Search Unity

  1. Welcome to the Unity Forums! Please take the time to read our Code of Conduct to familiarize yourself with the forum rules and how to post constructively.
  2. Have a look at our Games Focus blog post series which will show what Unity is doing for all game developers – now, next year, and in the future.
    Dismiss Notice

2019.3 entered the final stages of beta testing

Discussion in '2019.3 Beta' started by LeonhardP, Dec 12, 2019.

  1. LeonhardP

    LeonhardP

    Unity Technologies

    Joined:
    Jul 4, 2016
    Posts:
    3,021
    Hi everybody,

    We have entered the release candidate phase of the 2019.3 beta and are hoping to officially reach the release status in January 2020.

    Thank you all for your contributions throughout the alpha and beta phases of 2019.3 so far. Your help is much appreciated and we take your reports seriously. Please continue to reach out via bug reports and this forum if you are encountering any unknown issues with this latest release.
     
  2. nsxdavid

    nsxdavid

    Joined:
    Apr 6, 2009
    Posts:
    475
    Congrats!
     
  3. Xtro

    Xtro

    Joined:
    Apr 17, 2013
    Posts:
    481
    Car companies deliver the new car of next year in this year.
    Unity delivers this year's version number in the next year.
    :) :) :)
    No offense. Just kidding :)

    Thank you for all your efforts.
     
    MegamaDev, fherbst, gcteam and 7 others like this.
  4. Endlesser

    Endlesser

    Joined:
    Nov 11, 2015
    Posts:
    73
    Now since its release date gets quite close to 2020.1's,Im so confused to use which.
    Out of preview HDRP is definitely a huge step in next 2019.3f.
    Full package of Entity and Netcode are also mostly wanted in next next 2020.1.
    Always support new unity and hope it gets better and more product-ready.
     
    JohnC_Unity likes this.
  5. TextusGames

    TextusGames

    Joined:
    Dec 8, 2016
    Posts:
    424
    You should have written it 2 weeks ago then you misleadingly publish 2019.3.0f in stable branch of Unity hub. And many users upgraded their projects.. But, thanks for letting us know, better late then never.
     
  6. Kolyasisan

    Kolyasisan

    Joined:
    Feb 2, 2015
    Posts:
    372
    ...something's not stable, am I right?
     
  7. MadeFromPolygons

    MadeFromPolygons

    Joined:
    Oct 5, 2013
    Posts:
    3,625
    Before anyone says this is off topic, this is in response to this specific release claiming HDRP is production ready, with URP having been production ready apparently months ago:

    When is URP going to be actually production ready? Im sick and tired of waiting for it to be properly usable and your official position seems to be that it AND now HDRP are production ready. Like, what dreamland do you actually operate in where you cant see what everyone is saying and feeling?

    Make an official statement via the blog about the state of URP and what you are doing to get it to be actually production ready, and make a statement that in built will not go UNTIL all and I mean ALL of the issues surrounding URP are gone. If you can do it using in built but not using URP, then its not ready.

    Right now there are comments on the forum and twitter from unity staff stating in-built will be removed at some point, but no indication of when or what state URP will be in when that happens. This amounts to more uncertainty.

    Can you do custom FX in URP?
    Can you do custom lighting without some gimmick hack to get it to work?
    Can you 1:1 do every effect or visual style in in-built in at least one of the SRP?
    Is it clear what version of URP does what and what should be used? Is the docs for it up to date?
    Are breaking api changes continually happening + API in constant flux?

    Then where does that leave everyone? Because its not at a point that is "production ready".

    If your aware of this, make us aware that you are aware instead of relying on singular comments from developers all over the forums to tell the story of what is going on.

    2018 and 2019 can be summed up as: Great plans, terrible delivery and communication. Overall it amounts to a whole lot of headaches for users +mistrust and free very easy marketing for unreal at a time when they are swimming on piles of cash whilst we unity users are swimming on piles of bugs and unfinished or outright abandoned features.

    I am still trying to voice this in a constructive way, but seriously it is getting harder by the day for as long as no official announcement comes about you guys actually knowing how bad it is. We shouldnt be relying on reddit and twitter posts from people like @willgoldstone to determine that unity are actually paying attention. (Thanks will for taking the time to aleviate peoples worries, it is greatly appreciated).

    I <3 unity but even that has a limit guys. Take a look around forums, reddit and blog post comments and you will find I really am not the only long-term and professional user of unity who is feeling this way.
     
    Last edited: Dec 14, 2019
    remption, midian808, Leniaal and 44 others like this.
  8. Deozaan

    Deozaan

    Joined:
    Oct 27, 2010
    Posts:
    707
    2019.1 didn't release until about mid-April. So if that's any indication as to when 2020.1 is likely to release, it's not that close together.

