Please Unity, slow down.

If they want to save me as a PRO customer then they should pay attention to my feedback as a customer. If not, it's their loss.

 

All I got say is, its same old story, each week wait for patch, each week delay project from going live (mobile)
Splash Screen is bugged on iOS and 3 massive threads on it, yet its being ignored and some other crap is fixed that no one uses as much...So wait for another patch...Seriously just when one is ready to release and thinks things are good problems happen. Unity 4.6.9 is stable but its just sad that the new mobile stuff not been back ported. Get 5.2 stable before 5.3 really

 

I agree. If Unity wants to keep paying customers they should listen to paying customers

I also agree with the bugs, they feel an all-time high at the moment. I am also pretty disappointed in the docs, specifically the networking ones which are incorrect in some places due to API changes, and simply missing in a lot of other places or riddled with typos.

 

(slightly off-topic, but what's bugged about it? admittedly, I make my own LaunchScreen.xib file so maybe something is wrong and I didn't notice?)

 

I am also a pro user and I am not using Unity 5.2. It got pretty annoying after a bunch of recurring issues so I decided not to waste any more time with it. I also used the beta versions of it and it turned out that I got more issues with the official release, which should definitely not happen. It is out of question that Unity has to do a better job.
I just don't get this kind of mindset that someone has the illusion that the formatting of a post has any other effect than being ignored because someone tries to be louder and feels more pain than others.

 

That is your right to feel that way and to express that opinion, so I respect that. However, I respectifully disagree with your hypothesis. I am using the tools provided by the Unity forum to format my post to best reflect the emotional tone of how I truly feel. Since I am expressing both my own subjective opinion and emotions, I feel my formatting more accurately conveyed the urgency and passion I have in regards to these matters. Whether or not it gets more or less attention is not the point in formatting, the point of which, was to express my concerns as a customer in the most accurate representation I can through the tools provided by the forum. In those regards, my post has accomplished exactly what it set out to do. Since we're getting off topic though, I'd be happy to continue any further discussions regarding this via private message. Thanks for your feedback.

 

Yeah although moderators will edit posts if they're truly over the top. I figured it was just someone shouting Crom! and being generally irked by it all rather than trolling so I left it.

 

Wow, this thread is going wild and as some comments might be harsh, I really feel bad for all the guys working hard at Unity and trying to do their best to give us the best game making tool possible. Even if the current lack of stability can be quite annoying, it's far from being a total disaster. So for now, despite some recent bumps on the road caused by huge changes (like new UI system, mecanim, enlighten, etc), Unity is a great capable tool that we all love using. It's just that some adjustments are needed and I hope that our voices have been heard.

 

I have a large tolerance, I've put up with stuff for a while and logically I know it will get better. It's just that I really need to personally see some major reduction in bugs and increase in performance and better docs. I'd be very unhappy if from this point it gets any worse and not better.

I guess I'm just letting off a bit of steam hippo style.

 

I completely agree you with you!

I also believe that the fast release cycles they are having now are one of the causes for the quality of 5.2.x. As this is very new for Unity to have such release cycles, they still seem to be learning a lot about it. It is very difficult to do that in a small company and it isn't easier for big companies. But I don't think they should slow down! Instead, I am open to give them some time to learn and get used to this process. If they get it to work, this kind of release cycle is a huge advantage for us!
From my point of view, the beta of 5.3 is already in a far better shape than 5.2. This is a very subjective opinion, because I didn't make any systematic tests. I just opened existing projects and made some random tests that were tricky in 5.2. Because of that, my hope is that Unity learned a lot for their releases during 5.2 and that the 5.3 releases will be a lot smoother.

 

It's actually a very, very good post. Unity should listen. The current release strategy is simply wrong, given the resources Unity has at its disposal.

