What's your biggest problems/missing features with Unity ?

What's your biggest problems/missing features with Unity ?
From engine to editor, UX, code, etc.
Let it all out.

 

Closed Source. Even if only for pro users, full source code would be nice. Especially considering enormous performance issues with things like moving transforms or (when first released) a completely broken UNET.

Having to rely on Unity means you have to pretend the feature doesn't exist or the issue will never be fixed. i.e. You cant rely on them. It could be years until a fix comes. You cant make games on an "if".

 

Terrains and caves. I think the current terrain system is the oldest feeling part of the engine.

 

Ability to connect inputs to animations without the need to code. Though Playmaker makes this a moot point.

 

Thought this was an Arrowx thread when I clicked on it

 

Unity isn't technically 'Closed Source'. You can actually buy read or write source-code, for an amount of $$$. And there are areas that have been 'Open-Source-d' (UI System, Cinematic Image Effects, etc) via our BitBucket (https://bitbucket.org/Unity-Technologies/) in an effort to move more areas, that users want to modify and change, to become more accessible.

And which fixes have taken years to come?

 

You can already do this if your input is EventTrigger-based. Just bind the EventTrigger input to an Animator and set what you want it to do.

Although, even writing basic input (Click a Keyboard Key, Click a Mouse, etc) is very simple to hook into Animator, using C#.

 

If this would have been an Arrowx thread, it would have been called ´ All the things Unity sucks at and refuses to fix´

 

This type of stuff is so easily learned. If people spent half the time learning C# as they spend learning tools to avoid learning C# they would be much better off.

 

Unity is pretty great to be honest.
Sure, there are tons of things that are pretty S***ty at the moment, but it is whining at a (very) high level!

• .NET Runtime and C# Language Version upgrade
  
  There's an experimental preview and some Unity dev said on the forums that its a matter of months not years until this is finally done. But I'm still anxious about the possibility of more delays and I'm not sure if they higher ups really get just how important this is...
  
  
• Performance
  
  In the game we're developing we constantly have to worry about performance, the main problem here is how invasive pooling techniques become when you really want to get performance out.
  For example when you want to pool polymorphic classes, you have to add some something like `public abstract void ReturnMeToTheCorrectPool()` otherwise you can't make use of generics to let the language infer the right pool (because the type is dynamic).
  
  Also coroutines allocate. I know why they do that and Unity should automatically pool and reuse enumerators and c# generated statemachines. I believe this is an issue that Unity should take care of because we can't (since we have no control over the runtime and how coroutines are implemented)

All in all im extremely happy with Unity, but the scripting side of things and performance have some pretty deep problems.

 

More stuff for 2D. Especially build-in tool for 2D animation.

 

whats wrong with the existing animation + animator windows? Not saying there isn't a problem, I'm just curious. (I don't even use 2D at all except for the UI )

 

This could be the first response and the thread could then be locked as "Answered" as far as I'm concerned....

 

I miss skeletal animation (+IK) and sprite deformation.

 

Relevant to your Interests:

https://blogs.unity3d.com/2016/12/02/unity-expands-2d-offerings-with-anima2d/
https://blogs.unity3d.com/2016/06/13/2d-experimental-preview/

 

Thanks, I have read about anima. I am waiting for this tool as build-in solution, because I don't have trust in external tools

No need for something like this:

 

I've never had to write code like that and I use pooling extensively for 4 years (in fact we wrote our own plugin for it). Can you explain why you'd need to do that? PM is fine. I'm actually hoping that our plugin handles that better than whatever you're using, and if not, I'd hope to be able to add that so you don't have to code it.

 

I'm gonna say Unity should have its own included visual programming thing, "like" Playmaker but not necessarily an FSM. Maybe something like NodeCanvas or Blox2 are better / more versatile.

 

Disagree - people are better off specializing in there chosen field instead of trying to be marginally good in all fields.
If people wasted time learning someone elses craft they would not be as good at there own craft - as evidenced in marginal quality content from those who try to do everything themselves.

Don't assume it is easy for someone just because you find it easy yourself. I think character rigging and animation is easy - I don't assume it is easy for others.

