Search Unity

  1. Good news ✨ We have more Unite Now videos available for you to watch on-demand! Come check them out and ask our experts any questions!
    Dismiss Notice

Upcoming "Meet the Devs: Internal Production Team" -- Post Your Questions

Discussion in 'Unite Now' started by Mike-Geig, Apr 17, 2020.

  1. Mike-Geig

    Mike-Geig

    Unity Technologies

    Joined:
    Aug 16, 2013
    Posts:
    178
    Howdy!

    Next Wednesday, April 22, I’m hosting the next episode of our Unite Now Meet the Devs series, talking to our internal production team. You might know them as the authors of projects like FPS Sample, Megacity or DOTS Sample. We’ll chat about the impact of their work on how we develop Unity, what they’re working on next, and take questions from the audience. This session is live, and you can register to join us here.

    This is where I need your help! If you have questions that you’d like us to focus on during the session, please free to post them in this thread. Let's talk about the things that you want to talk about!

    More info about the session here.
     
  2. Kolyasisan

    Kolyasisan

    Joined:
    Feb 2, 2015
    Posts:
    321
    I'm very positive that the current Entity Debugger is not comfortable to use at all. Do you consider it to be so and will we get some kind of an overhaul to it or the DOTS ecosystem to make it easier to work with entities in the editor?
     
  3. Fluffy-Tails

    Fluffy-Tails

    Joined:
    Jun 28, 2009
    Posts:
    120
    Are there any plans for Real Time Hair and Fur or Procedural methods?
    Been waiting ages for this to happen and Sherman is the only thing I have seen so far.
     
    hayetham likes this.
  4. joshcamas

    joshcamas

    Joined:
    Jun 16, 2017
    Posts:
    1,061
    1. With open world features being more of a focus in the future, are there any plans to assess the current implementation of the occlusion culler?

    2. I think the idea of building larger "gamey" demos is a good idea, since it allows for an in depth understanding of what Unity does well and what it lacks. However, I noticed it has taken a very long time for features in the mega city demo (such as HLOD) to be implemented into the engine, sort of weakening the whole point of it being a demo in the first place. Are there plans on reducing this level of "custom code" later demos requires to make it be able to run? Aka will newer demos be more like "this is what Unity can do", versus "this is what Unity will be able to do in a few years"?
     
  5. Tomasz_Pasterski

    Tomasz_Pasterski

    Joined:
    Aug 12, 2014
    Posts:
    64
    My question is maybe not exactly to ppls you will talk with but maybe you can gather and present some information about........TREMBLE:
    Navmesh, Navmesh agents, agents local avoidance improvements, i heard there is new team working on Navmesh stuff improvements, curious to know what they have in mind for future improvements of this systems.
     
  6. JamesArndt

    JamesArndt

    Joined:
    Dec 1, 2009
    Posts:
    2,830
    Are there any plans for a mobile optimized lightmapping solution for 3d games? One that would allow us to pass through lightmap UVs completely untouched, implement externally authored lightmaps into scenes.
     
    joshcamas likes this.
  7. Marcos-Elias

    Marcos-Elias

    Joined:
    Nov 1, 2014
    Posts:
    115
    I would like to know more about open world plans. Unity never had a support for open world before, many of its native features require the static flag with fixed positions in world space... But these positions are so limited due to the usage of 32 bit float for storing transforms data...

    In recent moves it seems that this is about to change with DOTS. But it is still too far from the reality, by being in early stages...

    What are the real plans for open world?
    Will it be possible creating trully large scenes, without hiccups when fixing positions of all loaded objects? For example, to create a vehicle simulation system for a long road with more than 10, 20, 50km (be it with streaming or not)...

    Will PhysX be abandoned? No chance of supporting its great shiftOrigin function that already works flawlessly in Unreal? So we could improve the floating origin script to move all rigidbodies/colliders at once... This is something required after some km driven to fix floating point precision issues... There are many posts on forums detailing and asking for this feature. All always were ignored by Unity devs and workers :( The move to DOTS is certainly amazing, it's almost like a new engine, so powerful! But at the same time it requires that most games must be recreated from scratch... In meanwhile it would be great having a temporary open world support. If we could have access to shiftOrigin function that unfortunately is not exposed...
     
