Unity development kinda... slow?

Please excuse the slightly clickbait-y title...

I don't pretend to know how it's like to work on or manage a project as large as Unity and I can only imagine (and even then I'm probably wrong) the complexities and issues that might arise.

And this is nothing more than how I feel so I won't pretend that anything I present is fact.

But.

Doesn't Unity development feel a little slow these days?

I mean, I get it, the whole early 5.x was rough. Too many things changed and a lot of them kept getting changed, bringing even more issues. So everyone was busy trying to fix things to bring Unity back up to a usable state.

During that, Unity hired more people. But obviously you can't expect those hires to pay off immediately.

But I was thinking/hoping, that once Unity has a stable baseline (which I think Unity has achieved with 5.6 and after) and all those new hires had a chance to acclimate with the Unity codebase and culture, at some point all those things would start to pay off in a big way.

But I'm reading the 2017.3 beta release notes and the thing I'm most excited about is:

Which, is cool and all, but then I remember that the Animator was introduced 5 years ago, and zooming was an often requested feature which would be useful from the start.

And I feel a little sad.

I could bring up other examples, but I won't, since there are also other examples where Unity has improved a lot. Also, when I try to make a point I tend to exaggerate so take everything with a grain of salt.

But, I often can't shake the feeling, that 2 1/2 years after the release of Unity 5.0, every new release does not bring a lot of interesting things to the table and instead, we still just get a bunch of fixes for things that broke during the 5.x development while also managing to break a couple of new things, making Unity development feel like 2 steps forward one step back.

But maybe it's just me.

 

It's just you, mostly. There is a lot happening but it's spread out in different teams. SRP, Post effects, more networking. Navmesh work, lots and lots of work on getting transforms optimised, physics, both 2D and 3D, overhauled particle shaders, timeline, cinemachine, etc, etc, etc, etc


The problem is you see a little issue that could be fixed and so Unity must be slow. But in reality Unity is like "I got these huge jobs first! and all these bugs."

So they do those first.

But that's just it. It's stable for those features but after 5.6 there are some huge changes in the pipe that just reset the game all over again.

It's probably partly why you need to lock down to a specific version for a shipping title.

 

Yes, but none (or little) of it is polished or robust enough to be (very) usable yet so for me it feeds into the feeling that we're not "there yet" (whatever that means )

Well, fair enough

 

Darn I edited cos I thought maybe he'll take it the wrong way, but glad you were OK

I think if we look at UE4... There are big hitter changes. You get a big feature added but zero polish to it. It acts wonky, it has a poor workflow to it etc. I think Unity is kind of going there a bit. I prefer that because I want to use stuff, not wait years for some nuanced sparkle edge glimmer fx to shine along my nested prefab button.

 

There's a thing called Brook's law:
https://en.wikipedia.org/wiki/Brooks's_law

In general I'm pretty happy with unity not breaking more things.

I think this is normal.
You can't really expect new amazing feature with every release, especially if you want it polished.
I mean... you can take a look at any big product that has been in development for a long time and check release history.

Unity right now is at the "slow and steady development" phase due to being big. So, this is an expected release speed.

 

I'm not nearly as frustrated with Unity as I was in the 5.2~ (or so) era and I think I'm more calm in general these days, soooo no worries

I was actually feeling really good around 5.6, because I felt Unity was finally a solid product and the baseline of some features I wanted was there (namely the progressive lightmapper).

But then a couple of things happened:

I tried to use Timeline and I thought it was pretty undercooked (no events, pretty awkward workflow for some things and a ton of minor issues) even though it was in development for a really long time.

And the progressive lightmapper development feels like it kinda slowed down to a halt. (I mean, it still doesn't have full transparency support). I mean I was sort of expecting a faster rate of development on it after the initial 5.6 release. I actually think the developers expected the same as well, but apparently they ran into some issues which slowed them down.

