Godot 3D lack of polish

MagicLordMagicLord Posts: 578Unconfirmed
edited April 30 in General Chat

Hi there,

I like Godot 3, it's very easy to pick up and create any game; but i find it is too unpolished about 3D.

There is many bugs, a common one is a physic capsule getting stuck on a cube corner, or KinematicBody can become buggy and it's function move_and_slide() not reliable. The navmesh does not perform good path calculations, it is too primitive.
GI probes looks good but there is no blending between them.
There is many other problems encountered, but also the lack of tutorials, many things are not explained and not available on documentation but on other sites.

I wish there was less features announced (CSG tool), instead priority should to get better engine fundation.
Hopefully should get better.

«134

Comments

  • Ace DragonAce Dragon Posts: 192Member

    There are some rough spots, but pull requests are being made daily (once the core team gets back around to merging them in waves as in the past, that alone should bring improvements).

    GI probes do blend though, but to do that requires multiple probes that are overlapping (the way they work is a little different compared to reflection probes).

  • SeaFishelleSeaFishelle Posts: 39Member

    Documentation and tutorials are a matter of building a community, which requires time and prestige. It's happening though, Godot is becoming one of the names people put out there when talking about "games engines you should try". It being free (and open-source, but let's be honest, the general public cares more about "free" than "open-source") is one of its best selling points. There are a couple of tutorials out there, too, so it's not like Godot 3D is a blind wasteland here.

    Remember, Godot has only been on the public market for around half a decade, and 3.0 only hit market a couple months ago. Unity and Unreal have been around for AGES, and have both been used to create games everyone knows. Godot will get there, it just needs time.

    And, not to toot a horn or anything, but a strong community is a community that caters to people of all kinds. The live-chat crowd, and the wait-to-chat crowd that prefers forums to chatrooms. On that note, the forum might need a beefing up.

  • SchusterSchuster Posts: 295Member

    Godot 3 is a little behind the starting line......so polish needs some time :)

  • emo10001emo10001 Posts: 121Member

    I have no doubt that the polish will come. There are some good tutorials out there with a little bit of searching. Especially if you look in the tutorials section of this forum.

    I've actually been considering putting together a few tutorials myself, but we'll see if I have time :smile:

    But overall, this engine is already solid, and only getting better. It's not perfect, but none are. I suspect that 2018 will show way more growth.

    -emo

  • Ace DragonAce Dragon Posts: 192Member

    What Emo said.

    Godot right now has a lot of developers working on the code, and many of them are also creating games and making patches based on what they need (which includes a lot of stuff that improves on existing functionality). Godot's development rate has increased massively since the first year when you look at the github page and the number of contributors (it used to mainly be Reduz and a handful of other people).

    You should see the early versions of Godot and how easy it is to do various things now compared to back then. Back then, it was a bit more difficult to do things like changing scenes and getting objects to look at each other (which GDscript now has dedicated functions for). Godot also didn't have any API documentation (it's not complete yet, but the 3.1 development builds even have significantly improved docs. in some areas).

  • NeoDNeoD Posts: 135Member

    I guess the developer of a feature should be the more appropriate person to fix a bug on his code. Of course, I do not exclude that other people can help too.

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 1

    What you do when your character capsule is stuck on a corner ? I tried some work around but nothing good enough so i give up on the 2.5D platformer.
    Navmesh returns more than 30 path points to move along a very small distance, and sometimes path calculated was way longer going around obstacles when there was a very short path with no obstacle, i also gave up on another project using it. This is very few examples of problems i encountered.

    You can make very simple 3D games, but once you go further you'll face those type of issues.
    Godot on 3D is not there yet.
    This is what i mean about working on the fundation and let features for later instead.

  • emo10001emo10001 Posts: 121Member
    edited May 1

    @MagicLord said:
    This is what i mean about working on the fundation and let features for later instead.

    It's a valid point. We definitely want to see a healthy balance of "making sure the core basics work well and are intuitive" verses "new and shiny feature to help compete with [insert big-time game engine here]!"

    But I think that will happen.

    -emo

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 1

    @emo10001 said:

    @MagicLord said:
    This is what i mean about working on the fundation and let features for later instead.