    Last edited: Apr 19, 2020
  8. optimise

    optimise

    Joined:
    Jan 22, 2014
    Posts:
    598
    Any plan to develop realtime multiplayer mobile DOTS Sample project? If I rememeber correctly there is plan to develop asynchronous multiplayer mobile DOTS Sample project but I still hope Unity can change the plan to the realtime multiplayer and I believe the future multiplayer game should be fully realtime.
     
  9. Mikael-H

    Mikael-H

    Joined:
    Apr 26, 2013
    Posts:
    303
    What's your approach to clean code? Do you have a coding style guide? Do you run static code analysis? How important is clean code in internal and externally available source code?
     
  10. Baste

    Baste

    Joined:
    Jan 24, 2013
    Posts:
    5,013
    Any chance you'll do more to keep demo projects up to date?

    It's very nice to have demo projects to show off parts of the engine, but they generally get made for the version that's bleeding edge when the demo's made, and when the demo's done, it's stuck.

    The DOTS samples seem to be continually updated, while the only version of MegaCity I can find is for a beta version of 2019.1. Older demos (the tank one, the very old shoot-spider-robot thing) also generally got left behind.

    A big reason for doing internal projects is dogfooding. A thing that can be a big struggle for your users is upgrading our projects to a new version - it seems to especially be an issue for people using DOTS or the SRPs. It's be great to see you spend some time to keep your demos working on the latest Unity versions, so you get to feel the pain, and we get to have demos that we can actually run without having to dig through the download archive for beta versions of the engine.
     
    Cicaeda, SugoiDev, Egad_McDad and 3 others like this.
  11. optimise

    optimise

    Joined:
    Jan 22, 2014
    Posts:
    598
    Any plan to upgrade Megacity project to latest Unity 2020 cycle version and make it support Hybrid Renderer V2?
     
  12. andyz

    andyz

    Joined:
    Jan 5, 2010
    Posts:
    1,406
    Why do you use outdated C# naming like m_, k_ variable prefixes etc. in Learn code and is there a common C# naming convention internally?

    FPS, racing, (flying?) etc template/learn projects are great though and I hope you keep these up to date, though keeping an LTS version of each would be sensible (no 2018 LTS)
     
  13. transat

    transat

    Joined:
    May 5, 2018
    Posts:
    772
    Are there any plans for a big multiplayer “open world” sample?
     
    MidniteOil, awesomedata and Mikael-H like this.
  14. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    +10000

    I know that Internal Production Team goal is to test and push new tech to the limits but checking how new features integrate and how stable they are with a project of Megacity scale is equally important.
     
  15. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    Mikael-H likes this.
  16. axeldaguerre

    axeldaguerre

    Joined:
    Mar 1, 2020
    Posts:
    3
    Does a plugin for XR games will be written or we will still have to use the XR plugin in place ?
     
  17. valarus

    valarus

    Joined:
    Apr 7, 2019
    Posts:
    203
    Could you update Book of the dead to 19.3 SRP?
     
  18. forestrf

    forestrf

    Joined:
    Aug 28, 2010
    Posts:
    99
    Could you explain more about the Playables API? It's not well documented, will we get a better documentation for it? It's pretty difficult to use as it is, even though it is used in both the FPS multiplayer project and the Megacity project.
    How should it be used? How do we avoid performance spikes when using it? What are the best practices for using it? How does it work internally?
     
  19. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    Are there any news on occlusion culling in regards to your newest large open world shooter project? Is it all baked meaning that occluders still can't be changed/added/removed in runtime? I find lack of dynamic occlusion culling a pretty big issue in any sort of procedural or interactive environment 3d games (for ex. fortnite, minecraft, battlefield).
     
    Last edited: Apr 22, 2020
    awesomedata and andyz like this.
  20. calemb

    calemb

    Joined:
    Dec 13, 2015
    Posts:
    1
    Maybe it's no directly about tech, but still for Dev Team. Do you hire engine devs only as remote workers?
     
  21. aut0mata

    aut0mata

    Joined:
    Aug 22, 2014
    Posts:
    4
    Is there any ETA for realtime shadows from Point lights on URP? What about post-processing Ambient Occlusion on URP?

    Thanks!
     
    Last edited: Apr 22, 2020
  22. Livealot

    Livealot

    Joined:
    Sep 2, 2013
    Posts:
    153
    Cool session. Would have liked it even better if there were some screenshare segments where the team showed us what they were currently working on, or tell stories by showing snippets of released projects.
     