 

For me, the post basically says, the current situation is not good, it was better before, so go back. It completely ignores the potential of the new release cycle and it also ignores the fact that there were quite some people who didn't like the old release cycle. With the previous release cycles, people were complaining that it took too long between releases and that there was no possibility to get a new version, even if Unity had already resolved a specific issue for quite some time.
Right now they are struggling, that's out of question. Unity 5 was rushed and especially Unity 5.2 was not really a smooth experience. But with the new release strategy, they can release new features a lot faster, as well as bug fixes. This part still needs to be improved, but keep in mind that there are a lot of people working for Unity and introducing something as radical as that simply takes some time. I am sure that a lot of changes have been going on internally.
What I personally like is that they have introduced the experimental namespace to mark some new features as unstable. The good thing is, that we can already test them early which has the potential to end up being more stable at the end. And there are more signs that they are improving.
Yes, Unity needs to make a better job with the new release cycles at the moment, but I believe that once they feel more comfortable with it, we can profit a lot from it! I believe, it won't take much longer until the situation improves considerably.

 

(sorry to be that guy, but @Dantus, I think you're using "out of the question" wrong )

Also, I think it has little to do with the release cycle but more with the fact that Unity is doing some major changes at this time and is in a transitional period. There have been things that have been the same since... well 2.x and are now getting updated/changed.

 

No, that's not possible, I don't make mistakes
Thanks for pointing that out. I am obviously not a native speaker.

 

I see a lot of comparison with 4.6. As far as I remember, UT didn't add new things to late 4.x versions, except for their new GUI system. I believe they spent more time on fixes towards the end of the 4.x era and new features went into the 5.x branch. Which could explain why 4.6 feels more stable than early 5.x builds.

When I compare Unity 4.2 with 5.2, I believe 5.2 runs more stable than 4.2 back then. I see this as an improvement already.

I agree that it would be appreciated when UT puts more priority on fixing bugs, as this is what really helps us doing our work. It's easier to develop a game with known missing features in Unity, or any middle-ware if you want, than with unforeseen bugs. We can workaround missing things that we know upfront, but it's not going to happen to release a game when run into a showstopper bug.

But I think the patch release plan is a step into the right direction, as it gives us access to fixes early and this works for me pretty well.

I believe no matter what UT is doing, they can't make everyone happy. If they add new features, people might complain they should fix bugs instead. If they fix bugs only, other people might complain why nobody is working on new features.

I guess from UT's perspective it's not an easy call where customers have a higher benefit from and I'm not sure if the feature/issue-tracker votes can be taken seriously for that, as some groups are more oral than others. I remember a few features had so many votes which should be, from my point of view, irrelevant for developers that create games for a business.

For me it's fixing bugs that has a higher benefit, on the other hand, I can hardly wait for nested prefabs and there is already a conflict.


I agree that there are a few features in Unity 5 that are questionable why they found their way into the release. On the other hand, software development is often an iterative process and from my experience with Unity so far, history has shown that I can't count on a new feature right from its release. I give it a few iterations (minor version updates) before considering it to be production ready.

Of course it would be good if this is handled with the beta cycle and consider a new feature in a release robust. But that's not what I experienced so far and for me, as I know that's how it is, I can work with it.

I was thinking maybe it lowers our expectations when UT flags new features as experimental. This should serve the purpose to communicate that we shouldn't rely on it right now and if the feature matured, they remove that flag.


Anyhow, I believe Unity as a whole improved a lot over the years. I started with 3.5 and if I look back and forth, I think UT did a good job. There were times where a lot of swearing was involved because something didn't work, but I can't think of any middle-ware where this wouldn't be the case. When things got really ugly we used what was or maybe still is called the premium support.


UT is much more open than a few years ago. I feel like they consider feedback I provide during the beta phase, they respond on bug-reports, some of their developers even provide help in the forum (I guess they do this because they like what they do, not because it's their job), they provide patch releases every week, they fix showstopper bugs often in a timely manner, they write entertaining blog posts such as how Il2Cpp is working, they gave us a road-map and so on.