    Also keep in mind that by the time 2020.1 releases you will still have about a full year of changes and new bugs to deal with. Meanwhile 2019.4 LTS will release which should (in theory) only get more stable and bug-free as time goes on.
     
  9. Deozaan

    Deozaan

    Joined:
    Oct 27, 2010
    Posts:
    707
    The blog post says 19.3 has over 170 new features, but the new features preview thread only lists ~14 of them as I write this.

    Can it really be that less than 10% of the new features are noteworthy?
     
  10. interpol_kun

    interpol_kun

    Joined:
    Jul 28, 2016
    Posts:
    134
    Yes, this is a common practice among all software releases. All the minor features count towards that number, some are worth an article in blogpost, some not.
     
    MadeFromPolygons likes this.
  11. LeonhardP

    LeonhardP

    Unity Technologies

    Joined:
    Jul 4, 2016
    Posts:
    3,021
    You can find a complete list in the release notes: https://unity3d.com/unity/beta/2019.3.0f3
     
    Deozaan likes this.
  12. BakeMyCake

    BakeMyCake

    Joined:
    May 8, 2017
    Posts:
    173
    A perfect setup for a joke about the rest 156 new features being regressions :p
     
  13. taylank

    taylank

    Joined:
    Nov 3, 2012
    Posts:
    182
    I have to agree with Isaac. The last two years have been incredibly frustrating, dealing with features that were clearly not production ready when they were advertised as such.

    Even simple engine features that are not tied to any fancy systems are breaking under the weight of too many things getting crammed in. I'm still waiting for a fix to the bug that breaks nested assets when you try to rename them. Let me say this again: renaming assets breaks them. The fact that Unity brags about new features when their absolutely essential functions are unreliable infuriates me.

    I have lost precious development time researching and reporting bugs that should have been obvious to anyone seriously evaluating the production readiness of Unity builds. I have ugly hacks in my code trying to cover my ass for Unity's own bugs. I have lost so much time trying to figure out how basic functions of these 'production ready' systems work because the documentation isn't there and onboarding is a joke.

    There is a lack of awareness that the people who will be the early adopters of these new advanced systems will be highly technical, professional devs, rather than hobbyists. And yet the quality of onboarding in most cases is at a "Hello World" level. The examples cover only the most basic use cases, which makes me think:
    A) Unity dev teams don't understand the needs of the professional dev community, or
    B) They do understand but are also aware the tools can't currently meet professional production standards.

    And yes, I am excited about the new features and I do believe they will solve real problems when they are seriously ready, but right now I feel like I'm being taken for granted.

    Thank you again, Isaac, for getting the conversation started. I hope someone from Unity will acknowledge what we're talking about here and talk about how they are going to address the problem.
     
  14. valarus

    valarus

    Joined:
    Apr 7, 2019
    Posts:
    399
    Is recommended to upgrade 2019.2 project to 2019.3, or wait for LTS release?
     
  15. laurentlavigne

    laurentlavigne

    Joined:
    Aug 16, 2012
    Posts:
    5,747
     
  16. Deozaan

    Deozaan

    Joined:
    Oct 27, 2010
    Posts:
    707
    Where did that number (170+) in the blog post come from? I counted the bullet points in the features section of the release notes and there were only 52.

    Also, the length of the list of known issues in what Unity says is supposedly an RC3 gets a big yikes from me. Over 30 items in that list.

    3jfhq2.jpg
     
  17. Ramobo

    Ramobo

    Joined:
    Dec 26, 2018
    Posts:
    212
    It really says something when they even considered releasing 2019.3 with that list of known issues. They should have just stayed in beta. It's worse than 2019.2, except 2019.2's domain reload freeze is, according to my observations, ~3x that of 2019.1 and 2019.3 fixes that. Keep in mind that 2019.2 released with 18 known issues and still has 9 as of patch 16.

    I stayed in 2019.1 and am only using 2019.3 regularly now because of the skip domain reload option, new UI and because I only found one major bug that affects me: The editor randomly "soft-locking" — refusing to switch tabs and some other stuff —, fixed by restarting.
    EDIT: I just randomly came across this bug in 2019.1.14f1. I fixed it by selecting something in the hierarchy.
     
    Last edited: Dec 16, 2019
  18. Peter77

    Peter77

    QA Jesus

    Joined:
    Jun 12, 2013
    Posts:
    6,313
    @LeonhardP I always wondered what makes some issues end up in the "Known Issues" section of the release notes?

