"Unity is broken" - What are people referring to?

• Random freezing of the editor, requiring ending the process. Rarely go 1-2 hours without this happening.
• SelectionBase doesn't work properly, and is inconsistent.
• WASD + RMB stuttering.
• Physics.Processing spikes for no reason, even in empty scenes. (Showstopper).
• Many of these basic issues have been in the very first version of 5. No sign of them fixing them.

 

As none of that happens on any of our test machines, best you can do is report a bug since other users aren't getting the same problems... Maybe Unity is unaware they exist or it's project specific?

 

This kind of attitude isn't helping anybody.

That does not match my experience. I have never received a response to a bug report and has never seen a problem I reported being fixed.

That's why i HATE semantic arguments.
Unusable in which context? For which purpose?
Shall we waste eternity determining the One True Way of evaluating usability of the engine for the purposese of determining its brokenness?

The gi random tint glitch I linked earlier will make unity unusable for the purpose of constructing modular scenes with global illumination. Thus, making it "broken".

Encountering a glitch in mecanim (happened a lot during 5.2 update) once again, will make it unusable for the purposes of continuing the project with mecanim enabled. Again, broken.

People complain routinely (and have a good point) about enlighten being overly resource-hungry and unusable. That makes enlighten integration broken as well.

Anyone who says "broken", is very likely encountered a showstopper bug that hurt their project. Is it THAT hard to understand? Just because it still works for you, doesn't mean it works for them. Matter of perspective, you know? Different viewpoint, different projects, different needs? Circumstances? You are a human, so you shouldn't need that stuff explained to begin with.

Seriously, though, just go through the forums and looks for threads with complaints in them.

 

I still had the editor open from a test I did today, disabled my only 2 objects that had scripts on them and let it run:

18ms physics spike.

I created a new empty project, no evidence of spike, but it comes with vsync enabled by default, which might hide it maybe? I disabled vsync and saw a spike again.

Only tested in the editor, not sure if it would be noticable in a real game. Are we even sure that the profiler doesn't have bugs? Who profiles the profilers?

 

Unity is a game development engine, right? Well "broken" is slang for "op" which means over powered. People might be suggested that unity is too good.

 

Okey, i really want to enlighten ppl that have problems with Unity, the issues are small compered to other engines, Unreal engine example tend to crash due to improper memory pointers (Unity is based on managed code and doesn't really have memory leaks). Unity 5 is fast, stable great products. Tough i would like to see a newer versions that has improved Canvas GUI performance, i can live with 5.1.2 for some time, its fine.. It takes some time and understanding on how managed memory works and how you lay out of codes. Things can get very unstable and slow if you do something really wrong, yet it will work. On other engine it would probably not run at all. Pretty big difference, and other engines has more or less "blueprints" of the code structure making the games some what similar. Graphic render in Unity 5 is sharper and more precise.

But i have to agree with some here, that its better to narrow down supported platform if that can improve the most popular platforms. We all want quality and performance.

 

It is to good

 

Enlighten is most definitely too good.

 

This...

 

Broken perception ==> Brokenception.

 

Edit :

Broken adj (not working) : In game development, used to refer to a promised thing (feature) that does not work 'out of the box'

 

Hmm... I am apparently not seeing some part of this discussion.

At any rate, I agree. If Unity were "broken" there would be a lot of businesses in a lot of trouble. Ours included. Sure there are a few bugs here and there, but nothing approaching a show stopper (and this is across many projects). And we really aren't even doing much that would be considered work arounds. One thing I have noticed is that 5 (over 4) is a little less forgiving in terms of sloppy or awkward set ups. We had to to redo many of our mechanim setups during the upgrade to 5, but most were because we didn't set it up in the "proper" way in the first place. Some of the patch versions of 5.2 were shaky, but 5.3 has been solid.

Something I think that would help out, is if the docs were updated, cleaned up, and more information about best practices were shared. (I know Unity has some docs due out soon that should help out). For example, using resources folders can cause a lot of problems, many that are seen/reported in some of the other support sub forums. Unity's own best practice recommendations for resource folders in production is "Don't use them.", and lists several problems that can arise, and why they should not be used except in very rare cases. However the current manual and docs don't reflect this. There was some other issues that came up with incorrect docs related to shaders and such. I think these kind of things can lead to users being frustrated or perceiving bugs, when it is really a lack of info/poor docs. I know that there have been a few cases where we have run into problems, and were able to resolve them by using ILSpy/MD to look at what the actual methods do, instead of relying on what the docs say they do. (Handles is a particularly poorly/incorrectly documented area).

