Unity 5.2

So, what's next guys?

Sometime ago you commented something like publishing a roadmap. Have you changed mind?

Anyway, are you going to open a new forum group for the next release, like you did with 5.1 beta?

 

I think there is a small roadmap session next week at Unite so maybe they will put something on the web after that (finally). If not, then it's starting be as big joke as the UI was and we need a roadmap for the roadmap

 

I would peg 5.2 release around unite US, or maybe the beta at least. So far features that have been mentioned and are likely candidates have been hinted at in various tweets and forum posts.
A new UI rendering batch algorithm for improved UI rendering performance. A big issue since the UI was released.
They teased some improved 2d toolset features. Including a tile map terrain builder.
Aras was also working on some CPU rendering improvements (somewhere around 10% savings). I am not sure if those made it into 5.1.

I figure a roadmap for them is hard because there is so much R&D going on and the time might be reappropriated based on ever-changing priorities. Which means there will be a lot of broken promises going beyond a patch or two into the future. Having a rigid roadmap is dangerous in the fast-paced area of technology where support for the latest rendering API or the latest piece of hardware might need to take priority over another feature that patch.

 

Yes, I completely agree that publishing a roadmap with a reference timetable could be dangerous. But at least a prioritized list of demanded features, like SSR and many others so we can stop requesting over and over the same things.

I would say that WebGL should be prioritized, so the "preview" label can be removed ASAP.

 

The requesting will not stop with a roadmap in place, that's for sure. People still would have the possibility to complain about the priorities and if something goes wrong they will criticize that and will ask why feature X is already done which had a lower priority than feature Y which is still missing, though it is without question the most important feature ever. The amount of requests would likely stay, but the discussion would shift into a slightly different direction.

Regarding WebGL, it already seems to have a high priority as it is frequently mentioned in the release notes. Removing the preview most likely depends more on the browser vendors and other external forces than Unity itself.

 

WebGL preview tag isn't going away for quite some time AFAIK. It's not only about Unity priorities but also about browser vendors, emscripten team etc.

For roadmap there could at least be a list what is being worked on without timeframes or with rough ones. Now the info is only in random blog post, dev posts on the forum etc.

 

Someone should write a bot that collects posts from people with the black unity tag, could give lots of hints about whats being worked on without searching for it.

 

Yeah some forum system have own dev tracker threads and/or all threads with dev answers marked with custom symbol.

I have most devs followed here with the forum feature so it provides good stalker feed too

 

A dev post tracker for the forums is actually a pretty good idea! Although I'd want to ignore some dev's postings, like all of the lockable topics that Aurore and Andy Touch needs to post in for example.

 

I already have created the "isthenewguioutyet.com" and I swear... if they're not publishing the roadmap, I'll grab the "istheroadmapoutyet.com"...

 

Well why not use dynamic task system, where the most wanted features are tracked and colour coded to indicate their state and with what release they will arrive. And as it's a dynamic system so nothing is fixed in stone.

 

Let's hope that cool temporal antialiasing will be possible in 5.2.

 

Lets hope some of the bugs the community is complaining about are fixed in 5.2, no new features even needed.

UI issues priority + Performance fixes are a must for mobile development 4x + 5x
Android related issues to low FPS 5x + BlackScreen reports on 5.1
Mostly mobile performance fixes in 5x builds + common bugs.

 

Definitely need features...

1. Option for 32-bit vertex indices.
2. Option to cap mip-map levels.
3. GI that can work on procedurally generated terrain.
4. Metal support on OS X.
5. IL2CPP on desktops.
6. Better multithreading support.

I think that sums up my list of wishes. Granted, maybe those aren't needed in 5.2. Hopefully soon though!

 

1)Okay I agreed with Metal Support on OSX and Retina Support - both confirmed to be coming (doubt in 5.2)

2)What do you mean better multithreading support? For mobile development I see this option but currently no one has been able to confirm what exactly happens when this tick box is checked (in my experiance nothing)

3)IL2CPP Desktop, its taking them long time to even finish iOS but progress is being made each week, so I think this would take some time, maybe months. I not sure if they plan to do Android next?

 

I find myself quite annoyed at the "so and so function can only be called on the main thread" even when I'm doing perfectly thread safe things that (at least in my mind) have no possible way of messing up any internal systems, and then always having to work around it by using slower and more obnoxious code.

 

Can i haz dis please? O_O

 

Would you mind providing an example?
I think it's pretty hard to say that something shouldn't cause problems without knowing how Unity manages the underlying datastructures.

 

Here's the roadmap: http://forum.unity3d.com/threads/whats-unity-roadmap-look-like.326144/#post-2115168

 

That's true, I may very well be wrong. I just feel like doing something such as reading a position from a transform shouldn't cause problems. I could imagine problems if I'm changing a transform from a background thread, but reading a position from it?