These are all things that evolved over the years which I don't want to miss anymore.

 

I actually think they can. What everyone is actually asking for is pretty universal: faster with less bugs (and possibly more learn tutorials).

The number one complaint I see (and I do see a lot being a mod) is either a bug, or the perception of a bug (fixed with more learn materials - which are rather excellent, just not enough of them). People like me tend to moan a bit more for optimisations or speed though.

 

My post was not intending to say anything remotely close to that, but I guess I can see how you came to that conclusion.

My post was intended to say that STABILITY has been lacking, and that no new features since 4.3 have been worth trade off of the decreased level of STABILITY that those features have brought since then (I do enjoy the new features, but most of them shipped knowingly half broken, and thus in my opinion, shouldn't have shipped in a main release until they were fixed.)

What I want is for new features to be isolated in a different branch for however long it takes until they work out all the major bugs in their new systems. The current "beta" and RC system is rushed by hard deadlines, and thus less effective than a full on public development branch. These new systems ship almost always universally broken in major ways, and usually not completely fixed even after being patched and repatched for months, or even several years in some cases! That is why I strongly believe that all these gigantic new systems should be sitting in a BLEEDING EDGE branch where they belong until they mature enough to be complete and solid solutions. Then, once the bugs are worked out and the new feature is mature enough to be known to be stable, then it should be moved to the STABLE branch that stays STABLE all the dang time so that we can ship stable games on stable, well tested, mature code. Did I mention stability?

I find it shortsided and not professional to include so much work-in-progress buggy code in the supposed "final release" branches! Unnacceptable is the word I am looking for. It means that every single release from now on until they quit this flawed model is going to be shipping with BLEEDING EDGE buggy code that I absolutely don't want in my supposedly stable release! UNACCEPTABLE! I have to jump to 4.6.x or really for my tastes 4.3.x to get a somewhat stable build, and it looks to me that those might be the closet thing I get to a stable Unity release for years! UNACCEPTABLE! Which means if I am right, then I need to start looking at Source2 or Unreal for the next crop of games we've got planned because the current state of Unity and their development model is UNACCEPTABLE! I'd rather stay with Unity, but I want a LONG TERM SUPPORT branch of stable code or I am definitely jumping ship as soon as our current Unity projects wrap up.

I feel like every build since Unity 5 has just been one long broken beta Please Unity team, keep the development pace going fast, keeping patching weekly! Both of those things are great, but keep your fast development pace away from the stable code, make a long term support branch (you know, a stable enterprise quality branch that developers can count on to be rock solid) and then pump out all this new stuff into a bleeding edge branch (basically your bleeding edge branch is what you ship as "stable releases" already, as I personally consider all the 5.x releases as bleeding edge and unstable.)

Look, I am for progress, but progress shouldn't come at the cost of stability! There are development models that allow for rapid improvements that would fit in nicely with Unity's current development philosophy, the only difference is separating the quality well tested code from new code that the team just wrote that hasn't been tested by hundreds of thousands of users in their unique use cases. I am posting my opinion here for the sake of continued progress done non-destructively! Continued fast-paced progress that doesn't introduce a myriad of new bugs each and every freaking release! Hopefully things improve, and I think they can, which is why I am posting! I really just want a rock solid Unity build with long term support and I'd be happy and then I would buy more Unity PRO licenses and invest heavily into increasing Unity based production pipelines! That's what I want! Right now I feel like a beta tester who is spending thousands of dollars on a moving goal post of a platform that has no foundation or stability, and that has me concerned that I backed the wrong horse.

Keep working hard Unity team! Thanks for your hard work!

 

Sorry, I simply got it wrong in that case. It was not my intention to confuse what you wrote.

 

They don't do that, though. Since the .x releases are now time-based, that means that any feature which is still work-in-progress/buggy gets delayed until the next release. The notion that they are pushing "bleeding edge" code into final releases seems to be based on a misunderstanding. Take a look at the roadmap page, and you can see that features are color-coded for on track/at risk/delayed. Anything that's not stable enough gets pushed back.