    It's a valid point. We definitely want to see a healthy balance of "making sure the core basics work well and are intuitive" verses "new and shiny feature to help compete with [insert big-time game engine here]!"

    But I think that will happen.

    -emo

    Exactly, before trying to race against [big game engine], first get fundation right.

  • Ace DragonAce Dragon Posts: 192Member

    @MagicLord said:
    What you do when your character capsule is stuck on a corner ? I tried some work around but nothing good enough so i give up on the 2.5D platformer.
    Navmesh returns more than 30 path points to move along a very small distance, and sometimes path calculated was way longer going around obstacles when there was a very short path with no obstacle, i also gave up on another project using it. This is very few examples of problems i encountered.

    You can make very simple 3D games, but once you go further you'll face those type of issues.
    Godot on 3D is not there yet.
    This is what i mean about working on the fundation and let features for later instead.

    Don't forget to report genuine issues to the Github issue tracker (providing it is not already there).

    Also about the physics/collision problems. Godot has an option to switch the physics engine to Bullet now (which should be far better in almost all cases compared to the original one). It's not set by default, but anyone experience such issues should try the switch before reporting.

  • newmodelsnewmodels Posts: 119Member

    I hope godot will continue to have their own built in physics, not become dependent.

  • TwistedTwiglegTwistedTwigleg Posts: 701Admin

    @Ace Dragon said:
    Godot has an option to switch the physics engine to Bullet now (which should be far better in almost all cases compared to the original one). It's not set by default, but anyone experience such issues should try the switch before reporting.

    I think Bullet is the default physics engine for Godot 3, at least that is the impression I get from this article back when it was announced. Also, according to that same article, it appears Godot's built in 3D physics will be removed in Godot 3.1.

    @newmodels said:
    I hope godot will continue to have their own built in physics, not become dependent.

    Personally, I'm glad Bullet is (likely) replacing the built in physics. Having worked with the built in physics in several projects, I can confidently say that Bullet is much better than the built in physics and much better for performance too. Using Bullet will also reduce the work load for those contributing to Godot, since they no longer also have to write a physics engine as well as all of the other stuff in Godot.

    That said, I do get what you are saying. I think replacing systems internally (like the physics) is a slippery slope and should be carefully considered before replacing internal/built-in systems.

    @MagicLord said:
    Exactly, before trying to race against [big game engine], first get fundation right.

    I totally agree! Without getting the foundation right, you end up in a situation like Unity, where you have to backpedal and focus (almost) entirely on performance, but with a huge code base instead of a small one. I'd much rather have to code 'fancy' features myself but have a solid foundation to build on, then have a bunch of 'cool features' but a shaky/ill-performing foundation.

    (That said, I have nothing against Unity, I'm just using it as an example. Unity has its faults, but honestly all game engines have their quirks/faults)

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 1

    @Ace Dragon said:
    Also about the physics/collision problems. Godot has an option to switch the physics engine to Bullet now (which should be far better in almost all cases compared to the original one). It's not set by default, but anyone experience such issues should try the switch before reporting.

    I always use Bullet physics.

    @TwistedTwigleg said:

    Exactly, before trying to race against [big game engine], first get fundation right.

    I totally agree! Without getting the foundation right, you end up in a situation like Unity, where you have to backpedal and focus (almost) entirely on performance, but with a huge code base instead of a small one. I'd much rather have to code 'fancy' features myself but have a solid foundation to build on, then have a bunch of 'cool features' but a shaky/ill-performing foundation.

    (That said, I have nothing against Unity, I'm just using it as an example. Unity has its faults, but honestly all game engines have their quirks/faults)

    This is what i also think, Unity have bad fundations for ambitious games with unoptimized features, they finally are working on it and bringing changes. I wish Godot would stop bringing new features for some time and work on critical bugs and polish features like the navmesh, work on good culling options, or work on the animation blending graph among others.

    You can have a GIprobe looking cool, but if the gameplay is broken shiny graphics doesn't matter.

  • Ace DragonAce Dragon Posts: 192Member
    edited May 1

    For the record, I do agree on not chasing a massive accumulation of features over a short period of time (at the expense on making sure people can make their games without battling bugs, regressions, unpolished features, or worse). Unity Tech. and Epic do place priority on the feature race and their communities are filled with comments about crippling bugs, major regressions, and slow performance.