 

When you have multi threaded code, you can access the position at any time, e.g. while the physics is being updated. This would most likely lead to awkward results. In order to get correct results at all time, it is required that Unity makes it thread safe. And as long as that is not the case, it is better to block threads to access critical parts.

 

I don't know ? Where is the roadmap ?

Unreal 4 delivers big releases each time, with big features, i would like Unity to do the same.
Let's dream , bringing a lot better terrain, better vegetation (to avoid purchasing stuff on the store), better full screen effects like better Eye adaptation or DOF, a shader editor also, a cinematic tool.

 

I don't get why it takes months to make a list what each team/person at unity is working on and how many months it'll most likely take to be finished

 

@Schubkraft Which Unite? ...Amsterdam or Boston?

 

Amsterdam next week.

"See into the future at the roadmap session"

 

I thaught new UI after 4.6 was great for PC and mobile ?

 

No, the uGUI is less performant than many 3rd-party GUI utilities that were already out there by a significant enough margin for many people to have to revert back to using 3rd-party plugins. While it's better than using immediate mode GUI (the old UI system), it isn't as fast as many people wanted/expected.

 

DX 12 support hopefully, Windows 10 is coming out in a month.

 

Because stuff changes all the time. Fixing bugs is pretty much what the majority of customers want. Hardware vendors want us to support their undisclosed devices. Something like WebGL needs infinite resource so people get moved around. It's trivial to come up with a plan of what everyone *wants* to be working on. We know a roadmap is something that many customers want to see, and I am sure once Unite Amsterdam has arrived there'll be news to tell.

 

I am absolutely no expert on this but when I look at Sketchfab I can see that their WebGL implementation DOES already work on the common browsers, even mobile.

 

+1000.

 

Perhaps a little transparency about how the decision works would be fine.

 

May I point out that Sketchfab is only a 3d asset viewer? I don't see it running physics, custom UIs, animations or any of the complex dynamics in any game. WebGL is still very much in development across the industry. Unity is working with the browser vendors to be pushing the boundary, but that doesn't mean automatic cross-platform support from the get go as Graham implied.

See latest blog post about improving WebGL: http://blogs.unity3d.com/2015/06/18/webgl-webassembly-and-feature-roadmap/

 

I was also about to say that Sketchfab isn't the best example since it's doing only stuff in the area that WebGL was designed for

 

Well, my project neither have any of those extras. So, does it mean that every WebGL project will be heavy-weight due to all those features? I'm sure that you didn't meant that. I think you refer to the fact you'll are too busy working on so many fronts.

Take this example:

It loads almost instantly and those reflections on the metals looks quite good. Can we make a simple asset browser like that in Unity as of today ?

 

I think Unity projects will always have at least the core engine included which means you will most likely never be able to do as light weighted simple asset viewer style projects like Sketchfab is with it. Also Sketchfab is just normal minified javascript and coded strictly for one purpose only while Unity is generic and asm.js style code which still isn't the bestly supported in browsers (load time, mem load).

 

But you already have a lightweight WebGL modelviewer - it's called "Sketchfab".

While I also look forward to improvements on the WebGL sector it's probably really difficult to strike a ballance and I understand that. Unity is super complex yet still tries to push for WegGL further and further. It's still a rather new technology in regards of expanding boundaries. I am sure things will evolve and get better but at the moment if you need a lightweigt, good looking fast modelviewer: It's Sketchfab.

 

No, we do stripping to remove all the pieces that a project doesn't use (or as best as possible. Work still being done).
I'm assuming your project uses _something_ more than just the renderer as other wise Unity is just a model viewer for you.

Why use Unity for that? As you point out Sketchfab is already there. Unity is focused on game development and making that accessible. If it's missing game elements, then we aren't doing our jobs. I guess the point is we're coming the other way, working on making games usable and playable on WebGL. A viewer would just be a benefit from that.

 

This kind of examples are great at least as an optimisation reference.

I work on visualization projects. So the biggest differentiation factor for me is rendering speed and visuals. I really don't care too much about everything else, I mean, I need networking, UI and lot of other features, but if the resulting product is a brick then I will have to choose another approach and probably leave Unity for fat clients.

 

@Thrawn75 BTW here is something to read about in regards of WebGL and the preview tag and such

http://forum.unity3d.com/threads/webgl-roadmap.334408/

http://blogs.unity3d.com/2015/06/18/webgl-webassembly-and-feature-roadmap/

 

Of course I knew that comparing Sketchfab with Unity is dubious, I´m just happy that this discussion has started.

 

YES! YES! YES!

Thanks for sharing!

 

I am confused. Why are you happy this discussion has started? I couldn't find anything new that wasn't already mentioned.

 

Hope SSR will come soon!

 

Yeah.

 

Also point out that SketchFab doesn't need WebAudio, which also makes it a poor comparison.

 

*cough* Terrain *cough*

 

WebGL Roadmap over here -> http://blogs.unity3d.com/2015/06/18/webgl-webassembly-and-feature-roadmap/