Plus, a ton of silly things break during patch releases (like the current "tab" issue, a weird parenting issue a while ago, audio filters acting weird, collab somehow breaking script compilation, vertex colors being broken on android sometimes etc etc), which then make it to the normal releases as well and sometimes take a really long time to get fixed.

And all of that (and more) eventually got me to start feeling like what I'm describing in the first post.

Cool, but I don't feel like we're getting polished features either. Like, I don't feel Timeline is polished for example.

 

For a large product, they're moving very fast. The new C# job system, scriptable rendering pipeline, Cinemachine, Timeline, lightmapper, postprocess stack, new .NET, big VR & AR improvements, new platforms, etc. Timeline may feel not fully polished to you, but the major part of this first go-round was the underlying changes such as playables. They're also finally tackling a lot of tasks that don't result in bullet points on feature lists but improve the core engine, such as optimization and restructuring. I like the fact that they're putting Timeline out early in its current state to get real world feedback rather than spending time polishing a design that the userbase may want to see changes in.

 

Well, sure.

But this is where the pessimist in me remembers the Animator example I brought up in my 1st post, where they got feedback that zooming in/out the animator view is kinda important, and then.... half a decade later, they finally did it!

The optimist in me wants to believe that the major kinks/missing features will be worked out in a couple more versions, but reading the release notes for 2017.2 and 2017.3 doesn't exactly fill me with confidence.

But I guess we'll find out.

I'm not trying to convince anyone that the development of Unity is slow. I just wanted to express how I feel (or, I guess, more accurately, vent a little ).

I'm actually kinda glad people are disagreeing with me.

 

Timeline is not for production IMHO, it's still needing a lot added before I would consider using it. The difference is... it's here now, and Unity wants people to hate it but tell them why.

I don't think it's considered finished by Unity either... but that's kind of my point with the UE4 thing I said above: It's better to get something out and people hating it early, so you know what people's use cases are. It's kind of almost pointless to ask people what they want, so just get live feedback I guess.

What I am saying is: Unity is moving at the expected speed I would expect them to move at, and they are still not very good at doing that yet. It's a big change for a company to go from 2 year release cycles to 3 month release cycles, and those cycles need new features, which is always going to be a pain point.

I do expect them to get better at it and ask them to do so.

 

Okay, yes, I like this. I think I'll try adopting a similar stance.

And I guess, the animator zoom example I've been using negatively so far, can also viewed as a sign that they're starting to establish a workflow that allows them to work on long-standing but minor things like that, which is a good thing.

 

Unity development has always felt slow. Remember how long we waited for the UI system to come out?

I would run with the assumption that anything out now can be used, and nothing announced will be ready in time for you to ship your game.

 

I mean, yeah, but I didn't really care about it (still don't) And... the sad part is that I don't feel like it's a great feature.

Actually I think I had a mini-epiphany. I think what bothers me most about Unity in general is how they deal with features. A lot of the features follow the following pattern:

Feature is introduced, it's kinda buggy and incomplete and maybe slow, but seems promising. (features that fall into this: Timeline, Mecanim, Umbra culling and its thousand different implementations they went through, UI, Shuriken, Progressive Lightmapper and more). Then the development team goes into bug fixing mode and ...never really get out of it. And the features never really evolve and never get as robust or complete as they should be.

(I mean, I always felt like the UI was abandoned, until a dev assured me they were doing fixes non-stop, but is the UI really that much different now than it was first introduced?)

And by the time they are done with fixes, the feature is outdated and instead of evolving it, they do a re-write.

But maybe that's the main reason they switched to the more incremental model, with smaller and more frequent releases. In which case as hippo said, I expect them to progressively get better at it.

Yeah, I read that a lot on here, but I don't agree. My games also have pretty long dev cycles, so maybe that's why I often times disagree with this.

Ehh. I don't know about that. I can't currently use Timeline. It either misses features I need, or its workflow is a nightmare even compared to silly tools we hacked together.

And I'm using the Progressive Lightmapper but I do have scenes where it's either way too slow to ever actually finish the bake, or where I actually *need* full transparency support, which is still not supported.