    At that point, Godot can gain a lot of ground (among indie devs. especially) if it just polishes its existing feature-set to a crazy level that work as advertised (as many users have resigned to the possibility that game engines are just buggy monoliths and battling issues is just part of being a game creator). That's not say Godot doesn't have issue now (there are various bugs in the area of graphics for instance and particles are limited).

  • emo10001emo10001 Posts: 121Member

    @Ace Dragon said:
    At that point, Godot can gain a lot of ground (among indie devs. especially) if it just polishes its existing feature-set to a crazy level that work as advertised

    >

    Exactly.....they'll win a ton of indie developers if they go slow and get the core right. They'll certainly keep me :smile:

    -emo

  • emo10001emo10001 Posts: 121Member

    We do need to keep in mind that we can't compare all of Godot's features and/or shortcomings to Unity 2017-2018. Unity's initial 1.0 release was 12 years ago and it's 3.0 release was over 5 years later. Godot's initial 1.0 release was a little over 3 years ago with it's 3.0 release about 56 days ago. I didn't use Unity in 2010, but I'm going to bet that, at best, it was about as polished then as Godot is now.

    Comparing Godot 3.0 to Unity 2017-2018 isn't really a fair comparison.

    Just my 2 copper pieces

    -emo

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 2

    @Ace Dragon said:
    For the record, I do agree on not chasing a massive accumulation of features over a short period of time (at the expense on making sure people can make their games without battling bugs, regressions, unpolished features, or worse). Unity Tech. and Epic do place priority on the feature race and their communities are filled with comments about crippling bugs, major regressions, and slow performance.

    At that point, Godot can gain a lot of ground (among indie devs. especially) if it just polishes its existing feature-set to a crazy level that work as advertised (as many users have resigned to the possibility that game engines are just buggy monoliths and battling issues is just part of being a game creator). That's not say Godot doesn't have issue now (there are various bugs in the area of graphics for instance and particles are limited).

    Unity do racing but it's not comparable to any level. Lot of games complex gameplay and physics games have always been possible with it and have been made because the fundation is good.
    But Godot is bad on 3D fondation.

    @emo10001 said:
    We do need to keep in mind that we can't compare all of Godot's features and/or shortcomings to Unity 2017-2018. Unity's initial 1.0 release was 12 years ago and it's 3.0 release was over 5 years later. Godot's initial 1.0 release was a little over 3 years ago with it's 3.0 release about 56 days ago. I didn't use Unity in 2010, but I'm going to bet that, at best, it was about as polished then as Godot is now.

    Comparing Godot 3.0 to Unity 2017-2018 isn't really a fair comparison.

    Just my 2 copper pieces

    -emo

    Some features are available from Godot 2 when i tried it, but they never got adressed.
    Most people didn't used it for 3D games and i think not enough people are pushing enough Godot 3 for 3D.
    Another bug i encountered was about a capsule collider unable to move on top of a plane surface, the answer was to make a cube; while [name your game engine] have no issues with planes collision.
    Or the car physics that are not stable and buggy.
    Those are only few examples among many; Godot fundation has never been adressed.

    Hopefully more people will use it for 3D pushing 3D and perhaps it will get a better fundation.

  • emo10001emo10001 Posts: 121Member

    @MagicLord said:

    Some features are available from Godot 2 when i tried it, but they never got adressed.
    Most people didn't used it for 3D games and i think not enough people are pushing enough Godot 3 for 3D.
    Another bug i encountered was about a capsule collider unable to move on top of a plane surface, the answer was to make a cube; while [name your game engine] have no issues with planes collision.
    Or the car physics that are not stable and buggy.
    Those are only few examples among many; Godot fundation has never been adressed.

    Hopefully more people will use it for 3D pushing 3D and perhaps it will get a better fundation.

    I certainly hope so too. My main focus is definitely 3D. I know you say you reported bugs and that will definitely fall on us as the users, making sure we report bugs on Git, and really, on all the forums (to make sure it's seen).