    I'm asking, because there exist many more known issues, basically the entire public issue tracker. But some of these issues end up in the "Known Issues" section. At first I thought these are issues introduced in that specific beta version, but that's not the case.
     
  19. Milo_del_mal

    Milo_del_mal

    Joined:
    Jan 27, 2013
    Posts:
    43
    They know what is happening, either it was reported or they found it... They just have not solved it.
    And I assume they are out in Know issues so you know that a system is available, but there is an existing issue so you can plan around preventing getting stuck with that issue.
     
  20. OLGV

    OLGV

    Joined:
    Oct 4, 2012
    Posts:
    55
    Please don't call this a "...3.0f#" release.
    Leave it "Beta", at least you admit the constant unfinished state of everything.
     
  21. Ramobo

    Ramobo

    Joined:
    Dec 26, 2018
    Posts:
    212
    Yeah... no. If that was the case, there would be hundreds of known issues in every single release.
     
  22. Lars-Steenhoff

    Lars-Steenhoff

    Joined:
    Aug 7, 2007
    Posts:
    3,259
    What I understood is that the known issues are the ones that are planned to be fixed in this release cycle.

    The other issues in the tracker may be fixed this cycle, or the next or never.
     
    Peter77 likes this.
  23. DoctorShinobi

    DoctorShinobi

    Joined:
    Oct 5, 2012
    Posts:
    212
    If I had to guess I'd say it's critical issues that can affect a lot of users. The role of the Known Issues section is probably to let users know if their project could break in the new version. It's definitely nice to have and it makes sense they wouldn't put all hundred known issues in that section.
     
    Peter77 and Ramobo like this.
  24. Marcos-Elias

    Marcos-Elias

    Joined:
    Nov 1, 2014
    Posts:
    136
    I think that Unity should stop doing three versions per year and make a rolling release, separating the most stable stuff for an annual LTS version by the end of each year. (just a suggestion, of course)

    The deadline of three versions per year in a complex project like a game engine makes it very hard to become stable.

    Lots of bugs, regressions, unexpected behavior. Everyone knows that it is for the evolution, but this is really bad for production environments. Tech releases are not stable, almost never were. Almost every update to a dot 1, 2 or 3 ending version got some hard problems to lots of projects or the editor itself. This kind of thing is not expected in a "final" release.

    Or maybe just provide a better communication by placing the beta label forever on tech releases, leaving the LTS as the only stable one.

    Even LTS it is not that stable, we have critical problems with il2cpp and the 64-bit version for Android with performance degradation on many devices, that are still not fixed... Since Google forced updating the app in 64 bit, performance is very bad in some devices and it seems that there's no solution for that yet. This hurts. One seems related to Debug.Log being a lot slower and other is about physics, for what I saw on another topic. It's something like this, but I don't know if it is the same problem: https://issuetracker.unity3d.com/is...-with-physx-on-huawei-mate-20-and-mate-20-pro
    Or this one: https://issuetracker.unity3d.com/is...the-project-is-built-with-64-bit-architecture
    That's an unacceptable thing to be in a "LTS" version :(
     
    Last edited: Dec 14, 2019
  25. laurentlavigne

    laurentlavigne

    Joined:
    Aug 16, 2012
    Posts:
    5,747
    can you elaborate?
     
  26. Velo222

    Velo222

    Joined:
    Apr 29, 2012
    Posts:
    1,436
    So this means there will technically have been 2 dot releases this year. :rolleyes:
     
    MadeFromPolygons likes this.
  27. wlad_s

    wlad_s

    Joined:
    Feb 18, 2011
    Posts:
    144
    I agree with Marcos-Elias, one stable release per year would be just fine. Having three unstable releases complicates things for Unity themselves, for us users and for asset developers. I can't even imagine how frustrating it must be for asset developers, having to constantly update their assets and deal with Unity bugs.

    By having only one official stable release per year, there would be less worry for everyone. They could continue having lots of beta releases during the year so that people who are interested can test out the new features, while serious developers and asset developers wouldn't have to waste their time with that.

    Also, bugfixing should have more priority than new features. When you start developing a game, you construct your vision and budget by what's available now, not by what features might or might not be there in three years time.
     
  28. Marcos-Elias

    Marcos-Elias

    Joined:
    Nov 1, 2014
    Posts:
    136
    The two links at the end of my post. That problem does not appear if we build as 32 bit with Mono, without changing a single line of code. For what I saw on some devices, at least...
    If the problem is not with il2cpp itself, my bad, but it's still strange since it works with Mono. Also in many release notes there are a bunch of bug fixes related to il2cpp, so it means that it was not "completely stable" before for certain situations...
     