 

We are working on it!

 

Most of the people here are pointing out the inner-workings of Unity such as the source code or .NET, but I haven't gotten that deep so here's what grinds my gears as a user:

• Slow light baking. I'm happy that the PowerVR is in 5.6 beta since the demo looks amazing, but I just can't get my head around how my GPU can calculate the shadows for my scene over 300 times a second (when playing) and then take 5 hours to bake. How about the people on MacBook Airs with dual-cure i3s and Intel HD graphics? If a quad-core i7-4790 and Nvidia GTX 960 struggle with light baking, I can only imagine the pain people with worse computers (and Mac owners with Intel Iris graphics) experience.
• Line renderers. In 5.5 they got a massive upgrade, but as I outlined my woes in this post and UT stating they still have bugs, they really need to be fixed. Because of them I've had to create my own system to draw lines between 2 points, and it still isn't finished the way I want it to be. Probably by the time I'm done with that Unity will fix the line renderers and all that time will be wasted.
• Terrain system. It's already common knowledge that the Terrain system is to be used in simple projects or prototyping only. The fact that you can't make holes in the terrain or draw overhangs shows that it's really lacking.
• Performance. We've seen the Unity vs Unreal performance videos, and Unity does really well. However Unity is a complete battery sucker on mobile, even in completely static scenes. This article really explains just how bad the battery drain is compared to Unreal.
• Physics bugs. This might be nitpicking, but small colliders at fast speeds can go thru objects or get stuck entirely. I have no idea how you're supposed to make realistic bullet simulations in Unity if a high-speed small collider gets stuck in objects.
• The UI. I like how simple the Unity Editor UI is. However, it is starting to look dated. Even worse, it lacks the "Make an awesome game now" button.
• Buttons in the In-Game UI. Seriously, how can you make a brand-new UI system, and give buttons only a OnClick function? How about OnButtonDown, OnButtonUp etc... like with the Input system for keys? To make an OnHold function, you need to use an Event System.
• Monodevelop. Come on, just bundle Visual Studio already. I'd even rather use Notepad++ over Monodevelop as at least it doesn't crash every other time I try to open it.

 

Light baking:
Those are completely different systems. Shadows are essentially a 2D projection of a light, while global illumination and light "transport" work in a completely different way. Light "transport" in general works on the whole scene, in "3d". It really can't be compared.

Physics:
Thats expected, and thats why unity has continous collision modes on colliders/rigidbodies if you need them. Or you can do custom movement if its a special game element.

UI: I sorta agree, the default button could have used some more default events, but its really no problem to subclass `Button` and add IPointerDownHandler.

 

Don't they already do this in Windows by default? I know that was one thing I had to get used to while I was doing tests on my PC.

 

Or just throw an EventTrigger Component on the button. Problem solved.

 

Yup, at least on Windows. When I downloaded for the first time in December it automatically downloaded and set up Visual Studio.

I think Unity should implement Blender shortcut keys for the scene editor, and make Z up....................(just kidding).

 

a proper design tab for doing ugui layouts one where you dont' have specifically micromanager and hide every other fking layer that is blocking the ugui components.. like terrain etc dumb as hell unity...or do some other someretarded workflow like having ugui in a separate scene..

Complete with better tools for dealing with varying screen resolutions., dunno wtf the aspect/screensize dropdown on game view doesn't have more of the common display resolutions listed.. yknow all that fcking google analytic S*** you bundle with every build that end users have no option to disable.. surely you can provide a list of most common screen sizes ...same with performance, half a dozen 3d benchmarks can do it, but not Unity collecting fk tons of data.. still can't be bothered to make use of that data in unity editor so its easier for developers to see where a game would perform on certain hardware configurations.

The Editor ux could do with some functionality improvements.. ie detached editor windows with a bit more options like always-top on/off and maybe have them also show up as taskbar icon.. like how Visual Studio works if you detach panel to a floating window.. I like to be able to detach the game view while its at a larger resolution.. and have it in the background, but noo can't have that because its treated as always on top.. so you have to fk around moving it all the time or having it maximized in the main editor window, but then you can't see the effing inspector panel, unless you detach that, and if you do have it detached it will always be on top of the main unity window.