    I just feel this engine has so much potential and momentum right now, so I'm all in :smile:

    -emo

  • emo10001emo10001 Posts: 121Member

    To that point, and just fyi for everyone, this is a quote from the Github page:

    "To get in touch with the developers, the best way is to join the #godotengine IRC channel on Freenode."

  • newmodelsnewmodels Posts: 119Member

    @emo10001 said:
    To that point, and just fyi for everyone, this is a quote from the Github page:

    "To get in touch with the developers, the best way is to join the #godotengine IRC channel on Freenode."

    It will not last forever. If godot's user base gets big enough, they will have to shut out everyone except for some elect.

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 2

    @emo10001 said:

    @MagicLord said:

    Some features are available from Godot 2 when i tried it, but they never got adressed.
    Most people didn't used it for 3D games and i think not enough people are pushing enough Godot 3 for 3D.
    Another bug i encountered was about a capsule collider unable to move on top of a plane surface, the answer was to make a cube; while [name your game engine] have no issues with planes collision.
    Or the car physics that are not stable and buggy.
    Those are only few examples among many; Godot fundation has never been adressed.

    Hopefully more people will use it for 3D pushing 3D and perhaps it will get a better fundation.

    I certainly hope so too. My main focus is definitely 3D. I know you say you reported bugs and that will definitely fall on us as the users, making sure we report bugs on Git, and really, on all the forums (to make sure it's seen).

    I just feel this engine has so much potential and momentum right now, so I'm all in :smile:

    -emo

    Most information about how to use something are not in the Godot docs, instead you'll find something on internet.
    And there is no game starters, many people don't know how to do something in 3D and many just give up and won't search on internet. I didn't find any step by step tutorial how could i make some plugin with right panel and buttons with mouse interaction with objects in the 3D view. Why announcing features without demos ? without tutorials nobody will care.
    https://godotengine.org/article/plugins
    I guess the new CGS tool will be thrown as it is without any docs and any tutorials, this is the problem with Godot.
    For example i didn't know GIProbe should be baked, i didn't find anything and i had to ask it, while the docs should have covered how the settings and how to use it in the editor.

    Others [your game engine name] have good docs, for each function you have a code example; for each editor panel and button you have a description or how to use it , they have game starters for anyone to right away make some level and test gameplay.
    With fundation, those needs are also lacking and why most beginners don't jump on Godot train.
    I hope they will slow down making new shiny things, instead work on things that are less enjoyable but more necessary.

  • Richard KainRichard Kain Posts: 20Member

    @MagicLord said:
    Others [your game engine name] have good docs, for each function you have a code example; for each editor panel and button you have a description or how to use it.
    With fundation, those needs are also lacking and why most beginners don't jump on Godot train.
    I hope they will slow down making new shiny things, instead work on things that are less enjoyable but more necessary.

    Documentation takes time, and it takes a different sort of expertise. And also, it is worth pointing out that Godot is an open-source project without corporate sponsorship. They don't have extensive funding. And they don't really get extra money if more people use Godot. Any impetus to make the software more user-friendly comes from idealism, not financial incentives.

    The skills necessary to make effective tutorials and documentation are not the same as the skills needed to develop a game engine. Being able to code an engine does not automatically make you good at explaining how the engine works, or producing professional-quality documentation. And taking time away from development means that less gets developed.

    You're not wrong in what you say. Godot would be easier to use with better documentation. But it is important to remember that there are finite limits. Things can't happen just because you feel they ought to. And ultimately, the best way that you can contribute is not to complain about the creator's lack of documentation. The best thing you can do is to create your own documentation based on your own experience. That is how a community grows.

  • Ace DragonAce Dragon Posts: 192Member

    There's already a full FPS tutorial as part of the online docs. In addition, there's a lot of example projects you can download that you can pick apart and base your own game from (you can download them all at once here, but you might need a nightly build for a few of them).

    In addition, code examples are already beginning to slowly make their way into Godot's built-in documentation, and people who know how functions work and have the source can contribute a patch to beef up the documentation in those areas (of course, you will need to use a nightly build to have the most up to date information, but rest assured that there have been many doc. related patches in the past months).

    Also, another point to make about polish and quality-of-life, let me reiterate the fact that there have been many pull requests made centered around polish and bugfixing for Godot in general (though of a number of them are not related to 3D, they are still important for workflow whether it's 2D or 3D). Here's an example of what just recently into Godot (a nice new group manager).

