What makes Unity v2019 useful (or even exciting) to you?

Hi,
I am working on two projects for two different companies. One is using v2017 LTS and another v2018 LTS so I have not seen much of v2019 so far. A Toon Shader client asked me recently for v2019 compatibility so I decided to have a look at v2019.

New features in v2019 that could be interesting are missing (network) or are gonna be abandoned (HDRP). Support of the latest platform updates (IOS-AR-VR-MR) does not seem so necessary to me (IOS profitability became lower over the years, AR-VR-MR are not there as products and will not get there soon). v2019 does not seem so interesting to me after a superficial inspection.

What features of v2019 do you find useful - exciting?

 

WTF? Why do you think it will be abandoned?

---
BTW, stuff I like so far (including all 2019 versions):
- HDRP support
- Incremental GC
- Enter play mode without domain reload
- The new UI is much better than the old small-text gradient one
- New Input system looks very good to me

But nowadays the editor versions and packages are separating, so more and more "excitement" will be and are shifting to packages rather than the "new version" of the editor. (like HDRP and Input system, Quick search, Addressables)

 

More ECS / DOTS ....

 

Thanks all for your feedback, it is much appreciated.

I should re-phrase this part of the post.
HDRP is still in preview and since Enlighten is gonna be abandoned-replaced, high-quality rendering in Unity is and will be an ugly mess for some time. Practically by removing the Enlighten backend, all the fine-tuning people do to achieve high-quality lightmaps, etc will need to be redone, just as it happened before during the Beast->Enlighten transition period.

I guess you are mentioning UIElements. We cannot use them in-game, only in editor. Why would someone use this to e.g. create an editor tool, since so many projects are using v2017-v2018 which do not support this feature?

This is an experimental feature in 2019.3 - which is in alpha. Having an ssd and a CPU with a large cache can help a lot with the delay until entering playmode in the Editor.

This will hopefully appear as a package in v2019.3, around the year 2020.

You are right. I try to adapt into this new paradigm - packages that are adding functionality instead of Editor features, which is great for adding Editor functionality - the FBX export package is a lifesaver. I should examine the available packages more in-depth.

What I find annoying regarding ECS-DOTS is that I need to re-implement Monobehaviour functionality to use something that might work. And this really sucks. It is easier than writing native plugins to gain speed but still sucks.

Please do not interpret my POV as a nihilistic one - I do not feel like every new feature is worthless, they are interesting. I dislike the fact that most of these features are experimental and in-preview - it's like "use it at your own risk" - I do not wish to risk when someone pays me for consultancy in a project, so these features are out.

 

Well, true is, not every feature if for everyone. Some features need start somewhere, before mature. And we are is such pleasant position, we can observe, how these evolve and try as well. We just see lots of previews and experimental stuff, in contrast, what we have seen in past, where you had only given things ready on the table. And yet couldn't try upcoming stuff, since where "hidden" behind Unit dev team curtain.

In current state ECS / DOTS is in preview, not ready for production. Simple as that.
However, you can use fully functional Jobs, if you consider performance gain. People tends to overlook it.
But that may require redesign part of your project(s), if you far in progress.

Either way, if you in middle of the project, you should focus on what you got, rather what is upcoming.

What are your Unity expectations for new project? Is there really anything you will need, what you can not achieve already in your case?

 

I'm pretty excited for the improved maximize button. It makes a big difference in how many WTF's per day are generated when trying to maximize a window.

 

I respectfully disagree. They have plenty of time to bring to the table either the GPU or the CPU progressive lightmapper. They aren't in that bad shape (they need some work, but not that bad). You need to follow the change? Sure. But every software works this way. You signed up for this when you decided you're using an engine instead of rolling your own. Especially if you're using a preview feature. (The preview-ness of the feature is not a problem, it shouldn't be, we should be glad to be able to play with these new stuff to learn what will come when they are out of preview)
With that said, I am too have some doubts how this will play out regarding the HDRP, but since the feature is still in preview I wouldn't bury it just yet. But I'm not that pessimistic. (disclaimer: I have started to play around with HDRP but I decided to not to use it just yet, it's too preview for my project, I don't have the resources to maintain it, but that's my specific situation, HDRP is not worse because of this)

Partly, and the actual editor UI. Those who don't care about 2017 and 2018. Those, who start a new project nowadays or in the immediate future. I'm not an Asset Store publisher. I don't need to minimize the work between versions. I do whatever I see fit. I skipped 2018 entirely (other than testing and stuff), when our project will be mature enough, hopefully Unity 2019 will be mature too, until then I am willing to pay the using bleeding-edge-price.