The actual gui style though is fine, I wouldn't want to see it become some sort of sharp cornered squared flat pastel colored spaced out padded winblows10 looking abomination. I think they handed the rest of the Windows stuff over to morons as can be see with the OS going down the toilet starting with win8.

unet anyone.. fps controller with working unet support and other decent features you'd except for a decent authoritative based server?

I still install monodevelop even though its awful compared to Visual Studio in everyway except it at least provides a decent built in assembly viewer so you can check the decompiled c#..which honestly helps a lot when documentation on someones asset package is garbage and they've packed it into a dll.. shame Visual Studio's version of that is a useless pile of crap.

But yeh is so many things Unity could do to improve baseline only have to look at UE default features, for Unity its either make your own or find an asset that can improve that area. Some things really should be built in though, but then they build them in and its... like.. ok , yeh, why even bother if you aren't going to do it properly.

 

Have you read this tutorial for Precomputed Realtime GI?
https://unity3d.com/learn/tutorials...uction-precomputed-realtime-gi?playlist=17102
Also, PVR in 5.6 should help massively with Baking Static Lightmaps.

Have you bug-reported the bugs that need fixing?

On the Roadmap! https://unity3d.com/unity/roadmap

We have a performance team who look at these sorts of things; I will raise it with them!

You shouldn't really be doing physics colliders checks on really-fast moving things like Bullets. Watch this video: https://unity3d.com/learn/tutorials/topics/scripting/intro-shooter-mechanics

Have you used the EventTrigger component yet? It supports all the inputs you have referenced.

As other users have mentioned before, we already do this on Windows.

 

Vertex painter
shader editor

 

Terrain + Vegetation!!

bu..bu..but it's been on research for years

I know things take time, i'm just really eager to see the new Terrain system (and the vegetation that is coming first, as it was said at Unite i believe)

 

Wow, I actually didn't expect UT to reply, so huuuuge thanks!

I have heard of precomputed GI and used it, however on mobile it is still a large performance drag. However I'm glad to see major improvements in the roadmap!

The UT representative in the post I linked did say that they have issues with the Z axis with Line Renderers, so UT already knows about my issue.

The Roadmap just says "Terrain System" in the Research section, so I hope that will turn out well.

As for the physics colliders, I'll watch the whole video, however from the scripts it looks like you're doing raycasting and AddForce. AddForce does work most of the time, but it's an issue when it comes to small colliders at fast speeds.

Now I'll add a spoiler as here's some boring math:

Spoiler: Boring physics calculations

So let's get some numbers to do a realistic simulation in Unity.
The muzzle velocity of a Russian AK-12 is around 900mps. The length of the barrel is 415mm. So, we know that in 0.000046 seconds, it can travel the length of the barrel.
T = Distance / speed, so time = 0.042 meters / 900mps = 0.000046 seconds

Now let's get the acceleration
Acceleration= 900mps / 0.000046sec = 19,565,217meters per second squared

And the force:
Force = mass * acceleration, bullet around 0.03kg
Force = 0.03 * 19565217 = about 590,000 newtons.

Good luck trying to do that in Unity, or even at 1 100th of that speed.

So I have to say, the physics system is incredibly good (I've used it to cheat on many physics assignments perfectly). But it looks like we'll be stuck with raycasting bullets.

And of course, to press and hold a button, I've had to use the EventTrigger, it's the only way
I just think that if the button already has an OnClick option, it would be nice to have an OnDown, OnUp and OnHold functions too, instead of using an EventTrigger.

As for the installer on Windows, I can only see Visual Studio tools for Unity, but not Visual Studio itself. However, Unity is in the Visual Studio 2017 RC installer. Maybe the tools include VS, if they do, I'm sorry for not knowing that.

Also, after some time I've had to think this over, I have to include Standard Assets. Yes, they include an airplane, AI and character controllers. However UE has much more, and it's really much more inviting for a new user.