  29. laurentlavigne

    laurentlavigne

    Joined:
    Aug 16, 2012
    Posts:
    5,747
    ok so IL2CPP is like Flash back in the days, it works but you gotta watch out every step of your code with it.
     
    phobos2077 and MadeFromPolygons like this.
  30. LeonhardP

    LeonhardP

    Unity Technologies

    Joined:
    Jul 4, 2016
    Posts:
    3,021
    It's pretty much exactly what @DoctorShinobi wrote.

    The Debug.Log issue is most likely this one (link will start working in a couple of hours): https://issuetracker.unity3d.com/product/unity/issues/guid/1161589/

    It's being investigated by the devs.

    The physics performance regression has been fixed and the fix should get backported to 2019.3 and the LTS releases soon.
     
    Last edited: Dec 16, 2019
    LostPanda, Peter77 and Marcos-Elias like this.
  31. LeonhardP

    LeonhardP

    Unity Technologies

    Joined:
    Jul 4, 2016
    Posts:
    3,021
    That's indeed a faulty bit of information. It's supposed to read "170 features and improvements", combining the release note entries that fall under the "Features" and "Changes" categories. The blog post has been updated.
     
    Deozaan and interpol_kun like this.
  32. nsxdavid

    nsxdavid

    Joined:
    Apr 6, 2009
    Posts:
    475
    I think we will all come to regard the 2019's as the painful transition years. It remains to be seen if 2020 is an extension of this. In generally, the direction is solid. The path, however, is rocky and treacherous. These growing pains are necessary... and yeah, maybe there was a way to do it smoother, but let's not forget the tremendous efforts being put into making this all happen.
     
  33. taylank

    taylank

    Joined:
    Nov 3, 2012
    Posts:
    182
    I don't think anybody's forgetting that. I mean, I'm paying Unity my pro fees as a constant reminder of the value of the engine. I'm not expecting the developers to be superhuman; of course there are going to be bugs no matter how brilliant the engineers are. And it's not even their fault. It's really a business/marketing responsibility to correctly advertise the capabilities of one's product and here Unity has been taking liberties with their marketing at my expense. That's the issue for me.
     
    midian808, transat, wlad_s and 2 others like this.
  34. nsxdavid

    nsxdavid

    Joined:
    Apr 6, 2009
    Posts:
    475
    Yeah, engineers tend to be particular about these things. Marketing folks, however, see reality as more... malleable. :)
     
    wlad_s and Deleted User like this.
  35. Suduckgames

    Suduckgames

    Joined:
    Nov 28, 2016
    Posts:
    216
    I really appreciate your work guys, the new features are awesome and the direction of the engine is perfect.

    However, the problem came with the "production ready" tag, I don't mind to wait more time to use a stable version of Unity, the problem comes when I select a version that is production-ready and I found lots of bugs and workarounds that need to be addressed.

    When I select a production ready version I expect to have almost no bugs, with the main workflow of features working fine (looking specially at you nested prefab)

    Please, the last two years have been a pain in the ass to select the version Unity because we can't trust the "Produciton ready/ Official release" tag.

    We just need to be more clear about what is production ready and which version has lots of bugs so we can choose accordily
     
  36. pbritton

    pbritton

    Joined:
    Nov 14, 2016
    Posts:
    117
    This version of Unity was the first time I actually pondered using Unreal. The experience with 2019 has been particularly frustrating. I hope that these issues are resolved.
     
    stonstad, midian808, PeterB and 3 others like this.
  37. chrisk

    chrisk

    Joined:
    Jan 23, 2009
    Posts:
    700
    It's not about how many releases we have each year. Look how Epic is doing. It's just that Unity is Incompetent and Lazy. And on top of that Unity is Stubborn, not listening to users, yet they always say they are listening. That's what makes me so frustrating.

    If I compare 2018.3 Editor to 2019.3, there isn't anything significant enough for me to migrate to 2019.3, except one feature, "Fast Enter-to-Play" This feature is under appreciated and not even highlighted in the United and yet I think it's the only worthwhile feature I'll consider significant for the past few years. Fast Enter-to-Play mode should've been the default mode from the beginning but we finally have it after ~15 years later.

    The rest are available through separate packages and they are either Preview or Beta. I wouldn't dare touch Preview or Beta product from Unity. I don't even trust the so called "Production Ready" unless it's V2 or above.

    Unity is developed by people who never release any major games and they don't really understand what developers want or need. When asked why Unity is not making a game ?, the answer I got is "We don't want to compete with our customers" What a lame excuse! Epic is doing it and no one blames them. Basically, Unity doesn't benefit from developers being successful, they are just interested having many newbies and selling Assets, and Event tickets.
    Unity Editor basically an empty shell and users are hopeless without help from the 3rd party developers.