Actually it's very usable right now. There are some problems with it with specific situations, but again, just started a new project, I can afford to use the preview package. And it's working for the project so far (currently I do not need anything special in the input front, but the editor and the input handling is so much better than the old input...)

And? Your question was what I find exciting in the 2019. I'm currently running this new project in 2019.1 and in 2019.3 (so far not too many problems but again, this is a brand new project from ground up, will see if the 2019.3 beta will be stable enough to go forward with it or not. Currently I'm cautiously optimistic about this.

 

??
Is that a windows thing?

 

I did capture the screenshot on a Windows PC, no idea how it looks on Mac. Maybe it's a Windows thing only.

 

This info is very useful, I should check it out when the time comes.

This is useful too (when people disagree). My point is, why invest time with HDRP since it will be transformed at a fundamental level in v2020? I could simply skip HDRP in v2019 to save the hassle of learning HDRP twice and having inconsistent results between v2019 and v2020. Following the change makes sense sometimes, not always.

I really mean this part. The fact that I personally do not wish to use experimental features means nothing at all, I am very interested in other people's POV and I appreciate the effort of responding in this thread.

I might have overlooked it as well. Could you please suggest some source-link that provides some hands-on useful Jobs implementation besides this one?

My case is that I am a senior Unity developer and I get paid to advise people with serious budgets on questions like "which Unity version shall we use in our project". One project required Win 7 32bit compatibility - only 2017.1.p3 fitted the bill. In another project the new prefabs system was the deciding factor to jump to v2018 while it was in beta, it worked well, after some minor bumps. I am looking for killer production-ready features in v2019.

I think what Peter77 points to is that now it is slightly easier to hit the icon

I try to decide if v2019 worth it while I am very short in time to explore it and your feedback helps me a lot to narrow down my efforts.[/QUOTE]

 

Nothin' on Mac.

 

Make sure the window is undocked. Those buttons seem to appear on Windows only if the window isn't docked.

 

Ah... heh on Mac it looks like old versions of MacOS.

Undocked.... eww.

 

I like these three dots. I hope they'll keep them.

Anybody could post an image of the Linux editor?

 

Alot of stuff a little less preview in 2019

 

https://blogs.unity3d.com/2018/10/22/what-is-a-job-system/

If you understand Data Oriented Design and technical aspects, you can figure out tons of applications. Otherwise, if you not familiar with DOD yet, you may need to read something about first.
You can also search Job System Examples.
But as an advisor, that should be obvious, so probably no need to mention.

 

Eww, how anyone can use Mac os as a workstation OS is beyond me. With all its animated stuff etc. It's a toy

 

Yup, Google wanna be my friend, I found this one, looks decent.
Personal recommendations still have a higher value than a google search in my book - if anyone wishes to provide a link to examples that he/she/whatever finds useful, please point me to these examples.
Thanks again for the feedback.

 

You can turn them off, just as you can in Windows.

 

But you still have that awfull wheel at the bottom that makes no one happy plus you are locked to overpriced hardware, oh well.

 

Maybe because you can't build a player for every platform on Windows, like you can't build a player for every platform on Mac?!

So you use whatever gets the job done. Ego and personal taste/opinion left outside the studio.

 

Quit derailing someone else's thread with irrelevant platform wars.

 

@Ippokratis, the only reason I went to 2019.1 instead of 2018.3/4 is because the UI allows Inspector editing of one or more Prefab objects without having to go into each Prefab's editing mode individually. It's a feature that should have been in the box when they brought Nested/Variant Prefabs out of "preview" stage.

 

This is very useful, thanks.

 

For me, I'm wanting to be able to access the new cool packages - specifically the animation + search tool. That, and being able to edit multiple prefabs by selecting them. However the latter is going to be added to 2018 LTS.

Also, excited about terrain holes ) I know you can hack terrain holes, but it'll be nice to finally have real support for it!

I do agree that it is a bit frustrating that most features are "in preview", but remember that if we didn't have the preview, it would stay internal until complete. I think that being in preview does definitely give unity much less of a reason to actually finish the feature, but it is very nice to get a feature early on to test it and even use it for people who like living on the edge

 

This sounds useful, but I am currently using search and replace asset which works ok in every version I tried it.

Yes, having experimental features is great, for 2-6 months or so, but it is just frustrating to see that some features just stay in this experimental state for years or are depreciated. Just consider how long HDRP or ECS stay in preview (too long).