Both these are features that are out now.

But, as I said, I'll try adopting a more positive attitude. I'm hoping we will see a semi-fast evolution of Timeline and that once the Prog Lightmapper team gets out of whatever troubles they got into, we'll have a more regular stream of improvements for those features.

 

I was thinking more nichevo then positive. There is nothing that you can do about Unity's development pace, so why worry.

 

Compared to Epic they have always moved at a snails pace, on the other hand they retain backwards compatability (so you can upgrade without a change of undies) and all tutorials aren't defunct in the space of three months.. In terms of reliability UE has come a long way, but some times you just wish they'd stop dicking around with stuff..

Good example was the bloom post effects quite recently, they went for pure "physically" based and all the cinematic atmosphere was gone in a second.. Short of re-writing the post stack (as their TDR heavily relies on certain components like AA which is tied into everything else (lighting etc.) you can kiss the cinematic look goodbye..

If they do get lots of feedback to revert then it's slow.. There are pro's though, you get an amazing set of tools and Unity is a little ancient in comparison.

But end of the day you want to release a game right? You don't want to upgrade for bug fixes and have half your game completely change how it works. So for me it's just another set of pro's / con's you have to consider.. I like both approaches for different reasons.

It'd be nice to have a happy medium and I'm sure at some point it'll happen..

 

I am not sure if I worry, but there are certainly ways I feel about it. It doesn't mean it keeps me up at night. But yeah, I get what you mean.

I mean, yeah, but Unity right now is backporting fixes (so you don't really have to upgrade to get bug fixes, well, if you aren't too far behind), and they are being conservative with how they deal with features. I'm hoping for example that they're not too afraid of changing Timeline around if it means a better workflow, for fear of breaking backwards compatibility, since we're early in that feature's life.

A happy medium would be nice, yeah, I agree.

 

In regards to timeline, I don't think anything usable will come of it until mid 2018 really, just guessing but it doesn't seem like a trivial task.

What does asset store have?

 

I've heard good things about Slate, but haven't used it, so I can't recommend it.

I don't know, I don't really like relying too much on the asset store these days, especially since we're generally moving to the latest stable-ish Unity version all the time and having assets block the progress is a bit annoying.

We have our own solution, which is based on the animator (since when we started there was no playables api), so since it's animator based, we hit quite a few limitations, making the workflow less than ideal. It's a bit awkward to use, but at the very least it has all the features we want (like, events, and weird ones like using an audio track as the cinematic's timing, so that even if I do weird things to Time.timeScale, certain clip changes and events are always synced with the audio etc).

Obviously I never expected Timeline to have exotic features like that, but I was kinda hoping I could use it only for technically simpler cutscenes and enjoy a better workflow. But even for those, the workflow for creating and editing a lot of short camera clips quickly is kinda bad and buggy and I need events, so it's a non starter.

 

"In general I'm pretty happy with unity not breaking more things. "

Yeah maybe if you're sticking to 5.6.. if you've moved through 2017.1,2017.2, 2017.3... things have been breaking.

Maybe its always been this way, though I seem to recall 5.3>5.6 was pretty stable even in the beta versions. Anyways I both agree and disagree with OP, is more teams doing more things, but seems like internal problems causing delays etc

I think the problem is that overall actual general editor improvements useful to 'everyone' seem to be very low.. as mentioned things like the zoomable animator window should have been improved ages ago, its a general editor improvement the vast majority could benefit from. And there is so many more things like that, though looking at what will be coming with 2018 should finally be some workflow improvements for all, feel like I've given two feature suggestion examples in the past that are literately coming into the editor for 2018 just wish it was sooner. Still not enough and while all the releases have some notable feature, the last big one that I can think of, that benefits an overall everyone audience (with it atleast being pretty feature complete) was the inclusion of TMPro.. and that wasn't even on Unity development time.