After all, this was about describing Unity's problems. Most of the stuff I described aren't problems, but annoyances. Anyways, keep up the great work!

 

I'll also second this - since its something I'm working on right now.

 

As others have mentioned before, Blender style scene navigation and shortcuts
They just make everything so much faster

 

I never used blender, but from what I've heard others say, its UI and default keyboard layout is a complete nightmare.
So that scares me a bit... what'd be the top 5 keys/key-combinations if this were implemented and what would they do?

May I ask what the scope of this thing you're brewing up is?
I never used playmaker or anything similar, since I'm a programmer through and through

Will we "real" programmers be able to abuse it for efficiently scripting small things?
Like people use mecanim to implement some basic AI, but it has a lot of overhead compared to directly scripting it.

For example letting people design their own spells, so they can drag-drop effects somehow, and it will still be efficent (no gc alloc and not 10x slower than if I'd have written it myself)

 

A few more "visual programming" posts and it will be.

 

visual shading/ visual programing. This means blind design where the developer has no idea what is going on, this will be the coast of flexibility...Don't hide the reality from the developer, give us flexibility not handicap options.. Everything is pretty visualized in the editor anyways..

The only feature i want, are new version without bugs, and that are better than the previews build. Also please redo the whole concept of Mechanim. its a mess to work with (because its visualized logic)


The real truth about revolution..

 

dadude,
I think I was the first to suggest Blender shortcut keys, but it was meant jokingly. Most complaints about the blender interface comes because it is different from industry standards. But if your like myself and thousands others who start on blender, it is not as bad as it sounds.

But to answer your question, it is really easy to navigate blender's editor scene using only keyboard commands and a minor amount of mouse work. There is a gizmo, like in Unity, but most of the time I hide it. So I do think Unity scene view could use a few more keyboard shortcuts.

 

But that is not what I am saying. What I am saying is that the C# that an animator or modeler would need to learn is so minimal that it is harder for them to actually learn a new tool to avoid that C#. It isn´t about ´clicking´ .. it is simply about time.

 

Open all new notifications in a new tab for the website gets my vote.

I'm a simple man.

 

I agree, the concepts are the same, if programming doesn't "click" for you at all, then a visual tool won't help either.
Its essentially the same thing in text form, and with C# its really so easy to at least get the basics down...

An official "Visual Programming Tool" can potentially bring us a waves of new S*** games on greenlight...
But to be fair, its not the fault of a tool then (Steam and the retards buying these games are at fault).

Im all for making stuff easy to use. It should be made so the available tools are as simple as they can be, not more!

Because that's when you enter a zone where you have to manage stuff for people, and if they come to see it as "no need to learn programming, do everything in the visual designer thing lol" then all you do is trap people in some false sense of ability, and once they need more performance and/or flexibility than the tool can provide, they'll have to learn anyway - but this time they have it MUCH harder because:
- they instantly want to start with a problem that is so hard, that they can't even solve it in the UI
- they have no experience in programming because they even did the easy parts in the UI.

So now they're stuck with either relying on Unity to keep making improvements, or they give up, which would be sad...

But I don't want to paint an all too gloomy picture here, maybe things will turn out just fine



edit: disclaimer, I'm using shaderforge myself. But I accept the fact that I will forever be limited by it in one way or another - and acknowledging that is the most important part when using such tools imho.

 

I have already both , but these should be inside Unity package as i find them almost essential to create better visuals.

 

I am curious about visual programming, exclusively because it hasn't taken over. Clearly that implies that it is not all that better, or potentially even worse, than conventional programming. But I have absolutely no idea why, having never touched the thing.

For syntax, sure. To understand 'how to program' or 'engineer code', not at all. A lot of people have problems learning this. At least until it clicks, perhaps?

The floodgates have already been opened wide. Additional waves would be the equivalent of any planet being added to the sun's mass.

​

100% agree with you here. Steam's lack of quality control is an enormous problem, and one of my biggest complaints. Really hurts indies a lot, IMO. Obviously it effects gamers too, as indies will experience fewer sales (less funding) AND the gamers will have a harder time finding the games they'd actually enjoy.