 

It's a bit like games that are in "early access" mode - the developer can still make money and fame, without having to fix bugs since "it's in early access". So why ever actually release it? I don't necessarily blame the developer nor unity for this, it's kind of a human nature sort of thing. But still, frustrating from our perspective. Regardless, nice to get new features early, especially ones that have their source accessible through packages(!!!!!!)

 

The package system all on its own is really awesome and one of my favorite quality of life tools Unity has released. I love that I can keep my project folders uncluttered while I import my own reusable code and assets from external repos.

Bye bye to the days of exporting “unitypackage” files and hoping everyone grabs the right one to import!

 

Because I am a professional and need power and stability . #NotAHobbyist.
Keep your opinions to yourself.

 

Plus there are some apps for macOS that are more affordable than their equivalents for Windows. Final Cut Pro is only a one-time $299 whereas Premiere Pro is a $20 subscription. Final Cut Pro pays for itself after a little more than a year.

You could potentially make up the difference in cost between an Apple computer and a PC if you had enough of these.

 

Hmm, mixed about it. There's many things to be excited about dependant on what you're doing of course, ECS / DOTS / HDRP / Terrain improvements, Timeline, Material editor, VFX Graph etc. etc.

Although some of the core fundamental issues still plague it exacerbated by what Unity are trying to achieve, navmesh definitely needs some work, PVS style occlusion culling (umbra) and frustrum culling is dire in larger projects (like getting five ms overhead for every terrain segment).

The profiler is still useless, the time domain thing needed fixing yonks ago (kudos for starting to address it). Also Unity has been in a state of flux for a while now. Sure I am just nitpicking at this point but I always say foundations first graphics later. I know it'll all get sorted out at some point and they're more on the right path now than ever (incremental GC (yes please))..!

I wish games would come together this easily.

 

Unity's Input API. While I'm comfortable creating systems to make proper use of the previous Input API, the new one has functionality that can't be easily replicated. Like being able to handle remapping for more than one player at a time thanks to being able to identify which player is using which controller.

 

Also, pretty much "uh.. macs suck and they are toys... blah, blah" is pretty much the the default troll comment from a 13 troll who has nothing value to contribute to the conversation. It's cliche, ignorant and unacceptable behavior here.

 

Runtime rigging, though I haven't got the chance to test it yet. It's however a feature that I was waiting for years.

 

If it works as a MonoBehaviour and the performance is acceptable, leave it alone. That stuff is intended for cases where there's so much data to crunch that a standard object oriented approach becomes inefficient.

 

Tried 2019 for LWRP but found that to be a little confusing to start with and the materials do strange things (white material with white specular turns black!?).
So sticking with 2018 LTS for now for robustness as most new stuff in preview only.

 

Also saying "macs suck" whilst simultaenously being a gamedev is like saying the entire world of iOS development is non-existent and pointless, when in reality tons of game developers have made a living developing for iOS (which by extension, requires a mac).

Anyway on topic, the most exciting thing for me would be runtime rigging and the general new terrain behaviours. But what I am most excited about for the future is ECS which I still need to learn.

Is anyone able to tell me whether ECS is worth learning yet? Or is it still in the middle of constant major refactors and changes all the time?

 

Nothing I can think of .. with exception to https://forum.unity.com/threads/on-demand-rendering-new-in-a7.701369/


ECS - barely know much about it, probably unfinished

LWRP - would like to use, completely pointless due to so many third party asset packages not having support for it.

New Input system - would like to use, am kinda of using and trying to add basic support for it with things I do now, but again suffers from zero third party assets even attempting to bother having support for it.. it is after all it is still unfinished, and off track on there own roadmap predictions for features/releases.

HDRP - Couldn't really careless - it seems to much of a hassle to support the highest shader path in stuff when you also want to support mobile, and switching between HDRP and LRWP is just a pain.. again not much third party asset packages support it either.

Incremental GC - It's free, you don't have to do anything, I like that.

 

Depends what you want to do.
See here for example.
https://forum.unity.com/threads/unity-ecs-and-job-system-in-production.561229/page-2
There was also other project in production, but can not find link atm.

But you can utilize it, to improve part of your game, where logic heavy. You don't have to make whole game in ECS / DOTS. Or use Jobs for simplification.

 

I think it's worth investing the time into learning about cache optimization and data oriented design as that's a very significant chunk of the gains made by ECS. It's also probably the biggest mental leap when learning it from scratch - the change in technical design rather than Unity's specific implementation.

It's not a tool that will instantly help every situation, but if you're ever thinking about making a simulation-intensive or data heavy game, it'd be worth investing some time into now, so that when unity's ECS is finalized you'll be able to jump in with a smaller barrier.