Search Unity

  1. Unity 6 Preview is now available. To find out what's new, have a look at our Unity 6 Preview blog post.
    Dismiss Notice
  2. Unity is excited to announce that we will be collaborating with TheXPlace for a summer game jam from June 13 - June 19. Learn more.
    Dismiss Notice

Official Unity 6 and beyond: A roadmap of Unity Engine and services | GDC 2024

Discussion in 'Announcements' started by LeonhardP, Mar 21, 2024.

  1. LeonhardP

    LeonhardP

    Unity Technologies

    Joined:
    Jul 4, 2016
    Posts:
    3,136
    Hi everyone,

    Our roadmap overview that was presented earlier this week at GDC is now available on YouTube.

    Discover the latest developments in the Unity Engine and how they’re supported through additional tools and services to maximize your creative freedom and efficiency, while enabling you to produce gorgeous, immersive games that will wow players over and over.

    In this video, we show you how to achieve fantastic graphics fidelity and flexibility through rendering pipelines, shader/VFX Graph, and AI tools like Muse. In addition, with new tools like Multiplayer Center and Multiplayer Playmode, you get a head start on streamlining your multiplayer development. Next to that, we will also cover great collaboration tools such as Unity Cloud (with Azure) and Muse project-aware chat.

     
  2. Lurking-Ninja

    Lurking-Ninja

    Joined:
    Jan 5, 2024
    Posts:
    537
    "We are seeing less bugs reported", sure Jan, this happens on two ways: you do a great job and reduce the number of problems with your software or you answer every single bug report with "can't reproduce", "by design" or simply "won't fix", and people just getting tired of it and don't report bugs anymore.

    Guess which one is happening.


    Wow, just wow. Roadmap, where Unity is bragging about adding keyboard shortcuts, meanwhile the competition is bragging about complete procedural world building toolsets.

    And it isn't even a roadmap, it's just talking heads reiterate what we already knew. Nothing new or exciting. Same mistakes over and over. "We hear you", sure, you hear us, but you don't listen.

    Hilariously the Networking stack is an exception. It was mishandled for so long in the past, finally it's something!

    It is possible though that among the AI, mobile or XR things there are diamonds in the rough, these are outside of my interests.
     
    Last edited: Mar 21, 2024
  3. DevDunk

    DevDunk

    Joined:
    Feb 13, 2020
    Posts:
    5,155
    Nice overview.
    Expected just a repeat of Unite, but there were quite some nice additions.
     
  4. BackgroundMover

    BackgroundMover

    Joined:
    May 9, 2015
    Posts:
    241
    Distributed Authority sounds wild. It'll be interesting to see which technologies work together (self-hosting, relay, netcode for entities, web player..)
     
  5. PixelDough

    PixelDough

    Joined:
    Apr 27, 2018
    Posts:
    56
    Excited to try out Unity 6 Preview, and learn all the Shader Graph keyboard shortcuts!
     
    timmehhhhhhh likes this.
  6. Lars-Steenhoff

    Lars-Steenhoff

    Joined:
    Aug 7, 2007
    Posts:
    3,537
    Thanks for sharing the outlook into the future! much appreciated
     
    PaulMDev likes this.
  7. CodeSmile

    CodeSmile

    Joined:
    Apr 10, 2014
    Posts:
    6,657
    A bit unfortunate that of all things the "peak performance" part at the end has the most noticable stuttering slides (looks like 3 fps).
     
    JesOb and Ryiah like this.
  8. impheris

    impheris

    Joined:
    Dec 30, 2009
    Posts:
    1,716
    I'm not trying to be rude (i really want to love unity again) but, is it me or the statement from the intro is like "this is pointless, just get out of here" i mean, i know the forum is not full of positivism right now but, what is the point of showing features if "unity is not commiting to deliver" any of those?
    Is there any other game engine that shows some fetures while saying "we are not commiting to deliver those"?
    This is a serious question, i've never seen something like that and again, i'm not trying to be rude.
     
    soargon and XiangAloha like this.
  9. AcidArrow

    AcidArrow

    Joined:
    May 20, 2010
    Posts:
    11,984
    It's a preemptive "don't whine when some of these eventually don't make it into the engine" and also a "don't make plans based on unreleased features" which the community has been saying for a while anyway and I think is good advice.