However I don't believe Steam curating content is a good idea either. No monopoly should have exclusive control over what games live or die. Steam as a whole is toxic to the industry, and IMO needs to end ASAP. GoG is doing a good job to help, but it simply isn't enough.

I'm actually surprised that other engines are less successful than Unity. Perhaps it is because Unity is a 3D game engine, and the majority of "Easy" engines, like Construct2, are purely 2D.

Although I have brought up the same point in the past, I feel the need to play devil's advocate and remind you of RPG Maker, and all of the games made using it. Some even very successful ones, like To The Moon.

​

And Game Maker games.



These engines are known for their "You don't have to program!" style.

This already sums up the majority of Unity users. In fact, gamedev in general consists mostly of the latter (giving up). I don't see helping non-programmers increasing this fact. Quite the contrary, I think it would lessen the number very very slightly.

I have very optimistic hopes for the future of game development and game engines. If I were to ever enter into Minecraft-levels of wealth, I would immediately pour an enormous amount of money into creating a game engine that focuses on both performance & usability (separately) combined with a community of collaboration (funding developers to build assets for the engine, having cash reward contests for best assets, etc.) with a heavy focus on quality & strict conventions/rules for all contributions.

That maybe some day we could have an engine where anyone, literally anyone, could create any game they wanted. Or at least get closer to that ideal. Of course... that could destroy the gaming industry if literally anyone could make anything, due to the unimaginable flood of games that would result. But maybe that wouldn't be a bad thing in a futuristic society where robots rule the world and human jobs are strictly of the arts. (I imagine it would take a long time to get to that point, or maybe just optimistic about robot inventions. Up until the point where the robots revolt and terminator us.)

 

An automatic event bubbling system for the UI would be nice. I can write my own, it is open source. But it would be nice if I didn't have to.

 

Story time: (sorry if you know this already)

Programming didn't click with me for YEARS before I finally got the abstraction and was able to write more than hello world.

The only reason I got past my roadblocks was visual scripting tools. I wrapped my head around basic concepts through my use of kismet and the material editor in the UDK. I was able to practice composing more traditional scripts in Stencyl using their lego script segments.

As I got better, I was determined to keep learning, because as we all know, You just can't make the game you see in your head without programming. So when I started to learn to code again, I was delighted to find that my baby steps had taken me far enough to learn how to code.

However you feel about Unity implimenting systems to make game development more accessible, it kind of stings to hear people taking on an elitist tone about programming.

To me, it's worth taking attention away from people who want to make games, and instead focusing on how Unity can keep serving our (programmer) needs while balancing the needs of other devs.

 

Thank you for this post. I think it is very important for the community's health to hear stories like this.

Game developers in general, especially experienced programmers, can be extremely elitist. It's one of the common characteristics I find associated with competence in a technical skill... and it's not a pretty one.

So it's nice to see elitist attitudes get knocked down a few pegs by the truth of the fact a lot of what exists, despite the elitist's disapproval, is valuable. (I've seen elitists bashing all container types besides arrays, as if somehow Maps/Linked Lists are ALWAYS horrible for game development, even though they're used in nearly all the books I've read about game programming.)

 

I'm sorry if what I said came off as elitist.
I know how hard it can be to learn how to program, I fought my way to knowledge myself

And yes the elitism is a slight problem under programmers. I think it has to do with the fact that you can measure every single modification that improves performance, but you will have a much harder time objectively measuring code-quality, maintainability, and readability. And of course you also can't really express how good a model / texture / sound is in numbers. Its a matter of taste, while counting cpu cycles is not.

Making Unity more accessible to people with ideas is indeed a good thing, and if a tool is being made that helps to more easily convert these ideas into a real game then I'm all for it.
All I meant to say was that people should always keep their expectations in check when it comes to these things

I don't really understand that part.

 

I mean, I use arrays as often as I can, but coding without anything else sounds horrible

 

Before even considering making every thing visualized "click to develop" in the editor. For the love of god give us higher than 10pt font in the editor, i have a 4k monitor..