    AskCarol likes this.
  23. PhilSA

    PhilSA

    Joined:
    Jul 11, 2013
    Posts:
    1,474
    A criticism I would have of the FPS sample was that it felt somewhat inaccessible due to the very large codebase + complicated setup required to make a project like this. In my opinion it was perhaps a bit too complex of a project to do a good job at being a sample that people can just easily dive into & learn from.

    Would you expect the next DOTS open world sample to improve on this front? (be a bit more simple, accessible, straightforward, down-to-essentials, etc...)? I feel like it could be worth it even if the end result sacrifices a bit of wow factor, but I realize there are pros & cons
     
    Last edited: Apr 22, 2020
  24. petera_unity

    petera_unity

    Unity Technologies

    Joined:
    May 23, 2016
    Posts:
    41
    I think this is a fair point. But I also think the scope is needed to at least to some degree validate the scalability of the solutions. In fact, one potential alternative criticism of the projects could be: they are too small. What we want to do going forward is to have a very small project kind of like Dots Sample that contains just the core aspects. And then a larger project where those core elements are used in a larger scale. It is quite important, though, that the small sample still has the robustness that comes from code that has been used in a larger setting. So it will require some sync. But I believe that is the only way to do this right.
     
  25. PhilSA

    PhilSA

    Joined:
    Jul 11, 2013
    Posts:
    1,474
    Seems like an ideal approach to me as well! Thanks
     
  26. valarus

    valarus

    Joined:
    Apr 7, 2019
    Posts:
    203
    Could you create for SRP (URP/HDRP) non-dots/dots an FPS template character ( or some other other templates) which can be used in existing projects and scale along all SRP versions?
    FPS sample is based on only one version of Unity/HDRP.
     
  27. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    My question about making Internal Production Team bigger was badly interpreted. The team lacks resources to keep projects updated, documented and making them easier to understand. Even multiple projects could be developed at the same time. Don't get me wrong I think this team very important and does wonderful job but it's not enough.

    Compare amount of senior programmers that make games and have access to source code of Unity and Unreal. Not only have access but also contribute back with their changes. My guess: several dozen vs several hundred.

    Fortunately it changed slightly with packages and their C# but still most of them are not on github and have dependencies on hidden C++.


    I wish your projects would be released as very small and simple games to generate some resources to keep projects updated. Imagine if in the future there will be 10-20 projects - no way of keeping them updated if they are just proofs of concepts.

    Not only that but it would be great marketing tool - imagine big content creator playing Megacity game.
     
    Last edited: Apr 23, 2020
  28. petera_unity

    petera_unity

    Unity Technologies

    Joined:
    May 23, 2016
    Posts:
    41
    Yeah. I actually had that thought too while we were given the answer. The answer we gave went in another (important) direction of the importance of us "staying small" on the individual productions because we strongly believe Unity should also be a tool that allows small(ish) teams to punch above their weight. Solving the challenge of building a massive world by adding a massive amount of staff to production can lead to unfortunate shortcuts in the tooling, for example!

    But -- not surprisingly -- I agree with your point as well: we should and could do much more in this space of demos and samples. I think I mentioned in the interview that we have recently gotten a new structure that will help with processes and resources to manage lifecycle of the projects better. And as we also mentioned we are building up additional teams to make content for other areas like for example mobile.

    Growing takes time, of course, but fwiw your wishlist sounds very close how I would answer "How does success look like for your team?"
     
    mandisaw likes this.
  29. optimise

    optimise

    Joined:
    Jan 22, 2014
    Posts:
    598
    Hi @petera_unity. I miss the livestream. I hope I can get the Youtube video soon. Anyway I hope production team can stick with one project and just keep updating it instead of create 1 new project then throw it away. It's something Unity always do that create demo project every time then no more new update after that which is really bad.
     
  30. joshcamas

    joshcamas

    Joined:
    Jun 16, 2017
    Posts:
    1,061
    I agree, especially in these ever changing times, a demo being made a year ago makes it very out of date in the present. Case in point, Mega City Demo. Obviously updating a demo takes time, and with everything changing on a monthly basis it can't be updated all the time, but even a yearly pass would be very very good.
     