    Now, whether there is any point in Unity talking about roadmaps is another matter.
     
    XiangAloha and Ryiah like this.
  10. impheris

    impheris

    Joined:
    Dec 30, 2009
    Posts:
    1,716
    i understand the point on showing that statement, i do not understand the point on showing features and "not commiting to release them"... then why are they showing them on a GDC? o . O
     
    XiangAloha likes this.
  11. Lurking-Ninja

    Lurking-Ninja

    Joined:
    Jan 5, 2024
    Posts:
    537
    They are planning to release these things. It's just corporate legal insanity, nothing more. Do not read too much into it. They are telling that they don't commit, because the chance that the new CEO says, "hey guys, erase that whole stuff and boot out the rest of the team" isn't zero. If that happens or for any other reasons they end up not releasing something they don't want to be held liable. That's it.

    Unity isn't the friendly indie company for a long time now. They are the proverbial evil corporate conglomerate from any random Cyberpunk or Shadowrun campaign. Just to keep in mind.
     
  12. CodeSmile

    CodeSmile

    Joined:
    Apr 10, 2014
    Posts:
    6,657
    The statement is a legal requirement so that they can‘t be sued on not delivering even the smallest feature mentioned in the presentation. As we all know, software development is not perfectly predictable. Legal sharks don‘t care. Hence the disclaimer.
     
    PaulMDev, Ryiah, PixelDough and 3 others like this.
  13. impheris

    impheris

    Joined:
    Dec 30, 2009
    Posts:
    1,716
    What you are saying makes sense, but isn't GDC supposed to be the place to show not experimental, not "under consideration" but planned to be released features? because, if not, then whats the point of unity's GDC? we have the roadmap for that (under consideration section) Is Unreal or Godot doing the same? (not wanting to make the classic vs thing) Is gdc now a place to show "under consideration" features?
    I guess is a way to avoid being sued as you said, but again it does not makes any sense to me .
    Sorry if i'm being annoying
     
    XiangAloha likes this.
  14. DragonCoder

    DragonCoder

    Joined:
    Jul 3, 2015
    Posts:
    1,745
    It is entirely a legal matter. Hence why it says in large letters "Safe Harbor Statement" which is a legal term: https://en.m.wikipedia.org/wiki/Forward-looking_statement
    Doesn't have anything with commitment. Intel, NV and co. also show that when they present roadmaps.
     
    PixelDough likes this.
  15. CodeSmile

    CodeSmile

    Joined:
    Apr 10, 2014
    Posts:
    6,657
    Again, you're misunderstanding the legal disclaimer for Unity saying "oh btw, we may not even deliver any of these". ;)

    Anytime a publicly traded company presents an outlook into the future, they have to safeguard this.

    Unity is committed to delivering all the features they talked about ... but, what if some unforeseen force prevents them from completing just one of these features, be technical or social or health or shifting resources/priorities? Someone could sue them, hence the safe harbor statement letting everyone know there's always a possibility that things may not turn out as planned.

    That's really all there is to it. ;)

    There's no reason not to expect them to deliver any of these features as planned.
     
    Daxode, PixelDough and impheris like this.
  16. andybak

    andybak

    Joined:
    Jan 14, 2017
    Posts:
    569
    Can you post the slides themselves? I'm not watching a 45m video.
     
    uzunkayacemil18 likes this.
  17. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    No Graph Tools Authoring Framework, no CoreCLR and no Visual Scripting 2.0. A new year but Unity is still the same. And now that releases cycles will be longer than a year, Unity 7 is probably some 3 years away from a production ready state. The timelines continue to be just nuts.
     
  18. MiTschMR

    MiTschMR

    Joined:
    Aug 28, 2018
    Posts:
    518
    Did you really expect Unity 6 (well, 2023.4 LTS with the old naming convention) to use CoreCLR? If anything that's for Unity 7 at earliest.
     