All the other stuff that has been coming is either still early in design and features or just really game type usage specific that not everyone will have a use for it. And big things like the C# improvements .Net 4.5 is still experimental, I still don't use it yet, probably won't until they get it to where we can use C#7 with ref returns now.. maybe I will sooner than that though.

Also the number of previews for things is growing, but overall feels like more previews = less actual shipped improvements in a release that is usable and not just testing/feedback release.

And like I'm disappointed with the input improvements not making it into Unity with 2017.2 or even .3... things like the 2D tilemap stuff is great if you're doing 2D and its pretty generic enough to be useful for a lot of 2D game development. However 3D games.. Timeline, Cinemachine to me seem more useful for much later stage in the game development.. Where as tools and improvements you see Unreal adding like in the last 3releases are not only similar things to Timeline with pipeline to Maya, but you see them also adding more editor like tooling for early/mid development like in 3d character improvements with ik/physics/cloth etc.. or terrain tooling (albeit VR) or terrain texture stuff.. though with Unity you can rely on the asset store for a lot of this stuff to complement what they don't directly add at extra cost.

The Unite FBX improvements with Autodesk 3ds/maya are good, but how does blender not get improved pipeline support first when that is not only freely available to a larger audience, its open source aswel... obviously anyone outside of Unity wanting to bother doing similar would be somewhat(maybe) hindered by Unity's closed source side of things.. it just leaves the question of why Unity didn't bother with that first or atleast add it for blender at the same time. (I don't even use blender but still makes you wonder)

And if you remember the Unite demos about this https://github.com/Unity-Technologies/VolumetricLighting last year, there it is like updated 6months ago, its not built in or anything... I literately only remembered it because it was mentioned in the recent blog post...

I compare that to Unreal's (yeh someone had to bring it up for comparison sake) release notes where they have volumetric fog added in engine and compare to Unity's release notes you might think wtf is Unity doing, but overal they seem very much on par, it's just funny that for a closed sourced product Unity now have a growing amount of stuff out there in various forms of preview or github/bitbucket and that's just its own internal staff... while Unreal seem to pull it altogether into an actual release and the release notes really pack all the stuff they've done with images and gifs to show it all.. Unity do aswel just not as good.

So again I both agree and disagree, also getting tired of the beta releases... Overal useful improvements here and there. However 2017. releases nothing is really finished and great unless you have some good uses for things they have put out in 2017.1/2 so far. Still lots of good things coming, I have no idea what Unreal have in the future pipe, but I honestly think is more exciting stuff coming for Unity.. Maybe they could cut releases down and just have 1-2-major and 1-2 minor releases a year outside of the patch releases and a little less of the constant all year round public beta testing, 2017.3 beta is already out and 2017.2 hasn't even had a proper official release yet.

Also they should step up on the training videos on new releases of features that come out with Unity, if anything it does showcase things better than expecting users to try explore it completely.

 

My only gripe is it seems the editor doesn't get any love. It still has the same outdated user interface since the 2.x, 3.x days and little has changed.

There are so many improvements I can think of Unity can make to the editor but yet here we are with still exactly the same context menus, limited heirarchy/scene search features, bug regressions and things from the 90's like the whole editor freezing up when Unity is recompiling scripts.

Hell even the alert dialog/building dialog seems totally outdated and clunky.

I could go on for days... really...

 

@Player7 From what I can see the scriptable render pipeline HD path supports skin, volumetrics etc and that's done now, you can probably download it and see where Unity is going with it. It's headed up by Sébastien Lagarde's R&D team, and can be found on Unity's github. I had a go on it and came away thinking that it would most likely be absolutely fine as a modern dice/doom hybrid with everything anyone would want really.

You CAN do this stuff now, just nowhere near as efficiently as the HD pipe potentially will be. The Adam demo thing, I don't know if its the same one used in SRP.

As for epic's releases, well good on them. If it wasn't for them, Unity might well still be sitting on its arse

 

I see this said all the time across reddit \ unity forums but have never understood it. Can anyone show what modern 'game editor' UI is? I find Unity is the best of the bunch in this regard.