  31. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    It's in the first post, direct link: https://zoom.us/rec/play/vcYvf7qprD...OBGZeZLN_mY2Cmry6ftDiGArhAK?continueMode=true

    That's not why this team exists, you can read about their goals here: https://blogs.unity3d.com/2018/05/2...creators-behind-unitys-most-stunning-visuals/

    I agree, that way projects could be also reused to show new features and improvements.
     
    joshcamas likes this.
  32. ProceduralCatMaker

    ProceduralCatMaker

    Joined:
    Mar 9, 2020
    Posts:
    61
    I totally agree. When I read from the team "What we want to do going forward is to have a very small project kind of like Dots Sample that contains just the core aspects" I believe they are heading in the right direction, even if in my opinion even that is already a too complex sample. For instance, I had to spend a day just to understand the Terraformer ShaderGraph internal workings, not exactly a piece of cake.
    Concentrate on the real innovative elements of the DOTS/ECS functions.
    Keep It Simple Stupid (KISS) approach is best lo let others understand the excellent work you are doing! No need to overshoot, just to show off how smart you are.
    And a word in timings about new samples or alignment of DOTS and Animation would be welcome. So we also can plan our timings.
    thanks
     
    Last edited: Apr 24, 2020
    Fruei likes this.
  33. MRNXGBRL

    MRNXGBRL

    Joined:
    Apr 13, 2019
    Posts:
    7
    Nice and insightful stuff. (Still need to make time and get that FPS sample working one of these days) There was this question about 'why not expanding the team' (and a fair reply: reasons for keeping it small) and also about collaborations with other Unity 'teams' like e.g. the 'audio team' This would also make for a nice laidback Q&A session? Something i'd be interested in hearing more about and maybe future developments?
     
  34. Lethn

    Lethn

    Joined:
    May 18, 2015
    Posts:
    1,256
    Do you have any plans to implement a stencil buffer node in the shader graph?
     
  35. l33t_P4j33t

    l33t_P4j33t

    Joined:
    Jul 29, 2019
    Posts:
    108
    netcode has been available for some time now, has it not? https://docs.unity3d.com/Packages/com.unity.netcode@0.1/manual/getting-started.html

    Personally i would really like to see a dots implementation of global illumination, especially one in which you can move light probes. currently, light probes can't be moved at all which is a big bummer for anyone trying to create a game involving procedurally generated rooms.
     
    Last edited: May 3, 2020
    Fruei likes this.
  36. Kamyker

    Kamyker

    Joined:
    May 14, 2013
    Posts:
    397
    That's preview and 2 months old. They are using unreleased version:
    Imo dynamic occlusion culling way more important as it would help all projects with runtime placed/generated objects.
     
  37. olenkaAtmoetsi

    olenkaAtmoetsi

    Joined:
    Apr 29, 2020
    Posts:
    6
    Multiplayer in general? Or are you interested in multiplayer for AR specifically?
     
  38. optimise

    optimise

    Joined:
    Jan 22, 2014
    Posts:
    598
    Multiplayer game in general.
     
    olenkaAtmoetsi and EmilKoel like this.
  39. EmilKoel

    EmilKoel

    Joined:
    Jan 19, 2019
    Posts:
    6
  40. ScottPeal

    ScottPeal

    Joined:
    Jan 14, 2013
    Posts:
    53
    Will you update the DOTS Sample to 2019.4 LTS? How about 2020 current beta?

    Also, we need a tutorial on how to create a rigidbody first/third-party 3D character controller in pure ECS with new input system also in pure ECS. Oh with animation in pure ECS also. DOTS Sample seems to be a hybrid, but not really sure.
     
    Fruei likes this.
  41. ScottPeal

    ScottPeal

    Joined:
    Jan 14, 2013
    Posts:
    53
    Is Unity preparing a response to the features Unreal has just shown in their next version? Will Unity have the ability to have basically unlimited triangles and no need for LoD in the future?
     
  42. ScottPeal

    ScottPeal

    Joined:
    Jan 14, 2013
    Posts:
    53
    What works best for me is to have a video series that walks through the DOTS sample to basically recreate it from scratch. Sometimes it is the little comments mentioned in the video that give me the biggest value.
     
    ProceduralCatMaker likes this.
  43. ProceduralCatMaker

    ProceduralCatMaker

    Joined:
    Mar 9, 2020
    Posts:
    61
    Totally agreed. I have been asking exactly for that, but DOTS Sample in basically undocumented. So, what is the scope of such a big effort if left abandoned like that?
     
unityunity