    DragonCoder likes this.
  19. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    I didn't expect anything. I've long since given up on that. And Unity 7 is like three years away at this point for an LTS equivalent. CoreCLR thread was opened 3 years ago. Visual Scripting and Graph Tools Foundations (now dubbed Graph Tools Authoring Framework) promises were made 4 years ago. 6-7 years sounds about right for Unity at this point.
     
    Last edited: Mar 22, 2024
    in0finite likes this.
  20. Ryiah

    Ryiah

    Joined:
    Oct 11, 2012
    Posts:
    21,560
    GDC is an advertisement opportunity for companies. Unity has always had a habit of showing off features and then not committing to their release. It's just that they're now much more open about it upfront rather than burying that info in a forum thread.
     
    Antypodish and impheris like this.
  21. Hertzole

    Hertzole

    Joined:
    Jul 27, 2013
    Posts:
    423
    This was just a big advertisement for paid AI services, paid multiplayer services, and SpeedTree (which is also paid btw). Meanwhile, the competition showcases new features available to everyone and they are much more powerful than anything showcased here.
    The multiplayer features (not the services), the ECS animation system, and the CoreCLR stuff were very good to see though, but who knows when they will be available. Certainly not Unity.
     
  22. impheris

    impheris

    Joined:
    Dec 30, 2009
    Posts:
    1,716
    Man, when i saw Unreal's GDC i was like "that is exactly what i need... oh that is impresive.... oh that's the future.... oh can i tried that?" even knowing that unreal is more suitable for teams... but it was so impresive... while Unity's GDC was like "ZZzz... is this pre GDC?, i hope this is pre GDC", i'm sorry unity but that's a fact, it was more like: "we are trying to make your game to run at 60 (;′⌒`) but hey! we have a bunch of cloud services and stuffs"
     
    XiangAloha and in0finite like this.
  23. iScriptz

    iScriptz

    Joined:
    Jul 4, 2018
    Posts:
    17
    GTF/VS is never coming man. :(

    I've followed this for as many years as you have and it's time to just give up on it and use something like Playmaker or BlooPr1nts. This GDC even showcased how they are unifying Shader Graph and VFX Graph's UI/UX but nothing about Visual Scripting. It's NOT coming, Bolt is officially dead. :( Ludiq made off like a King in all this. I'm happy for him.
     
    Kirsche, Ryiah and PanthenEye like this.
  24. impheris

    impheris

    Joined:
    Dec 30, 2009
    Posts:
    1,716
    I remember they were talking about Visual scripting on 2023 GDC, didn't they?
     
  25. PixelDough

    PixelDough

    Joined:
    Apr 27, 2018
    Posts:
    56
  26. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    No updates since then, which means nothing is happening in 2024 either. It would've been on the roadmap otherwise since it's a major update. In 2020 people were meming about Unity 2026, but at this point it'll actually be true.
     
    Ryiah and PixelDough like this.
  27. PixelDough

    PixelDough

    Joined:
    Apr 27, 2018
    Posts:
    56
    Yeah, since Unity 6 is technically Unity 2023 LTS, we're still enduring the 2023 release cycle until near the end of 2024... So we might not hear anything until late this year or early next year if they start sharing anything about Unity 7's development plans
     
    XiangAloha likes this.
  28. AcidArrow

    AcidArrow

    Joined:
    May 20, 2010
    Posts:
    11,984
    So what you're saying is Unity's ethos around features isn't a good fit for GDC :p
     
  29. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    I don't think we're hearing anything about Unity 7 for another two years at least since they mentioned they are ditching the yearly release schedule in favor of longer release cycles.
     
  30. runner78

    runner78

    Joined:
    Mar 14, 2015
    Posts:
    806
    I think there will continue to be annual releases, but we may get Unity 6.1 next year and not 7
     
  31. PixelDough

    PixelDough

    Joined:
    Apr 27, 2018
    Posts:
    56
    If Unity 6 releases later this year, wouldn't that make the next in development version Unity 7? If Unity 6 = 2023 LTS, then wouldn't new features be added to an in development Unity 7, while Unity 6 remains the stable and tech supported version?