Yes Unity development is fairly slow, I don't think it should have taken more than 6 months to get a bare-bones tilemap into a release. Hopefully they will be less afraid to 'release' things when all of Unity is modularised into the package manager because it will allow them to independently update \ add features fairly trivially.

 

Free dark skin!
Honestly, apart from my complaints about drag'n'drop and hierachy problems that I already mentioned elsewhere I have no complaints about the GUI. Don't even mind that it's not dark, as long as it isn't the eye-blistering white of a thousand suns that the forum likes to use. I would like to have more hotkey functions set by default.

My main editor-related complaint is lack of precision and sensible scaling of movement speed. I have one project where everything is super big, 2 x 2 km level area and every thing is several meters big. It's always uncomfortably slow to move around in scene view. And then yesterday I wanted to render some assets in Unity for my portfolio (just take screenshots from game view), and they are just a few cm big. It is impossible to move such small things around precisely. I ended up just scaling everything up by a factor of 10 to be able to properly place the things. Stuff like this I just can't understand, because I never have such problems in Blender, even when working in extreme scales. It's a night and day difference how uncomfortable just placing assets is in Unity compared to Blender. But I miss the WASD viewport movement in Blender, and the move/scroll/rotate button defaults in Unity probably make more sense as an industry standard. So it's not all bad.

In terms of development speed I don't have much of an opinion because most of the stuff they are working on isn't relevant to me. I just wish there was a bit more focus on opening up blackboxed low level systems to allow people like jbooth and other assetcreators to make cool specialized stuff for us, that can allow themselves the "luxury" of not needing to target all available plattforms Unity can deploy to. I'm sure if Unity was Desktop only, we'd see them move forward a lot faster, and personally I get nothing out of all that cross-plattform compatibility. Not complaining really, in fact it's remarkable that in spite of that Unity is still likely the best fit for me.

 

Couple of tips: hold alt when dragging a value in inspector for fine grain movement.
Hold shift while flying in scene view to fly fast. Alternatively, middle mouse pan. Alternatively click any object twice in hierarchy panel or click and focus with F over scene view.

Can easily script a few tools to make navigation around the currently developed game a lot easier, we do it this end from time to time, to morph editor to game requirements.

 

Thanks for the advice. Holding alt while tweaking position snaps it to .2 increments for me (20 cm if you take 1 unit = 1 meter, which I think is intended). I was needing more than 1 milimeter precision because the object was only like 3cm big. Camera movement with scrollwheel can't be influenced by any of the modifier keys. I know about shift while flying, it's still slower than I want it to be in my case. Middle mouse pan seems to struggle with floating point accuracy or something, because around an object positioned at 700,700,700 and the screen centered on it by pressing F, the MMB pan already behaves super wonky and juttery.
Editor scripting might potentially solve many or all of my problems, but that's something I can't motivate myself to dive into, because I feel like I really shouldn't need to have to deal with this to that extend, and I think I don't do enough manual placing that I'd ever get that time investment back.

 

I find the best thing to do is to use a stable release such as 4.7 and then just add the features you need with scripts etc.

I recently started a new game in unity 5.6 but now I'm regretting it because there are subtle changes. So I'll have to waste a lot of time testing on different platforms etc. and relearn all the quirks to do with 5.6.

What I should have done is stick with 4.7 which I know really well.

As an example. You want nested prefabs? There's a few scripts out there that do more-or-less the job. Better than sitting in a dark room for 5 years hoping Unity will be updated. Kind of reminds me of communist Russia. Where everyone just sat there wishing for the government to update there black and white TVs to colour.

 

My plan currently is to do main development in 5.6.3p4, because that one is very reliable on iOS. When I want to ship a macOS build, I will duplicate the project, open the duplicate in 2017.2 (for retina support) and build/ship it with that. Then continue development in 5.6.3p4. I don't plan on upgrading my project from 5.6.3p4 for a long time now, that is, until I am completely forced into it.