    Year 2019 has been a really frustrating year. Thanks to ever getting slower Editor with each release and DOTS fiasco that no one is really using or no one can use. I think DOTS will be irrelevant to the most for at least for a year or two.

    What Unity should've done is to create separate DOTS branch and work there. And optimize/stabilize the current Editor. Editor is so slow and clunky and I really want to kill myself everyday.

    Speaking of optimizing the current Editor, using Mono based Editor has no future. We should migrate to .NetCore immediately. .NetCore performance is as fast as Burst compiled without doing anything special. It's just faster and we will be free from AppDomain reload hell. And it will help Unity to build new features on top of it. Adding more feature on top of Mono will slow things down even further and I hope Unity doesn't make any worse that already is.

    Yeah, it's easier said than done but if Unity really wants to do it and spent 1/10th of the time they spent on DOTS, we should have .NetCore Editor by now. And it helps everyone immediately instead of selective few who adopts DOTS.

    Well.. I already mentioned that Unity is Lazy, Stubborn and Incompetent and I expect the least.

    If you want to prove that I'm wrong, please do this.

    1. Create a separate DOTS branch and don't get that interfere with the current Editor. You can call it Unity V2 or whatever. I don't care until you can release it.
    2. Migrate the Editor to .NetCore as if there is nothing more important. You will agree if you have to work on large project with the clunky editor everyday.
    3. Overhaul Editor Workflow. The Editor workflow is really horrible and it's still missing many simple stuff yet fundamental to any Editors. If you don't know what they are, you shouldn't be reading this. It's probably better to start from scratch as anyone can design better than the current Editor. Focus on making things Easy, not on Simple. They are fundamentally different.
    4. Start AAA commercial game project and deal with the pain first, instead of making all of us beta testers for your buggy product.

    Lastly, if you can't do any of above, just make the Editor stable and fast so that we can work on our games, not fighting with the Editor.

    Thanks.
     
    Last edited: Dec 18, 2019
    serpin, PeterB, Doddler and 18 others like this.
  38. SmartMediaNL

    SmartMediaNL

    Joined:
    Sep 29, 2016
    Posts:
    77
    I remember reading Unity would take better efforts regarding bug smashing 3 years ago. Now they have added a bunch of automated unit testing and so their has been much approved on that side i suppose. But i hope Unity will add a "sabbatical" year where no new features would be added but all focus would be on fixing bugs and smashing even more bugs and for fun smashing some more. for starters fix all known bugs before even thinking of moving on.
     
    midian808, PeterB, MegamaDev and 3 others like this.
  39. doarp

    doarp

    Joined:
    Sep 24, 2019
    Posts:
    144
    Yeah... that’s not gonna happen. They have so much in their plate with new developments in all directions that what you suggest isn’t an option.
    But yeah, a day barely goes by without unity crashing one way or another or some weird bug popping up (to be fair I’m on 2019.2 / 2019.3, but still those are too many too often).
     
    MadeFromPolygons likes this.
  40. DoctorShinobi

    DoctorShinobi

    Joined:
    Oct 5, 2012
    Posts:
    212
    Chrisk, there's a lot of reason to criticize Unity and there have been some valid points raised in this thread. Your criticism is the absolute worst one here.

    Instead of giving constructive feedback you call Unity incompetent multiple times, and you phrase controversial ideas as if you're giving orders.
    One of your points complains about the editor workflow. You say that it's missing many simple stuff but then you go and say "If you don't know what they are, you shouldn't be reading this". What kind of S***ty feedback is that lol? If you care about Unity improving the editor then you should tell them exactly what is it about the workflow is bad. I'd probably write something like "the editor hangs on assembly reload" or "dragging assets around different folders can take a lot of time in a large project".

    It's as if it is more important to you to insult Unity instead of giving constructive feedback
     
    Last edited: Dec 18, 2019
  41. patrykszylindev

    patrykszylindev

    Joined:
    Oct 28, 2019
    Posts:
    23
    The workflow within the engine has improved incredibly since the last time I touched unity (over a year ago now) in terms of features. I was excited to read all the stuff about DOTS and mobile development features (finally glad to see the Device Simulator and impressed by the new Input System) and I am now planning to fully commit to the Unity3D engine, but as other folks mentioned, the frustration with the engine still persists due to the bugs previously mentioned. Overall, a really good job by Unity Technologies, although most of the stuff I am excited to use isn't close to frustration-free development.