I feel fortunate, in having the ability to directly ping Unity devs, and even sit down and go over things with them. But even so, often times it does turn out some of our questions or issues are right there in engine, just not very well documented.

 

(slightly off topic, but why shouldn't Resource folder be used? And what replaces it? Asset Bundles? Why are those better (and weren't they originally meant for downloading assets from the internet?))

 

It gives you less control over memory management, increases startup time and build times, doesn't really allow for platform specific content/management, there's the potential for duplication and unused assets. Since it reindexed every time you hit play, crashing can break links, especially if you mix in prefabs and such. It was intended for quick iteration of dynamic content during development, but assetbundles for production. The new bundling system allows removes that need. Assetbundles are main way for asset management. For content not in bundles, direct references to the asset.

 

Unity has some really great docs coming at some point, that cover various topics like asset management, optimization and other things.

 

I'll believe it when I see it.

(or... more accurately, I strongly believe that "some point" will be when they have changed the systems already and the docs will be outdated)

 

I can't wait for the asset bundle change. @zombiegorilla wouldit be possible to write a script, compile it, then import the assembly and use it in a game without restarting?

 

It's probably worth noting that just because you haven't experienced issues, that doesn't mean they don't exist. There are a lot of arguments here expressing anecdotal evidence, and that is 100% useless when considering a game engine as robust as Unity.

We're all designing very different games. Some people are going the "retro" route with 2D side-scrollers, others are making FPS games, some are making racing games, RPG's, etc. All of these genres are going to make use of different components that Unity offers. So you can't really use your personal experience with Unity to say that there aren't showstopping issues.

So you didn't encounter any bugs making your 8-bit side scroller? That's fantastic, I'm very happy for you. Meanwhile, over here in 3D world, making an RPG that uses just about every system Unity has to offer, I can tell you there are definitely large issues, some of which were showstoppers not too long ago.

With Unity 5.2, Mecanim had an issue where you couldn't smoothly transition from an upper layer animation to the base layer animation. It would just *pop* into the new position and rotation. Sure, that may not seem "broken" to you, but if that person was releasing the game right then, they're going to get some flack for crappy animations. I still remember at least three specific instances where someone claimed they had been unable to release on time due to issues with Mecanim in 5.2. Sure, works fine now, which is totally cool if you want to flip the bird to any customers who actually have to adhere to deadlines.

I really wish people would harden-up and deal with the fact that there are certain things that just shouldn't be tolerated. In what other industry is it okay to say "Meh, it's not that badly broken?" I can tell you right now if you tried to do that with any other product, your customers would laugh you out of the room (or wherever you happen to be).

Sorry, I know this is a violently unpopular opinion among other vets, but it really does get my knackers in a twist when I see things like "Well I don't have any problems. Stop complaining!" It's not logical, it's not productive, and it really pisses people off who are having legitimate issues.

 

And the flip side: just because you have experienced an issue, doesn't mean it affects everybody. Sometimes it's an edge case that only happens in certain circumstances (which doesn't mean it shouldn't be fixed, but overstating the issue doesn't help either). Sometimes it's user error, and occasionally it's not even a Unity bug; I've seen a number of topics end with "it was my AV causing the issue". Of course, sometimes a bug is actually just a plain ol' bug that does affect many users. Also, nobody ever said what you're claiming, about "I have no problems so stop complaining". Again. Aren't we all sick of straw men yet?

--Eric

 

A simile using "like" or "as" is used to note that the following statement isn't meant to be taken literally. Or, to turn the tables back to you, I never said anyone said that. Two straw men don't cancel each other out. I never said screeching "broken" is an appropriate response, nor did I say that just because someone does experience issues it affects everybody.

 

They know. And it's a showstopper because it also appears in builds.

Feb 2015 - https://issuetracker.unity3d.com/is...cs-spikes-in-profiler-on-some-android-devices
Jul 2015 - https://issuetracker.unity3d.com/issues/physics-spikes-on-android-cause-frame-rate-drops

Here's a picture of it at its worst:


And why is there an unsearchable Fogzbugz website, and a separate Issuetracker website? My submitted bugs are on Fogzbugz but don't appear on the issue tracker.

I'm assuming your test machines have a range of hardware specs with up to and out of date drivers?

 

Unity 5.2 : The Quickening.

 

Yes you did:

You are claiming that you see people say things like that. Those are your words right there! When I counter that nobody actually says that, of course I'm not talking about the specific exact words, that's just silly. But nobody says anything "like" that, either. Hence, straw man.