 

I've been out of the game-dev loop for about 8 or 9 months, and the first thing I did when I decided to get back into it was review whether we should stick with 5.5.x or try jumping to 2017.x -- mostly for the C# and .NET Framework bump, which has been a major pain-point for me.

Even having a pretty good idea of what was coming (mostly thanks to Unite talks), I still found the number and scope of changes somewhat daunting. Nonetheless we'll try to make the jump once the .NET updates are stable. Easier for us being solely desktop-focused but I was also surprised at the number of changes that do impact our project.

 

This is worrying.

 

Yes. Very slow. A lot of the things they're working on would be side-projects in other engines (and/or already be included with the engine).

They seem to be taking less "risks" and doing more "dink and dunk" stuff as a result. I still think Unity does a lot of things right, and as a developer features can never be implemented fast enough , but their innovation seems to have slowed quite significantly.

I mean, personally, I'd like to see Unity have the best terrain system, and a vegetation system that matches engines like CryEngine. But maybe it's just me.

I always want to push the envelope in game development, and right now, Unity doesn't seem to be that engine unfortunately. That being said, I still use it all the time because it's ease of use and flexibility is excellent. And I don't want this to be a "hate on Unity" post, because Unity does fill a niche in the game engine market that no other engine does right now.

I simply want to see Unity be the best engine it can possibly be, and it seems slow right now

 

I dont think development is slower at the moment, it's just a lot of the most important developments have been at various foundational levels for quite some time and most of these are not quite ripe yet (but are clearly getting much closer). Stuff that changes what we can expect in terms of performance and modularity of unity and the stuff we build with unity. And many of the future tools and higher level systems they will build for unity. I'm happy they are reworking some of these foundations rather than just building more obvious higher level stuff on top of old parts that have old limitations.

Yes it has felt a little strange at times to have unity event keynotes/presentations which dont feature as many 'headline features' of the eye-candy and/or tool variety that we we used to hearing about in greater multitude in the past. But I can more than live with that so long as the new foundations that eventually arrive live up to their promise eventually. I'm saying 'eventually' a lot in part because Unity were also upfront that one of the things the switch to a subscription model meant was that they didnt feel the same degree of pressure to include all the shiny new toys in the first version of a major new release. I'm fine with not having timescales and priorities distorted negatively due to sales/marketing pressure, but really it's still too soon for me to judge whether any new problems arise from the change of release model.

Personally I will get one of my first real chances to judge how well they are doing in the 2017.x era by watching the progress with scriptable rendering loops stuff far more closely in the next 3-6 months than I have been in the past, in the expectation that at least one of the example rendering loops they are providing is nearing initial maturity.

 

I would be happy if they would stop adding any new features for a year or two. And get rid of all bugs and problems instead. Anyone who ever tried to publish on Mac App Store with support for Game Center knows what I mean.

 

Not really; they specifically recommend not using patch releases unless absolutely necessary, precisely because they haven't gone through the usual QA.

--Eric

 

Funny.

For me it's always basic things that are broken with most (or far too many) updates. Things like reflection probes just not working at all, or in 2017.x now the move/rotate/scale gizmo doesn't update to camera rotation, so grabbing the x+y axis handle is hard. But at least it turns yellow when dragging (who asked for that?). Then there's the basic scene view picking problems (SelectionBase seems to do nothing at this point) and scene view stuttering that won't be fixed ever.

 

I thought patch releases are meant to fix bugs only?! If fixing bugs tend to introduce new bugs, then that's something to worry about.

 

Yes, fixing bugs can obviously lead to new bugs. It's happened to me, and probably everyone who writes software of any complexity at all. The point of patch releases is to fix a show-stopping bug ASAP, but again, it's advised not to use them if you're not directly affected, or can work around it. Therefore it's not something to worry about: if you don't need patch releases, don't use them, but if you have to, be aware that they didn't go through the usual QA. No worry involved, just basic info/logic.