Bugs still get in, of course, like they always have, but the current release model isn't flawed, it's just different. Although personally I've found 5.x releases to be typically somewhat more stable than 4.x releases were. For example, I can minimize/maximize the game view without expecting it to crash every third or fourth time, which 4.x was prone to do for quite some time.

--Eric

 

I do see where you're coming from, and maybe the truth of the matter is somewhere in between my position and yours.

I know they say the philosophy is "any feature which is still a work-in-progress/buggy gets delayed until the next release" but the practice doesn't seem to be matching that from my perspective, as I see bugs get reported & acknowledged by staff during beta/RC right here in this forum, only to still ship with those bugs in 5.1.0f and 5.2.0f regardless. Perhaps my perception is wrong, but if that costs Unity my business, it's still a lost customer regardless of the accuracy of my perception. It's almost November and Unity 5.x still seems to have a beta level of stability to me. I still feel like I should only be basing games for commercial release off of 4.3.x or 4.6.x based off of the experience I have had developing with all the major releases in the 5.x branch (I do update to the patch releases generally.) Again, even if I am wrong with my assessment, this is the impression that Unity has left me with during the 5.x branch. That's not good.

Yes, bugs do still get in, tons of them. That's my gripe! I've had so many days of work where I end up finding a post with a Unity representative posting "wait until patch X for the fix" and then I have to switch tasks for a week or two, all because Unity rushed yet another broken system into their stable branch again and again and again and again. I'm sick of it! And no, I don't expect to move a game from one Unity version to the next, but I've yet to touch a 5.x build that's felt stable, so I feel like I got to keep trying the next 5.x release because hopefully it will have the patches I need and not be a giant crash or bug fest. It's frustrating that Unity hasn't ironed out the kinks in 5.x yet while it continues to add even more features. I know that's basically the plan, but it's a flawed plan that isn't working for at least some of Unity's customers like me.

I want Unity to fix bugs and not introduce new features into a branch for longer than a year so that it is actually stable and supported! I also want them to keep up this breakneck pace of development, but put it into it's own bleeding edge beta branch where all these kinks can be worked out for however long they really need (and lets be honest, several of the release systems should still be in beta) instead of pushing them out because they're "close" to done. Honestly, although I know it's not going to happen, I feel like Unity should just keep polishing 4.6.x and 5.x should've been the bleeding edge branch, because 5.0.0f, 5.1.0f, and 5.2.0f all felt like and performed like buggy betas to me due to the high number of crashes I got even on new projects with new scenes.

It seems to me that the Unity team is trying to do too much, too fast, for too many masters and then pack all that up on a regular schedule, briefly test it, and ram it out the door as a "stable" release to meet the roadmap. Yes, I see stuff get pushed back on the roadmap (good job!) but yet in the last year there have been "stable" releases with show-stopping bugs to lighting, lightmapping, GUI, and animation. Release after buggy release has eroded my faith in Unity 5.x and the development philosophy behind it.

I'll say it again, I believe the development philosophy is fundamentally flawed and is incapable of producing a stable release due to the introduction of new complex buggy systems each major release. Barring a development philosophy shift and a move towards a long-term stable branch, and a long term testing branch, I don't see this situation improving or becoming reliable enough for me to continue to try and base future commercial games on such an unpredictable and volatile codebase.

PLEASE UNITY, SLOW DOWN!

 

I think one of the issues is really the expectations of bug fixing. We're currently working on an VR game using minimum recommended machines.

We encounter different problems in different areas and whenever there is a patch release we carefully check what has been fixed. And it really got to the point that we even have to consider accepting a patch with an highly awaited bugfix because we might see regressions on other parts.