I really shouldn't be having to explain this...I thought most of us were programmers, who used logic. I know it's frustrating when you run into problems, but going down this path doesn't do much to solve them.

If not successfully reproduced yet, they aren't on the issue tracker.

--Eric

 

Arguing for the sake of it gets nobody nowhere, I couldn't care in the slightest if anyone else is having a specific issue. It's not their game I'm trying to make, secondly I find the method of "voting" on bugs frankly ridiculous.. If my problem doesn't get upvoted or fixed by our targeted six month release date, because it's not "popular" or Unity just wrote off a bug as (non re-producable) then where does that leave someone?

I've had to bounce back and forth trying to roughly point them in the right direction, which at return rate can take forever. I'm more than happy to do the "ground work" to fix an issue, I had problems "cooking" packages in UE spent near enough a week trawling through source to find dependancy issues in the project / engine framework, found the issue / fixed it and submitted it to Epic through UDN.. They gave feedback "instantly may I add", giving some pointers on how to improve / automate my findings.

You can argue semantics all you like, but my definition of "broken", your definition of broken and / or the Oxford dictionary definition of "broken" is frankly irrelevant. When a customer says he has an issue with phsyx for e.g. and I can re-produce said issue. I could argue with customers, no it's not broken / just has a bug.. But it's a complete waste of time, it has to get fixed irrelevant of your definition. If a customer says it's broken, it's broken...

Then saying, ok guys it'll get fixed when the engine manufacturer can re-produce it and my bug can be upvoted. Check back in six months or so.. Yeah, that'll go down a treat.

Personally, for any commercial based project I wouldn't touch Unity without a decent support contract.

 

I was providing an example of the type of argument I tend to see.

It's classy stuff like this that makes the community feel appreciated.

 

Hell, most to all of UnityEditor is poorly documented with what amount to boilerplate descriptions, and the closest you get to best practices are the unicorn examples that use Undo or SerializedObject/Property. Then I wonder how many people blame the editor for exploding and vomiting color in their face every ten minutes, when it's some crappy extension they won't get rid of.

 

UnityEditor documentation... sigh. Yea, there are times I think it would almost be better if it were just not documented. Handles is a great example, is a mishmash of old, broken and strangely pointless examples. Incomplete or missing documentation. Often random guessing or using ilspy/md is much faster than the docs.

UnityEditor is powerful and a huge strength of unity as a whole. I just wish it was better (or at least correctly) documented.

 

I like the post because I think every engine could do with better documentation, but credit where credit is due. There API / example documentation is still one of the best I've seen. The tutorials for Unity rock, very simple / effective and easy to understand.

 

No doubt. In general unity has pretty good docs and examples. I just think that for whatever reason UnityEditor docs have been left twisting in the wind. Maybe just because it is a more advanced area, and it is assumed that folks delving into this area are capable of foraging for themselves, or maybe it is just an area that doesn't update as often as the core API. This one for example got a good laugh from our Unity rep.

 

LMAO "and stuff".. Mighty descriptive there haha!.

 

Well, for me it was broken...I had lots of issues as they changed how physx was implemented in Unity 5. I reported many bugs and eventually I was given a weird workaround with a promise of a future solution. I changed my game behaviour to match it but then they changed theirs again and broke it again... By broken, I mean that my gameplay was literally destroyed as our gameplay relied on physx.

I think all that is because :

1) They usually move on to new features without perfecting the old ones due to lack of time.

2) its difficult for them to account for the different ways a person will use a feature.

3)Because of 1) and 2) when a system needs upgrade (for instance the physx system), it often gets broken heavily -meaning the newer system might have more features but might not work the same way as before, which means will have different restrictions, and they have to compromise. If you were using the physx one way they did not account for, then it could become broken for you and in need of fixing.

Its all about perspective. In the end it depends what you do. If you use it generically in ways most people do, then you can usually avoid the issue of it being "broken" and needing fixing. If you do stuff they don't account for, then when there is an update, everything can go to the waste.

 

If it ain't reported broke, don't fix it...

 

I don't think that's the problem. There are plenty of reports for multiple issues and the solutions and fixes keep getting postponed.

Whatever Unity's issue is right now, it's not that they don't enough bug reports.

And if really their system is made in a way that they can't lift a finger if there isn't a relevant bug report there, then there's something wrong with their system.

 

@AcidArrow is right - unity needs all the bug reports they can get, they don't want to ignore your issues.

 

I agree that they usually fix most bugs. I'm a Unity fan and I want it to be good but in our case they just could not fix that one.