    I've recently tried to use the new Input System together with Device Simulator, and I struggled to simulate "Touch" events on PC and Android platforms when interacting with UI elements on Device Simulator Window, I have ticked the "Simulate mouse clicks as touch events" or something along those lines in the Device Simulator window. I have also updated the "Event System" GameObject that gets created within the hierarchy to listen for events coming from the new input system by selecting the Input System auto-generated C# class... I didn't get anywhere, ended up downgrading my unity3d version and stuck to the old way of doing things.

    Regardless, good attempt from Unity but not quite there yet.

    EDIT
    I forgot to mention that Unity Remote 5 is definitely bugged out as well. Not sure if I'm the only one who struggles with Unity Remote on 2019.3, but it definitely does not like my Samsung S9+ device, sometimes it connects to the editor's game window but most of the time it doesn't
     
    Last edited: Dec 18, 2019
    tonialatalo and Abrasive like this.
  42. RomanBanana

    RomanBanana

    Unity Technologies

    Joined:
    Jun 23, 2015
    Posts:
    51
    Thank you for your feedback. We haven't yet implemented support for the new input system and haven't tested that scenario. We plan to work on new input system support in Device Simulator early next year.
     
  43. interpol_kun

    interpol_kun

    Joined:
    Jul 28, 2016
    Posts:
    134
    Yeah, but I think he is very, very frustrated, like almost desperate. He has some valid points in his post too, but it is really important to elaborate on them. I will do it for him (and cry for a bit too):

    This is true. .NetCore 3.1 and C# 8 have tons of new features which will help with the performance (Span<T>, intrinsics, under-the-hood optimizations, AOT, improved GC, Collectible Assemblies to remove DomainReloadTimes without clunky workarounds like Fast-enter-play-mode) and the stability (Nullable Reference Types). AFAIK some Unity guy tried to port the Editor back in 2018 or even 2017 for fun, so with a medium-size team, it won't be really hard.

    .Net Core should be the future of Unity, not Mono. The more you ignore it now, the more it will cost you and us, devs, in the future.

    That is true. When I started using Unity I was already experimenting with UE4. I chose Unity for C# and simplicity, but I was always wondering, why is the Game Engine lacks top-hierarchy tabs? Why can't I pin the animator windows on the separate first-class tab, so I can easily change my workflow environment? That's one of the major UX flaws for me. But I saw something from the Unity Hack Week (I think from the Will) which could be implemented in the 2020 cycle. Better than never, but it says a lot about overall Unity's understanding of the good workflow.

    The next personal pain point is the lack of dedicated editors for different types of stuff. It's good that nested prefabs came into life (and it's again the feature which should be in the engine from the start, it's very perplexing that you ignored it for the decade), but the runtime GUI workflow is a total mess. Soon the UIBuilder will come, that's good stuff, I hope. But again, Unity just has had to work with their own GUI system for a few months on a medium-size project with lots of complex widgets to truly understand the total horribleness of a developer's daily routine.

    That's true. And when on the Unite 2019 the CEO of Unity (former EA btw) said "We're not here to make games, we're not interested in creating experiences that compete with your hobbies or your businesses." I cringed like really hard. If the man himself can't understand the benefit of making real games with your own technology, that says a lot about the overall direction of the project.

    He tried to make it sound beneficial and reasonable for developers like Unity won't take our slice of the players. But in reality, Unity will take our time and money by creating an unstable product with lack of QoL stuff which you can only notice by using unity "in the field".

    One year ago I saw the post from Joachim Ante about dogfooding. He replied that they in Unity do make games and use it for themselves. But let me get it straight: making demo projects, cinematics and research projects in the spare time is not dogfooding.

    Dogfooding in gamedev means you have a real product based on your own technology, which will make you money, and satisfy the customer. It won't be abandoned for a year like FPS Sample (which, as I understand that right, is being updated only to LTS versions), or like other projects, it won't be a demo, it won't be a tech showcase with lots of workarounds like MegaCity. It will be supported with patches or, if it's a game-service, with constant updates. You don't have any big projects from your team. You don't support them properly. It's just a pseudo-dogfooding.

    Your CEO said it for you. You are here not to make games. I think this is a real problem.

    I will be sticking with Unity for another year, I won't blame it on "the transition phase". You were always like that, from the very time I started using Unity you had lot's of abandoned tech, unstable releases (which by nowadays Unity's quality standards will be mega-stable), lack of understanding from devs and overall "if I ignore it maybe it will go away" attitude. It shocked me when I realised that the community was asking for prefab workflow and terrain updates from 2009. They were among the most voted topics in the feedback section which you moved to a forum for no reason.