As an example, we're integrating the multiplayer part of our game. UNet has seen the light of day and we integrate it. Now, we initially used 5.1.X for our game and had great performance, no performance problems on the GFX side. Now, with 5.2 we have seen a regression in performance (according to Oculus HUD we went from +10% available to ... -50% in our test scene). But we needed some of the fixes in UNet...

So now we have to investigate where the performance hit comes from (WaitForGPU seems the culprit which didn't exist before)...

So, I can understand the frustration of my developers and their caution to even accept patch releases...

 

I'm sure this is something that Unity struggles with on a daily basis. They have to balance with supporting new platforms to remain relevant (like AppleTV, VR, etc.) or improving some of their previous design decisions (monobehaviour lifecycle, Input system, etc.) or fixing existing systems (GUI, Enlighten, etc.). This is a tough tightrope to walk with everyone at Unity having different priorities. They have serious competition right now and they need to remain both relevant and stable. So them dropping all new features for bug fixing isn't really a viable option. It works OK for some users trying to finish and ship your project, but they are constantly trying to acquire new users and grow. New users are much more interested in shiny features than they are with an object lifecycle system that actually makes sense (at least when they start).

 

Precisely.

 

how does that help?
what is the solution and not what is the general idea of a solution?
has anyone suggesting solutions run multiple teams all collaborating on one piece of software like unity?
like Unity?
ideas are easy, real solutions are hard

 

Well with all the people Unity employ they hardly will take advice of a few community members expressing concerns, they have people internally needing to sort these things, this is what they paid for!

 

Not at the same scale, but I actually have years of experience running multiple teams of programmers on a single codebase. It's not easy, and we went through several systems and development philosophy's until we found an efficient and profitable routine that got the job done "best."

No one is saying that Unity's job is easy, and I certainly don't have a hard and fast solution to fix all their issues. However, as a PRO customer who is either going to sink $xx,xxx in more PRO licenses or find a different platform to invest in, I need the Unity team to be aware of my gripes, so that there's an off-chance they fix it. I have outlined my proposed solution: a STABLE branch (no new features, only patches) and a BLEEDING EDGE branch (new features.) I recommend this out of direct experience managing large software teams and as a lifelong programmer and as a last cry for help before I begin moving on to more stable platforms.

 

yea i got nothing near that level of experience,
just hope the community can be positive constructive about suggestions and solutions

 

I have been developing games 16 years professionally, and I have work on game engine dev of similar capability. The main issue is Unity does not have a project in production, if they do, they will see all the issues we are facing immediately. The current QA, no however how much automated test, does not test as well as in a real world production. A few of the feedback I given to features/fixes, are just plainly ignored.

 

Well said - I created a thread in general discussion two weeks ago " do unity performance test there releases" some intresting replies. my main arguement was simple, have a project internally and upgrade it to each release and see the problems.. But of course was shot down. Unitu can show fancy QA presentstion at Unite but then why is it obvious things are worse.. Until the mood of the community improves, only then will general attitutes..

 

The counter-point is that they probably have too many people telling them what they should fix. If they worked on their own game, it may be even worse. They might prioritize the features and problems they ran into for that game rather than another issue in the engine. For example, they could develop a fast-paced 1st person PC shooter with a heavy emphasis on destructible environments. Or if they focused on making a unit heavy RTS they may focus on a hardware instancing of units. Either genre is something that their engine should support, but they would prioritize the game that they are trying to make leaving other areas of the engine lacking. The advantage of having so many users pushing the limits of their engine is that they actually get that feedback on all these various use cases without spending the time/money making the game themselves. I think that they can get feedback from dev teams of all types (from 20 person teams to solo developers) and wager what is important based on the feedback received. Unity wants to support all of them, and every team has issues that they want to see fixed.

Sorry if I'm white-knighting Unity here. The engine is far from perfect. However, I empathize with Unity's predicament. They are receiving feedback from all of these sources saying "my feature/bug X is more important than feature/bug Y" and they are often contradictory. It's up to Unity to gather up all of the feedback and try to prioritize as appropriate.