    I guess if they do minor versions, but then what determines a major version from a minor version? A major point of the new system is to avoid confusion for people not knowing what version to use, so I would guess that releases like 6.1 and 6.2 are just updates to the LTS release, while Unity 7 receives the new features in beta
     
    Last edited: Mar 25, 2024
  32. MiTschMR

    MiTschMR

    Joined:
    Aug 28, 2018
    Posts:
    518
    With the new naming convention, they also mentioned that they first have a beta stage, then preview and only then will it be released. Unity mentioned that "preview" is the "tech stream" version, so it is unlikely for them to use a Major.Minor versioning scheme.
     
  33. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    As far as I understand, we'll see 6.1, 6.2, 6.3, etc in the coming years. 7 won't follow 6 immediately in alpha/beta or otherwise. Basically how Unity was pre-yearly releases or how Unreal does it.
     
  34. Ryiah

    Ryiah

    Joined:
    Oct 11, 2012
    Posts:
    21,560
    I don't know about GTF but I suspect the VS team was laid off.

    Or just learn C#. :p
     
    Antypodish and impheris like this.
  35. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    I've given up on visual scripting in general. Might do something custom when they finally drop GTF/GTAF, but that hasn't been available for the same amount of years.
     
    iScriptz likes this.
  36. Welfarecheck

    Welfarecheck

    Joined:
    Jun 14, 2020
    Posts:
    121
    For the most part, UVS works alright and is good enough to make a networked game for production with 1000s of players.

    Fishnet for certain you can have a large scale game. Mirror to a lesser extent. Haven't used Photon much but it would probably work out. NGO, nope. I've used many solutions but prefer Fish with VS as it's the easiest to use and most efficient.

    It would have been nice to have Unity update VS's performance for loops and whatnot. It's fantastic for developing a game rapidly as long as you structure your scripts in a way that you never use update and instead use events, while loops, control scripts, and FindGameObjectsWithTag nodes for an ECS approach to managing game objects.

    I think the lack of interest in VS is mostly driven by the lack of understand of what VS is really about. If you follow the tutorials, you're most likely going to fail miserably. It's an entirely different approach to programming. Like comparing ECS to the game object model. Folks are using it wrong and they fail because the nesting, running on update, etc. all become incredibly confusing and convoluted as you progress. Have to come at it from a completely different approach than what's recommended.

    I tried some of the other assets like uNode and whatnot. I still prefer UVS even with its shortcomings. If they would fix the loop performance. That would be the biggest update in history. I make little worker objects to split large loops and that cuts the performance cost by 80% or so but that's extra steps that shouldn't need to be done. A native UVS to C# converter would be nice too.
     
  37. PanthenEye

    PanthenEye

    Joined:
    Oct 14, 2013
    Posts:
    2,129
    Not by studio standards. I'd like to see any studios actually using it at scale, haven't heard of a single one in recent years. There was one Warhammer B-tier game some 4 years ago since WOTC licenses IP to anyone, but that's pretty much it. And GIGAYA folks also had to abandon it for various reasons. It works alright for solo indies who can't regular code.
    That's one of the things they are supposedly working on unless the team was let go, which we don't know. Seems like at least some original UVS people are not at Unity anymore.
    Like any dynamic language - it has fast iteration but poor maintenance and extensibility long term and lacks expressiveness. And writing code with something like Rider is much faster than wrangling nodes one by one so the rapidness is questionable.
    The vast majority of nodes, like 99% of them are automatically generated straight from Unity's C# scripting API. The difference is, typically there are better ways of achieving something in code than mimicking the way UVS does it. The lack of interest is precisely because it's just a C# code visualizer and not some high level design tool like Blueprints.
     
    Ryiah likes this.
  38. iScriptz

    iScriptz

    Joined:
    Jul 4, 2018
    Posts:
    17
    UVS is not ready for primetime by any stretch of the imagination. It became extremely unwieldy and the service worker started crapping out after 4 months into a mobile game project with 50 custom nodes baked into it.

    I had LLMs taking plain C# scripts and putting them into UVS custom nodes for me.
     