    But I wish you good luck with your huge plans. As again, your CEO believes that "traditional businesses— such as construction, car design, and film making — will eventually replace video games as Unity's main customer base and source of revenue". It makes it more clear to me. It's really sad that the only people in Unity who's gonna read that are the one who can't do anything about the management. You are developers, I believe you too, at some point, dissatisfied with all of that stuff going on. And this is why we shouldn't be very rude and aggressive towards you like Chrisk. Those posts should be read by your CEO and other people in charge, but only feedback they are interested in is numbers.

    Have a good one, my fight against the windmills will last till the end of 2020. But I am already feeling stupid.
     
    Last edited: Dec 19, 2019
  44. Ramobo

    Ramobo

    Joined:
    Dec 26, 2018
    Posts:
    212
    Tell me about it. 2019 was disappointing to say the least. What I... uh... noticed? In 2019:
    2019.1: Clickable stack traces, inspector prefab editing, incremental GC, DOTS hype, UI Elements for the Editor, bye bye built-in launcher, Hub 2.0. Currently using in a team project.
    2019.2: Launching with 18 known issues and still having 9 as of patch 16, insane compilation + domain reload times (~3x 2019.1), TypeCache (slightly shortened some code in one occasion in 2019.3). The first occasion of me skipping a Unity version, and I've been using it for three years.
    2019.3: New UI, disabling domain reload, considering releasing it with 42 known issues (rc1). Currently using in personal projects, as I'm yet to run into major bugs in this version and the benefits (the thing .NET Core can replace) are just too great to wait.

    Now I'd like to highlight two points of that:
    • UI Elements: Editor only (not that it matters to me since I'm not using it) and they chose XML + CSS (dialects thereof) instead of the unified and arguably better XAML. I guess they want to apologize to the angry web developers after they finally threw UnityScript in a volcano. Additionally, how the hell are we supposed to work with this thing at runtime when everything else is game objects? Yes, they'll be converted to game objects at runtime, according to a developer, but it doesn't sound like we'll have any control over that representation.
    • DOTS: Oh boy. Oh boy. It's barely usable. They threw the DOTS Editor out the window when they had already done it in Project Tiny and then pushed their hot garbage "conversion workflow" that doesn't even have a two-way link between the inspector and the entity world, so you have to rely on the Entity Debugger. Yes, you have to rely on a debugging tool for basic usage. And their roadmap says "Entities 1.0 in 2020.1". Like hell.

    If you can stay calm about what has turned into a huge pile of hot garbage of an engine, good for you. Not everyone is like you, though. Some are long-time users of Unity who can't take it anymore, some have used the engine for a while and are realizing what they've signed up for, and some are beginners who don't realize what waits for them below the tip of the iceberg.
    Remember, Unity is the easiest engine around: You can run, but you can't hide unless you can handle Python (indentation-based, which can only go very well, I imagine) or C++ (low-level). You might disagree with me, in which case I point you to the asset flips, most of which are made with Unity, giving the engine a bad reputation (not that it's undeserved).
    Others engines support C#? As far as I know, the only ones with decent C# support are Godot (not the main language) and Xenko (extremely lacking and bad Editor performance that manages to rival Unity, written in C# though). Both are also Mono, as far as I'm aware, and Godot just has a Mono version, as C# is not its main language. Additionally, neither have a play mode, meaning that the application opens as another program at the beginning. You thought working with preload scenes in Unity was bad? Wait until you try Xenko or Godot.
    And a like-minded person:
    Yet here we are, with them only contemplating the idea of throwing Mono in a volcano. Pause everything and migrate to .NET Core, I say. Also:
    That was @xoofx in the May 2017 Hackweek. Here you go. Yes, 2.5 years ago and only now are they considering it.

    Same here. As I said:
    See my points on UI Elements at the long-gone beginning of the post. This turned out much longer than I expected. Originally it was a reply to just @chrisk's message, then @DoctorShinobi and you popped in, not that it's a bad thing.

    [Heavy breathing...]

    Let's see... Paid dark theme, lots of engine features (like render textures) locked behind a paywall until a few years ago (I think the Unity 5 pricing scheme change) despite the Personal tier being advertised with "All Engine Features" until sometime ago...

    Oh, there's also their inconsistent API. Why do you need three different conventions in one engine? And even their "new conventions namespace" (`Unity.*` except `Unity.Mathematics`, can't find source) is inconsistent.
    Another one: We have `UnityEngine` and `UnityEditor`. Why not `Unity.*` and `Unity.Editor`?
     
    Last edited: Dec 18, 2019
  45. phobos2077

    phobos2077

    Joined:
    Feb 10, 2018
    Posts:
    349
    I just can't agree with this enough. Full C# support is the main selling point of Unity right now (before it was graphic assets on Asset Store, which I believe is dying now because everything is suddenly not supported with new rendering pipelines). Yet, UnitTech's continue to ignore the needs of coders. Yes, they add some nice things here and there, but C# scripting improvements should be their main focus, as this engine is chosen instead of it's competitors mainly by coders for coding reasons.