I personally think a better use of their time would be to spend more targeted developer time examining/interviewing use cases of their engine in a more in-depth manner. On an online forum and email feedback it's easy to get important features and issues drowned out in the noise. Get more real-world feedback and use-case experience directly from the users, and spend the time to really understand the problems being submitted. This would ensure that Unity developers understand more of the underlying problems that the users of their engine face. This experience could then be used to prioritize bug fixes and feature development.

 

There is no project that would cover Unity's whole range. Should it be for mobiles or high end computers? 2D or 3D? Single- or multiplayer? Physics driven? Be on the edge of what's doable? No matter which one would be chosen, the test coverage would be questionable and that's why I doubt that they would be facing those issues as well.
The QA has been improved significantly and it seems like an ongoing process. It wasn't unusual that bug reports were not considered at all, but my experience is that the situation is now a lot better.

At the moment, there is a lot of justified critique. And my impression is that this is being used by many people to blame Unity for random things that go wrong on their end.

 

No, we are not talking about features, we are talking about bug fixes. Some are pretty generic. Of course you might not see them if you are making flappy bird type of games but for other games that are more complex, the issues are pretty apparent.

 

Then it should be sitting in beta, not a stable release version.

Look, guys, I think we can all live with the occasional bug in a release version. None of us are perfect either and understand how development goes. Unity's case is exceptionally complicated and we can all appreciate that. Lord knows I write my own share of bugs.

But occasional bugs are not what I'm seeing. I'm running into something game breaking in nearly every version I try, usually stuff that anybody who would have spent an hour or two making a basic, experimental game as a test before releasing it should have immediately noticed. I.e., create two different GUI's in two scenes, have one scene load the other and watch what breaks. Basic stuff like that. Maybe hire a couple guys that do nothing but try like hell to break the system if you haven't already. I can't do it, I don't have time to make a repro for every bug I find, there are just too many, and I'm not usually going to bother when it's an older release anyway so as not to waste Unity employees' time with bugs they might have fixed already.

While I'm happy to be part of the beta as a Pro user, I didn't pay $1500 for the priviledge of being a tester for Unity's stable release versions. No, that needs to be done internally the same way I've always tested my software to make sure it's as bug free as possible before I release it to anyone. If it's too broken, don't release it. Just keep it in beta where it belongs until it's sorted out.

My most recent pain: I settled on 5.1.2p3 a couple weeks ago, but just realized tonight after spending the last two weeks writing and implementing a native DirectInput plugin so I can read more than 20 buttons on a controller (why Unity is limited to 20 I'll never know) so all the buttons and hats on a modern steering wheel will work, have force feedback,etc., that this version appears to be one released prior to the implementation of reflection probes in deferred mode. Oops.

Now I need to decide whether to skip reflections, go to forward rendering mode, or go on what was last time a day long adventure of trying all the various Unity versions again in hopes of finding one that works well enough. Say goodbye to any productivity on Friday. If I'm lucky there might be a patch or some version of Unity that works.

This is my biggest gripe at the moment: I'm struggling to even find ONE version that works on my project well enough to even release my product. I've probably looked through six to ten different versions of Unity in the last month and am not currently aware of one that works acceptably for my project. All the 0.7 Oculus stuff is way too slow. That should never have made it out of beta as far as I'm concerned, but whatever. I'm currently just hoping that I can find one that is ok. Hunting through versions for one that works well enough would be fine for beta releases, but I'm sorry to say that this is a lousy batting average for public release versions.

Develop new things, that's fine, but please don't release stuff until it's tested better than it currently is and you're pretty sure it works. Keep that stuff in beta so the folks that want to experiment with the latest whiz-bang project breaking stuff can do so, but the guys like me that currently value "it works like it should" above everything else are not driven crazy by every new patch that fixes one thing and breaks two more every week or whatever. I'm at the point now where I'm forced to try to find a new patch because some previous version that I thought was ok actually isn't.