    Ryiah likes this.
  39. Baste

    Baste

    Joined:
    Jan 24, 2013
    Posts:
    6,346
    Can we hire some runners to hack into their systems and introduce some Playables patches?
     
    PanthenEye, impheris and timmehhhhhhh like this.
  40. Ryiah

    Ryiah

    Joined:
    Oct 11, 2012
    Posts:
    21,560
    If you want anyone to believe that you're going to have to provide a source. Because the GIGAYA team tried it out and @Andy-Touch very publicly mentioned that it had to be ripped out due to performance problems and that was a single player game. Maybe it's usable if the entirety of your logic is server-side code and UVS is just driving the basics like UI and passing input to the networking but then you're not really making a networked game with it.

    No, it's a side effect of how UVS is one of the worst VSes and C# being one of the easiest programming languages for game development. Blueprint is far beyond pretty much any other VS that isn't highly domain specific because it provides abstract nodes that implement many lines of code rather than a one-to-one translation.

    If I remember correctly Bolt 2 looked like it would compete with Blueprint but Unity shut down that project before it left alpha instead focusing on the one that was worse than the competition on the Asset Store. It's a shame but it's also typical of a company that doesn't understand what people actually want.
     
    Last edited: Apr 3, 2024
  41. PixelDough

    PixelDough

    Joined:
    Apr 27, 2018
    Posts:
    56
    I still cry at night thinking about how amazing Bolt 2 looked... At this point I wish they could just put its source code online and let the community finish it...
     
    POOKSHANK likes this.
  42. Welfarecheck

    Welfarecheck

    Joined:
    Jun 14, 2020
    Posts:
    121
    Lot of people fail with UVS because they use update everywhere. 5000 updates doing nothing will destroy your performance. And, 5000 objects needing scripts isn't many in a game. You should have a handful of updates in your VS scrips and those should be only be control scripts using FindGameObjectsWithTag with an ECS approach if there's multiple objects. Everything must be done with events and while loops. You can fire any client side action with an RPC event and while loop with very little performance cost.

    Update node with a single if statement connected and not connected. The update cripples your project.
    upload_2024-4-6_18-41-37.png

    While loop objects vs 5000 update objects.
    upload_2024-4-6_18-34-20.png

    If you followed the Bolt/UVS instructions, you did it wrong. There's a misunderstanding of how to use UVS and it leads to failure for most who try to use it.
     
  43. iScriptz

    iScriptz

    Joined:
    Jul 4, 2018
    Posts:
    17
    I really do wish that just steering clear of the Update Node was enough to be able to use UVS. There's just the issue of serialization breaking ALL THE TIME. There was also the issue of the editor/IMGUI not playing nicely if I had over 50 variables stored on a single gameobject's blackboard. God forbid I made a custom object and tried to get it to display in IMGUI/Blackboard that introduces lag and weird issues around that.

    In my game specifically I used a mix of UVS and C# scripts. C# to do the heavy lifting and jumping back/forth between C# and UVS. Was a wonderful symmetry. After so many custom nodes and imported scripts and building for 4 months it just kept causing the service worker so many issues. Intermittent crashes, lag spikes, slow to rebuild the nodes, and god knows what would of happened if I continued with it.

    I really wish it were just a matter of steering clear of the Update node. I'd be ecstatic. But its not. :(
     
  44. POOKSHANK

    POOKSHANK

    Joined:
    Feb 8, 2022
    Posts:
    368
    visual scripting will sadly never be viable again, if unity had just waited a bit we could've had bolt 2, but now we only have something that can barely support a small game with few scripts. i had the same happen to me, figured out after a couple months that it just isn't ready for a full length game at all. not to mention the more nodes the slower traversal becomes

    luckily c# is basically just as easy, just takes a bit more thought and with well documented code, it can be easier to understand.
     
    iScriptz likes this.
  45. Welfarecheck

    Welfarecheck

    Joined:
    Jun 14, 2020
    Posts:
    121
    I dunno, I rarely use the UVS variables. For network objects I have over 50 variables multiplied by Nth cliens/npcs/objects and there's no performance problems at all.