--Eric

 

Biggest problem is that there are game break bugs that still exist, some that are down prioritized beyond reason like

https://issuetracker.unity3d.com/is...hile-colliding-the-edge-of-another-gameobject

 

I totally agree. I would also love to see Unity have the best terrain system. I hope they are working on it now. I also would love to see a really robust vegetation system as well. I just wish sometimes we would get a sneak peek into their new terrain system and see how they are retooling it and what improvements they are making it. Maybe just a teaser trailer or a acknowledgement on how far they are on the new terrain system. I also would love for them to improve the UI of the editor and address some accessibility issues (Editor Font Size). I love Unity and it has been fun to see all of these new features being added. However, I do wish they would also spend more time on improving the editor UI and work on improving the core engine.

 

Terrain and foliage that work in VR *drooling*

 

Slow? do you realize how hard it is to make a engine? and how much work is required? plus understanding og math and how you actually do everything dynamic? jusy Fyi Unity 2018.1 will have a insane new update which basicly rewrites the entier engine. and allow us to use multithreading and new compiler etc. from what they showed its about 15x faster then mono

 

Quite the contrary. I would say Unity pace of development is really fast now - in fact probably the fastest of all engines. There are 500 devs working on Unity core. Just look at the 2017.2 release notes. Huge number of features in it. There will be more in 2018.

 

yes so fast that metal tesselation is still not there, and 2d smart sprite, and many more things that were annouced long time ago

 

Have you noticed yet that the newer release notes have images showcasing the features? Older release notes didn't have that. They were purely text release notes. Go read the release notes for prior releases (eg 4.3) and you'll see the lists were far bigger back then than they are now.

 

Do you know why the list is shorter now? I started using Unity 5 in July 2016 and I didn't know how much features they had implemented in previous Unity versions.

 

I can imagine, but I would probably be far off from reality. Same as you, I guess.

All I know is that progress in the 3.x and 4.x cycles felt faster and the progress of the competition feels faster as well.

There's a ton of fragmented development going on in various githubs and bitbuckets and experimental previews, but I can't really count on those, right? Unless I see them in a proper release, they might as well not exist (because, they could get cancelled, or completely changed or whatever).

I said it before, I was expecting that once Unity got somewhat stable (which I think it recently did), and the new hires had some time to acclimate, that all those things would start paying off and we would get an accelerated development speed. Which doesn't necessarily mean new features, but fixing and improving long standing issues and old features (I mean, when was the last time we got a decent update to navmesh? If they keep not improving it, I think we'll quickly get to a point where it needs to be completely replaced by a new system)

 

I think the thing that bothers me is that quite often, a lot of (admittedly not super-major) issues that get introduced in patch releases, make it into the final releases as well.

 

I often think how neat it would be if they could do a "bug-fixes only release", working an entire release cycle on nothing else than fixing bugs.

Then I remember they have the Sustained Engineering Team, which as far as I remember, is dedicated to fixing bugs and making sure they get back-ported and whatsoever. I didn't refresh my memory by reading the blog post again, I could be wrong.

Thus, it seems they actually do bug-fixing a lot already. Be it their separate patch releases and of course during a beta cycle.

How come that quality often feels sub-par then?

 

Have you guys used other engines? For me Unity is the most stable from what I've tested and I tested UE4, Cryengine and Lumberyard.

 

Ah yes, a thread critical of Unity shows up, so everyone comes out of the woodwork to go "WELL ACTUALLY" about other engines, all while ignoring the very real issues Unity has.

Here's the deal: when big new features come out, they're typically a pile of hot garbage until several releases later. Mecanim, Enlighten, and the new UI all come to mind (and the new UI is absolute piss still). There's also the issue that hotly requested features (new input system, new terrain, general prefab improvements) seem to go completely ignored. It honestly feels a lot of the time that Unity is focusing big on implementing new features that look good as bulletpoints on a release list rather than fixing the things that have been needing attention for years.

Honestly, it feels like Unity needs something like MacOS had with Snow Leopard, where new features are largely ignored in favour of fixing what needs it.