Will I find a version tomorrow that works so I can release my product? I honestly don't know.

 

This *= 10.

 

Here is my perspective. I have been on this train as a professional user/customer for a looong time now (since 2006!).

I used to find Unity solid. It was perhaps a bit limited in the early days, but what it did, it did well and in a controlled manner.
I fully agree with the OP. Unity is no longer a stable, well-performing, or predictable platform.

As the popularity of Unity has grown (and the company itself) it feels as if each release has been an effort to add features, at the expense of stability and performance. Many features are not even valuable addons.

The race for more features come at the cost of stability and performance.
My pet peeve is SpeedTree and Terrain, neither of which work well. Last time Unity paid attention to nature or terrain it was in the form of the Unity Tree editor, which showed promise. But how long did that live until Unity decided to drop it in favor of SpeedTree?
Now we have SpeedTree (yay, another feature on the pre-sales website page!), but it is not production-ready. After what..a year almost?
The terrain engine itself has not been touched for years. Nor has aspects of the PhysX integration that have impact on terrains or large scale objects (collider mesh generation). Terrain is slated for R&D later on, but who knows when that will be.

My point is that rather than improving the existing functionality until it is impressive and efficient, feature after feature is left to die in favour of new features that are sort of half-implemented. And so the train accelerates towards certain doom.

Please do not alienate your users, least of all those of us who are forking over serious money for your software. Please ensure that the features you have work as well as they can. After that, and only then, should you add new features.

Thanks.

(And sorry for the rambling post. I am sleep-deprived after having worked around poorly functioning "features" for the last few weeks....)

 

Totally agree.
Also, more times than desired, after upgrading a complex project to a new Unity version, you get poorer performance.
I remember upgrading from 3.4 to 3.5 and from 4.2 to 4.3...

 

4.6 = 'Unity Mobile'.

Can we just accept that and get extended support for 4.6.x?

Currently working on a project that's relying very heavily on baked lighting, and moving to 5.x simply isn't viable. Lightprobes don't work, and even if everything did work, it'd take maybe a month just to relight+rebake all the levels. And the results may well be inferior to 4.6.x.

And as well as problems caused by the loss of Beast, I'm seeing lots of moaning about UI breakages, performance regressions, splashscreens, lightmapping issues, and lots more. In my brief experiments with it, the editor has seemed crashier than 4.x, and I've run into other issues with some versions, e.g. broken VS integration.

Does 5.x have anything to offer to serious mobile developers other than pain? (Support for Metal and IL2CPP seemed like they were supposed to be it's mobile selling points, but we got those in 4.6)

 

What extended support, they not planning to put in any new features into this build for sure.
All the new mobile stuff, hence forcing one to upgrade to 5 in due course.
Also look at the 4.6 last few patches, nothing there - slowely given up on it.
As much as I like the idea myself, we all got to move on, just want to move onto something more mobile friendly and that feels like 4.6

 

I'm enjoying the stuff in the beta so far, Unity keep up the good work..
Keep that steam train a' chuggin' and turn it into a maglev on the go

 

I just saw that Unity 5.3 is now in a release candidate state. Isn't that a bit early? I still see a lot of ongoing issues and perilous last minute changes in the API that make me feel that this update is also going to be problematic.

 

They are sticking to the roadmap, which has fixed release dates from the day it was announced, so it's time for a RC.

Edit: I guess they are going with the fixed updates/release dates model that Apple is also doing (yearly updates for iOS and OSX no matter if they are ready), which has really messed up OSX, it has really gone downhill ever since they switched to that.

 

We're constantly re-assessing and figuring out what's ready to ship or not.

Thus far, SSRR is being pulled for having edge cases we'd like to still address and OpenGL Core update is being considered for being to default back to the Legacy renderer for final. The graphical/OSX issues are mostly related to the OpenGL bits but we're trying to see if it still can be shaped up for release. We think we're quite close. Worst case is we go back to what we have (legacy) and allow folk the option of switching to the updated core.