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 2

    @Richard Kain said:
    You're not wrong in what you say. Godot would be easier to use with better documentation. But it is important to remember that there are finite limits. Things can't happen just because you feel they ought to. And ultimately, the best way that you can contribute is not to complain about the creator's lack of documentation. The best thing you can do is to create your own documentation based on your own experience. That is how a community grows.

    Why sponsors money matters for features instead of Godot echo system ?
    I don't complain, i know how to use 3D in Godot, but it has been lot of trials and errors, some times i just wanted to turn on [your game engine] because it was so more easy with docs and tutorials.
    This is why i think tutorials and game starters are more important than new features (CSG and new ones).
    I hope at least critical bugs will be worked on otherwise release after release it will become a mess like some few other open source engines.

    @Ace Dragon said:
    There's already a full FPS tutorial as part of the online docs. In addition, there's a lot of example projects you can download that you can pick apart and base your own game from (you can download them all at once here, but you might need a nightly build for a few of them).

    In addition, code examples are already beginning to slowly make their way into Godot's built-in documentation, and people who know how functions work and have the source can contribute a patch to beef up the documentation in those areas (of course, you will need to use a nightly build to have the most up to date information, but rest assured that there have been many doc. related patches in the past months).

    Also, another point to make about polish and quality-of-life, let me reiterate the fact that there have been many pull requests made centered around polish and bugfixing for Godot in general (though of a number of them are not related to 3D, they are still important for workflow whether it's 2D or 3D). Here's an example of what just recently into Godot (a nice new group manager).

    Tutorials should be made by the engine team, they are the best skilled to work on those.
    You are right , there is an FPS tutorial, and this is the only one , while there should be many to showcase how you can do other game genre in Godot, or game starters.
    Requests about polishing the editor are nice, but the engine features needs polish also.

  • SeaFishelleSeaFishelle Posts: 39Member

    @MagicLord said:

    Tutorials should be made by the engine team, they are the best skilled to work on those.
    You are right , there is an FPS tutorial, and this is the only one , while there should be many to showcase how you can do other game genre in Godot, or game starters.
    Requests about polishing the editor are nice, but the engine features needs polish also.

    We've been over this, an engine developer isn't automatically a good teacher.

    The "if you made it, you're an expert of it" philosophy is incorrect. Look at eSports games; the developers of any given major game get their butts handed to them by professional teams. Why? Because those teams spend their time practicing the game and learning how to play it, while the teams spend their time writing code, making assets, and doing business things. When they play the game, they don't play it to practice, they play it to make sure it works.

    On that same level, look at Godot's documentation. It's a mess. There's a lot of content, sure, but it's unpolished, vague in many areas, and an English grammatical nightmare. That's because it was written by a French engineer who's trying to explain a codebase he's partial to and knows too well to be able to describe to a total newbie.

    Imagine having to explain the concept of a chair to an alien that doesn't have legs or knees. Not easy, is it? That's what you're asking the Godot guys to do.

  • Richard KainRichard Kain Posts: 20Member

    @MagicLord said:
    This is why i think tutorials and game starters are more important than new features (CSG and new ones).
    I hope at least critical bugs will be worked on otherwise release after release it will become a mess like some few other open source engines.

    I do agree that focusing on issues other than shiny new features would be the proper focus. Frankly, I think Godot's current feature set is far more robust than it needs to be. Since Godot doesn't have any tangible need to "compete" with game engines like Unity and Unreal, it doesn't need to sport the same level of graphical fidelity. Stability, bug fixes, platform reach, and performance are the areas that I would focus on if I were developing the core engine.

    However, I can't agree that the priority should be on tutorials, and we'll get to why...

    Tutorials should be made by the engine team, they are the best skilled to work on those.

    This is the primary point that we disagree on. Developers do not necessarily produce good tutorials. While it can happen, it does not always work out that way. As SeaFishelle pointed out, having back-end technical knowledge is not the same as hands-on experience. And hands-on experience with a tool is what you are actually looking for. That's the kind of knowledge and experience that makes for good tutorials and documentation. Building a tool is not the same as using the tool. Someone who is good at making wrenches is not necessarily going to know how to use a wrench effectively. Making a good wrench involves carving out a good mold, and getting some solid metallurgy and smelting. It does not involve wrenches. It is entirely possible to know exactly how to make a really good wrench, while at the same time being piss-poor at using wrenches.