    Performance is quite good when used correctly. For example; I am moving 1344 server authoritative Network Transforms randomly for all clients and this is my profiler. No spikes or craziness as I iterate large numbers of objects in UVS.

    upload_2024-4-9_19-15-23.png

    Every 1 second I am checking all movable objects to see if they need to move. If it needs to move the object then moves and shuts down to use no resources. If I had updates running on those objects or ran them on a big loop, the profiler would be 30ms++ crippling the game.

    upload_2024-4-9_19-17-35.png

    It's important to treat objects with an ECS approach while splitting the lists and spreading the functions over a few frames. In this case, I am splitting the enemy network objects into 10 lists, and running the lists over half a second leaving wiggle room for other processes that may come into play later on.

    FindGameObjectsWithTag is by far the most effective and efficient method of running processes. The arrays are managed by unity and you don't need to manipulate lists in the flows at all. Just split the list and isolate the objects you need to run for any given function using while loops.

    Splitting objects into lists;

    Code (CSharp):
    1. using UnityEngine;
    2. using System.Collections.Generic;
    3.  
    4. public class SplitObjectsByTag : MonoBehaviour
    5. {
    6.     public string tagToFind = "Tag"; // The tag you want to find objects for
    7.     public int listCount = 5; // List Count
    8.     public List<GameObject>[] splitLists;
    9.  
    10.     public void UpdateLists()
    11.     {
    12.         GameObject[] objectsWithTag = GameObject.FindGameObjectsWithTag(tagToFind);
    13.  
    14.         splitLists = new List<GameObject>[listCount]; // Array of Nth lists
    15.         for (int i = 0; i < listCount; i++)
    16.         {
    17.             splitLists[i] = new List<GameObject>();
    18.         }
    19.  
    20.         // Splitting objects into lists based on their index modulo 5
    21.         for (int i = 0; i < objectsWithTag.Length; i++)
    22.         {
    23.             int index = i % listCount;
    24.             splitLists[index].Add(objectsWithTag[i]);
    25.         }
    26.  
    27.         // Outputting the counts of each list
    28.         for (int i = 0; i < listCount; i++)
    29.         {
    30.             Debug.Log("List " + i + " count: " + splitLists[i].Count);  // remove me later
    31.         }
    32.     }
    33. }
    34.  
    Using UVS to make a MMO is fully doable. People just use it wrong because they followed the instructions and ended up in some incredibly inefficient convoluted process of having a zillion scripts everywhere that becomes unmanageable and unusable due to performance.
     
  46. AcidArrow

    AcidArrow

    Joined:
    May 20, 2010
    Posts:
    11,984
    And there was a dude that did paintings with excel, "doable" doesn't mean much.
     
    PanthenEye and POOKSHANK like this.
  47. POOKSHANK

    POOKSHANK

    Joined:
    Feb 8, 2022
    Posts:
    368
    so the only way to use it viably is by doing all this research yourself? of course people are going to follow the instructions given to them, they wouldn't instinctively assume that it's wrong if it's straight from the people who "should" know best.

    why bother jumping through invisible hoops when c# is simple? i just don't see a reason to do everything in counterintuitive ways, and i'm a firm believer that limitation blossoms creativity

    not to mention all your techniques are just as viable in c#, but for you it's a requirement to add, not a choice
     
  48. AcidArrow

    AcidArrow

    Joined:
    May 20, 2010
    Posts:
    11,984
    Does UVS also take forever when adding or removing a single node, or god forbid, undoing? Or is that a shader graph special?
     
  49. POOKSHANK

    POOKSHANK

    Joined:
    Feb 8, 2022
    Posts:
    368
    the beauty of uvs is that the more nodes you have, the slower the editor runs, not to mention undoing will randomly decide to move your nodes just enough that the lines aren't straight anymore
     
    iScriptz and AcidArrow like this.
  50. futalihua

    futalihua

    Joined:
    Sep 13, 2023
    Posts:
    48
    Thank you very much for strengthening and upgrading the engine. We are fortunate and I am witnessing a moment in history. Defeat Unreal and wish you success.;)