That leaves mainly the multi-scene editing work which is still improving APIs based on feedback. We feel this is a pretty important improvement to people's workflow, even with late API changes that it's worth while to have.

Most everything else is in a rather good position that we know of that's been reported for 5.3. We are down to a pretty low number of reported bugs we've deemed shipstoppers that we think we can wrap up in this RC period. There is always the possibility there are edge cases or areas we missed in our QA (be it process or automated tests), so unless it's been reported we won't have seen it so make sure to raise the flag if so!

Let's not forget all the items that we're are not seeing significant issues in that are part of 5.3. We're aiming to get all of these into people's hands.

 

Ok that's a good sign, but what matters the most to me is the ratio of reported bugs / fixed bugs. Is this ratio good enough right now?
I think that the small and frequent edge cases are as important as the big shipstopper bugs and I really hope you do the same. It's those frequent small bugs and quirks that sometimes drives me crazy when I use Unity. Please take time to consider those "edge cases" that play a huge part in the perception of Unity's stability.

 

Let me rephrase: "pretty low number of not-yet-fixed reported shipstopper bugs" Most have been fixed, some small portion deferred to a patch as those bugs we'd still like to address for this round, and some kicked out to 5.4 (surprisingly not that many.) Internally we've been in fairly serious bug fix mode for the past year and each release has been passing fewer and fewer bugs to the next release. (Though, that's not to say some key bugs crop up that create serious pains)

We do try to take this into account. In fact, we'll still consider them shipstoppers in many cases. However, being human, there's always a chance of error of the assessment. But please to recognize we do watch for these as much as we can.

 

Yeah, I think I am done with Unity PRO. I've got multiple projects in Unity PRO that will take at least another year to wrap up, but Unity's development is an unstable mess with no solution in sight. Maybe by the time these projects have wrapped up I'll have changed my mind, but I don't see any reason for hope.

 

Ok, as the beta of 5.3 was still a bit chaotic for my project (and 5.2 unusable) I've decided to stay with 5.1 for the moment and release the game with it. After that I'll spend more time upgrading the project to newer Unity version and I hope I'm going to be surprised in a good way by Unity's stability's improvement.

 

I agree stability is much more important than features. Just look at the release notes for "5.3 Release Candidate:"

• Crash previewing (playing) a Mecanim Transition in the Inspector preview window. Fix expected in next release.
• Multi-scene editing: Unity crashes when using the GetSceneSetup API of the EditorSceneManager class. Fix expected in next release.

I just tried installing the RC and got a "Unity.exe not found". In what universe is this considered a candidate for release?

I understand, you now have a fixed ship schedule, but all you're doing is killing your brand. The reason people choose Unity is because it's easy to use and it works. Instead you're choosing to release things prematurely that are prone to change and don't work (e.g. the lighting fiasco, and it looks like Multi-Scene Editing is going to be along the same lines judging by its current state).

You have the advantage on Unreal Engine, don't give it up by releasing broken software and spending months to patch up the issues.

 

Seriously? Being deeply buried in work I have not checked the RC, but "fix expected in next release"? What the h*ll is that? Have Unity stooped so low as to release versions with fully known crash bugs in it? In what universe is that OK? We would be (rightfully) crucified by our customers if we did that with a game update release. Hell, they crucify us when we release them with *unknown* bugs, which happens sometimes.

I am really starting to lose faith here....

 

Perhaps terminology isn't clear here. Next release == RC2 (5.3.0f2).
Only one of the items explicitly says we're fixing unfortunately in 5.4.

edit: Will go edit release notes to be more clear.
Realizing release vs. release candidate wasn't obvious.

 

Well, better. But one thing then (that is known) will be fixed a full point release later? To say that I am surprised and disappointed is a vast understatement. I hope it is not a feature we are using (have not checked the notes, starting to feel I have to. Who knows what else is lurking in there... :-/ )