    Expanded documentation and tutorials are usually better when they come from the community. This happens with other engines too. While some of the official documentation for engines like Unity and Unreal are decent, they are never the better options. I always end up looking to the community for answers, as opposed to the actual development teams. This is why the documentation issue is one that will sort itself out over time. Building a decent community will lead to more and better documentation. You mainly need more people using Godot, and sharing the experience they have from using it.

  • Ace DragonAce Dragon Posts: 192Member

    About the idea of tutorials and documentation being done by the developers, pretty much every FOSS application in existence has relied on the community to create tutorials, documentation, and other forms of training to help new and existing users alike.

    Granted, the result can and will vary, but it is also somewhat dependent on how the dev. team sets up the ability to edit the docs. (if it's easy enough, you get more contributors). The core team's job here is to mainly make the documentation process as approachable and as open as possible (because again, it attracts contributors who will eventually fill a number of things in).

  • emo10001emo10001 Posts: 121Member

    @MagicLord said:
    Most information about how to use something are not in the Godot docs, instead you'll find something on internet.
    And there is no game starters, many people don't know how to do something in 3D and many just give up and won't search on internet.

    Yes there is a game starter, and it's in the official documentation. There are demos on the official website that can be picked apart. From the official docs...how to start your first 3D game in Godot:

    docs.godotengine.org/en/3.0/tutorials/3d/fps_tutorial/part_one.html

    I do agree with you, there needs to be tutorials that can be found. And they can be found, right here in this forum in the Tutorials section. All of these are "how to get started on your first 3D game in Godot" tutorials. Again...not having to go scour the internet, but right here.

    https://godotdevelopers.org/forum/discussion/19165/game-tutorial-by-jeremy-bullock
    https://godotdevelopers.org/forum/discussion/19143/fps-tutorial-by-jeremy-bullock
    https://godotdevelopers.org/forum/discussion/19107/fps-tutorial-on-the-godot-docs
    https://godotdevelopers.org/forum/discussion/19292/godot-3-tutorial-series

    And again. Godot is young. Give it some time to mature. More and more tutorials will come. I'm actually working on some myself.

    Plus in this day and age, a youtube search of Godot 3 brings up a lot.

    -emo

  • MagicLordMagicLord Posts: 578Unconfirmed
    edited May 2

    @Richard Kain said:

    Stability, bug fixes, platform reach, and performance are the areas that I would focus on if I were developing the core engine.

    This is the problem, the fundation is not good if you push 3D, trying to make vehicles, collision problems and many others.
    While new features are pushed on, it's attractive, but when you start your gameplay you'll find critical problems.
    Godot 3 and 3D are not new, it had 3D on V2 but i think almost nobody used it for 3D.

    About graphics, Gi Probe is something important but it's not optimized.
    A low poly scene with Gi Probe was running 4 fps at 1080p, it's not usable and there is not enough fine tuning settings. I disabled GI probe and increased the ambient light to get some similar lighting.
    Or the lack of culling solution in Godot, i had to create a very rough culling portals system.

    I do agree Godot 3 is new on graphics and physics with other features have not been worked on.
    It lacks polish, there are still about 1500 bugs , some are minor other are critical.
    I hope this will get better after 3.1, because quality i think is better than quantity.

    You must be right about tutorials, engine coders may not be the best people to make them; but why there is monthly tutorials and demos votes if nobody makes them :D

  • Ace DragonAce Dragon Posts: 192Member

    I have to ask a couple of questions here to see if we're on the same page.

    1. What is your definition of a "good foundation", what does that look like in a game engine, how do you know when the foundation is good and new features can be added?
    2. What are your hardware specs, I ask because scenes with GI Probes can run at decent framerates on my humble AMD HD7700 1GB GPU at 1080p (not at 60 FPS, but decent)? If you are running Godot using an older iteration of Intel Integrated for instance, then you will have to pare down the graphics no matter what engine you choose.
Sign In or Register to comment.