    I should say DOTS is overall a nice option to have. We use it in a commercial project and it does solve some problems and gives great potential for simulation time optimization. I just don't like the fact how they neglect the existing technologies (GameObjects) that people really use in production on-mass.

    Also I think it would be great if Unity management would distribute more resources to the "core" team or "Unity Editor team" whatever they call it, so they really focus on the overall usability of the Editor itself. Unity game engine is good, but I gotta say the Editor leaves a lot to be desired in terms of UX.

    Please do less hype for individual features (which seems to be eternally in preview, even when they call them "production ready") and more on improving Unity as a software product.
     
    p87, tonialatalo, PeterB and 7 others like this.
  46. Ramobo

    Ramobo

    Joined:
    Dec 26, 2018
    Posts:
    212
    Yup. I don't see a lot of good programmer tooling. Visual Scripting? For DOTS? Who thought this was a good idea?

    It's sad. It looks like they just abandoned MonoBehaviour's only hope.

    Ahem.
    • Addressables: No synchronous loading API yet.
    • Collaborate: Link loss issues last time I tried it.
    • Input System: Questionable design choices, actions that don't work like you'd logically expect them to.
    • NavMesh Components: Abandoned, apparently. Not even in the package manager yet. Oh:
    • Package Manager: Lacking features such as Git dependencies and NuGet support.
    • ProGrids: Stuck in official preview for as long as I can remember.
     
    Awarisu and elias_t like this.
  47. snacktime

    snacktime

    Joined:
    Apr 15, 2013
    Posts:
    3,345
    2019.3 is the biggest release I've ever seen, they are trying to cram a ton into it. But equating what is important to you to what is important to someone else, that's where it goes off the rails I think. A whole lot of what was recently mentioned here I could almost care less about. DOTS is what we care about the most, and 2019.3 has a lot of DOTS and performance related stuff we care about a lot.

    For someone that doesn't care about DOTS, I can see how 2019.3 doesn't look too appealing.

    What bothers me the most given our specific focus is the whole performance first mantra appears to be more performance eventually. With getting what they want in being a priority. Asset db v2 being a good example. It's caused so many performance issues either directly or indirectly it's a bit ridiculous.

    Another issue is previews, specifically DOTS previews. DOTS core is pretty stable and they have been pretty clear about people are using this on new games bound for production and they think it's ready for that. Yet they also hide behind the preview mantra when they really want to put something in that's not really ready. And it's all tied at the hip with 2019.3.

    Performance first is pretty clear, you can't have it both ways.

    I'm reserving final judgement until they actually release. I hope they take as long as it needs to take. That we are into RC's makes me nervous.
     
  48. Lars-Steenhoff

    Lars-Steenhoff

    Joined:
    Aug 7, 2007
    Posts:
    3,259
    It's normal to be in release candidate in December 2019 if you want to release a version with this year's number on it, I would even say its a bit late.

    I mean 2019 LTS version is going to be out halfway 2020. This makes no sense to me.
     
    Awarisu, Ramobo and Abrasive like this.
  49. laurentlavigne

    laurentlavigne

    Joined:
    Aug 16, 2012
    Posts:
    5,747
    There is a lot of frustration on this thread, way more than back in the days of 4.7
    Back in the days, Unity was at least feature complete and the vital parts were stable or those bugs were founds ways around and I think it's due to the monolithic and slow release schedule.

    How has Unity agile-machinegun release changed the way you work?
    And who among you have released a game with 2018 or dare I say 2019 and how was the experience?

    I'm asking because I started a thing in 2019 and so far I've hit no problem apart from the usual Mecanim bug or sluggish editor.
     
  50. Roni92pl

    Roni92pl

    Joined:
    Jun 2, 2015
    Posts:
    395
    Im currently finishing/releasing game in 2019.2 and I must say it's not bad, and Im in good spirit generally for it. I mean yeah, domain reload times are real PITA, but overall I can't complain much about stability..
    however there was some VERY strange episode where Unity started to crash like mad in basically random gameplay moments; after few days of debugging in desperation I reimported whole project, resterted Unity, restarted PC.... and guess what? ALL crashes stopped again, literally had not one crash after that, lol.

    I also switched to IL2CPP builds from mono builds recently and to my surprise it was flawless(I was basically expecting this transition to fail miserably considering my earlier exp lol), especially considering I didn't prepared my project in any special way earlier, so it was only matter of changing and preparing serialization solution to be AOT.