but it's shader code anyone can find examples.

I also have reasonably complex level geometry where collisions work well because of the Bullet physics engine. I am also using the Better Collada exporter which does a reasonable job in importing assets in a working state. My games have scripts with many hundreds of lines of code across up to 2 dozen files with reasonable performance (noting techniques such as segmenting scripts into functions and invoking a "passing game" using the call_group function).

Anyeon is able to import animated models with Better Collada , it's an old feature, while Glft2 was buggy and crashed.

I don't mind about your thousand lines of code, what are your games about? what far is the gameplay pushed ? how far Ai is pushed ? How 3D level is complex and can run without culling ?

By complex indie game i mean, tiled 3D terrain you sculpt and paint in Godot directly, good navigation, pushed ai, interactive foliage with physics, 3D ui, physics that does not jitter or jump on small slopes, good indie looking water system optimized, some time of day and perhaps some dynamic clouds system, inventory and stats and some others things. We don't have the same definition of fundation tools to make a game :D

The point to make is that your experience with Godot does not speak for everyone, and neither does my experience. There are things I have done in Godot which actually turned out to be a bit trickier than expected, but I didn't stop work in favor of posting on how the engine failed me.

What Godot 3.1 allows you to do in 3D is not comparable to what indie game UE4 or Unity can achieve, that's it.

Godot will grow lot more with 3.2 and new versions, while i don't find 3.1 enough complete and not significative until i would make small 3D mobile games.

We don't need another thread with many pages, i already had the discussion about Godot fundation features and polishing things. Good for you, i wish you succeed in making your game in Godot.

Ace_Dragon

i already had the discussion about Godot fundation features and polishing things. Good for you, i wish you succeed in making your game in Godot.

Well then, I guess we diverge a bit on what features we desire to have before really getting down to make a nice and complex game, I do agree by the way that nearly everything you listed in that last post would be good to have (though it will depend on whether they are enough of a foundation feature to be high priority and what the patch writers need for their games).

There's no direction to take the discussion from here, but nearly everyone here would probably love to see a major bugfixing race in this final leg to 3.1 final.

I too wish you success in making your games, and don't hesitate to report bugs (especially if you can put together an example file) along with requests to remove the more pressing limitations.