There is another bug also that I reported, and that one was less critical, but the workaround was really stupid requiring to do extra calculations for nothing. Others complained about it and the unity dev first said that its not possible to fix because its too much work... That was not very professional but then, after other people joined in, he said that he changed his mind and would try to figure out the best solution to fix it, but it was eventually forgotten. That one wasn't that critical for me though so I dropped it. Sorry if it sounds so terrible as that was my bad experience with the bugs. On the bright side, I reported maybe a dozen others that did get fixed.

 

There's still basic issues like the GUI not updating until you activate/deactivate an object, or press play.

 

As Spock would say :"The needs of the many outweigh the needs of the few."

 

I don't know guise, for me Unity is da beez neez!!

Hope any issues you're having with Unity get solved fast. For me coding anything is so difficult, I can only admire Unity, on my side it's pretty darn stable and issue free, luckily, so I'm grateful.

 

Its some peoples moral and attitude that is broken.. Games i have made in Unity are darn more stable than most games out there made with other engines. Nothing is broken here, move along.

 

I'm glad this is your experience. Ours is vastly different. Constant crashes, bugs, issues. That's our consensus at work. We are trying to do something vastly unique and different and it's really struggling to keep up. Sure, you can blame it on our user error, and some of it surely is. But that's heavily outweighed by the errors on Unity's part.

 

Unique stuff will stress any engine though. Try for example, getting toon shaders working in UE4.

 

That's an ill thought out, flipant statement based on zero context. Some games will never use a terrain system, some will never use mecanim, some will never use Umbra and so on / so forth. The higher the scope of a project, the more dependancies you'll come across therefore technically what you're saying doesn't mean a thing.

There's no such thing as a completely "stable" engine, simply doesn't exist. So if it's not completely "stable" there has to be bugs, if there are bugs customers can complain "it's broken".

It's not a matter or "broken" or "bugs" unless it's out of control, what matters is what gets done about it. In a version of 4.X, there were CTD issues when removing trees from the terrain system. Simple example, which in no way can be blamed on a games developer. I checked back a while after to see if it had been fixed, it hadn't therefore the terrain system was "broken" and is it's a major component in an RPG game for e.g. that would of been a showstopper.

Again, having bugs / broken systems isn't an issue. It's what gets done about it..

 

Have UT ever officially commited to sticking long-term with enlighten? If they internally are already integrating another light baking/GI engine, that would explain why the current solution seems to keep being broken. Just a random thought, I have no idea how feasibly it would be to switch a light baking/GI engine. Might be a very unrealistic idea.

 

I never had any issue with Terran system in Unity, sure there are many things that needs improvement, but it ain't broken just because of that.

Calling things broken is by definition "Jump to conclusion"

 

AFAICT Unity are invested in Enlighten for the long term, with a view to improving it constantly. I don't know if they plan anything in addition to it but given they've blown a ton of money and engineering effort on it so far, I don't think it's getting chucked tomorrow.

 

Claiming that it isn't broken because you haven't encountered problems... and you say others are jumping to conclusions.

 

encountered problems != broken

 

It's broken for the purpose they want it to fulfill. That's the definition of broken being discussed in this thread.

 

There's a difference between struggling to get something to look right and having it crash or encounter massive bugs that cripple even your ability to work.

Our gripe with Unity isn't that it doesn't support what we are trying to do, it's that some core elements break down or bug out too easily. While we have a huge procedural element to our game along with a huge simulation, there is no reason for constant memory leaks or crashing coming from Unity's internals.

Old outdated systems that don't allow much flexibility is our other gripe, but that's not what we are directly talking about in this thread even if it may be considered "broken". Input system without remapping for 7 years?? That's a broken system in my book because it's practically unusable for 90% of the demographic without introducing a considerable layer on top of it. This basic of a request is unexplainable of why it didn't happen almost a decade ago. I make the argument that Unity is broken because of the inability to respond to such a basic user request more so than the shortcomings of the actual system.

 

Well, it's not quite "stop complaining", but it sure is damn near close to the mark. Maybe now, some of us can actually deal with the arguments being presented instead of name-dropping logical fallacies like first-year Sociology students? I know that's salty, but it's just something I don't take kindly to.

No one is saying it's broken "just because of that." And, in fact, no one is saying Unity as a whole is broken. I'm pretty sure everyone here can agree that Unity is a fantastic game engine. What people are saying, is that there have been issues in the recent past that have lead to massive problems for developers.

I think certain people are glossing over the arguments because they interpret "Unity has issues that should be addressed" as "Unity sucks and so does everyone who likes it!"