Unity 5.0 - A shot in the dark

I really like extrapolating and speculations. This should not invite to beg for X happening at a certain date but to guess when or if it will come. Some facts:

3.0 - September 2010
3.5 - February 2012
4.0 - November 2012
4.1 - March 2013
4.2 - August 2013
4.3 - November 2013

2D tools were a big suprise and ranges in the same category as our DNFGUI imo, which heavily relies on them. So I expect it to come together with nested prefabs and events who were demonstrated all together at Unite. If they don't drag it out even longer [...] 4.4 would be a well fitting release around March 2014.

No idea what they might put on a 4.5, fixing the FBX-Importer maybe, decal system, node shader editor etc...

With 4.0 they changed their release model to include all ready features at this time to shorten the cycles. So maybe there will be a 4.6 this time.

What will likely be limited to the 4.x cycle:
- Flash (was abandonened)
- Webplayer in its current form (Chrome drops NPAPI)

What's unlikely going to be implemented during 4.x, means it has to wait at least until 5.x:
- PhysX 3.x upgrade (webplayer backwards compability)
- Mono upgrade (Xamarin licensing)
- 64 bit editor (broken dependencies everywhere)
- Linux editor (well...)

I'd be very curious if they give up on the webplayer completely for 5.x or not. In favor of HTML5 (or any other independent technology that might pop up) they'd only need to keep the player available to download for legacy exports. The other option is to upgrade him and put massive effort in keeping backwards compability somehow. That being said, new plugins having a tough going lately *cough* flash *cough*.

But till 5.0 in March 2015 there's much time to change the circumstances

 

+1

 

Under the speculation banner;

Seeing that Google gave Unity a 16 month delay to implement a different solution, meaning December 2014, I hope to see that solution implemented before 5.x. I think that being forced to pay for an upgrade (5.x) to be able to support the "new" WebPlayer is a no-no.

If they stick to a 24 month per major release... we have 10 months before the next one. However, a lot has changed since 3.x->4.x. I suspect the UAS is now providing the major part of their revenue, making the company less dependent on license sales. So they could wait longer...

I mostly agree with the OP's assessment about what to 'not' expect, except for the reasons behind two of them;

- PhysX 3.x upgrade (webplayer backwards compability)
-- Unity already uses different runtimes loaded at content initialization for 3.x and 4.x. If an updated PhysX comes out for the 4.x branch, they can load that runtime instead. All we know is that Unity has heavily customized PhysX for their needs and I suspect they will switch to another physics engine altogether in 5.x, like they did with the 2D framework (hint; it's not PhysX obviously).

- 64 bit editor (broken dependencies everywhere)
-- I think this is still in the realm of possible for 4.x, and they said they have one internally... However! Apparently most of the trouble to get a 64bit IDE working is because of 3rd party software / plugins not playing well with Unity (old or depreciated and still in 32bit).

Cheers

 

- PhysX upgrade
- 64-bit editor

Those are a must have for 5.0 release IMO.

 

If that is true, then we might see a change in the licensing structure. I would like to see the licence broken down into components that you could buy individually through the Asset Store.

But I suspect licensing is still the major revenue generator, but with the Asset Store as a very significant part of the income.

 

What? No! Why?
Maintaining and mixmatching individual components and how they interact with each other is a pain to work with. The only advantage is the price point but in the case of Unity the price for a license is just fair and right the way it is. I really think that a component based licensing system has way more disadvantages than advanteages. For the company as well as the consumer.

 

Well, why not have both?

You could have the full licence version where you get Unity Pro with all its' advantages or you can pay a fixed price point for some Unity Pro-exklusive components like the Profiler or the ImageEffects. Personally, I would pay for these two. And if you want to you can upgrade to Pro anytime.

 

They need to support the Physx destruction. I aint gon write my own script if it already done! By professionals! Also, I would reprogram all the lighting and give it to the community for free, but I believe Unity should be fixing their lights. Their lights burn my eyes with its extremely bright shine. 64 bit editor would be a plus, so is the physics engine update. One thing that REALLY needs to be added.... wait for it..... 2D ragdolls. Those are soooo fun to play around with. I used to play with 100's of 2D ragdolls on the Flash engine but that kinda got boring. I would love to implement ragdolls in my 2d game. Would be cool hehe

 

For someone who works with character heavy projects, I would not only love to see an update to PhysX but for god sake, give us APEX! I would hate setting up my cloth now for my games because it was too heavy to use. And definitely gives up a viewer such as diffuse only, lighting only, detail lighting... etc. It is too painful to swap out shaders just to see how it looks for my game.

 

The Asset Store? I'm pretty sure it's been said elsewhere by Unity that while the Asset Store does indeed make money, it's only a small portion compared to license sales.

It's probably a good way for them to monetise the free license users, but lets do some maths.

As soon as people are using Pro they're immediately spending $75/month or $1500 as a minimum for people only targeting a single platform with a single developer. That's equivalent to Asset Store spends of ~$220/month or $4500 once splits are taken into account, which I'm going to go right ahead and suggest are already many times more than most people spend on the store. And I'll reiterate that they're minimum spends for license purchases - even a small studio with only a couple of developers is likely to spend at least 4 times that (two developers with Pro and 1 add-on).

 

I imagine some expansion of the 2D support - tile maps etc. Probably some new kind of lighting or something high end, and the new GUI system.

 

I prefer unity to support Havok physic instead of nivda PhysX.
Havok physic is not depend from GPU made unlike nivda PhysX.
Most of games I play use Havok physic like Splinter Cell Blacklist and BF3

 

in the live learning twitch chat (23.12) they mentioned that there is "..game changers in unity 5.."

illustrator vector file support?
free rendertexture(s) and other new free features?
..

 

Vector support would be amazing but I doubt it will be implemented soon :/

 

-64B Editor
-Mono (update)
-PhysX

 

(speculation)

The problem with your calculation is that buying a Unity license is once per X.x cycle. If the cycle lasts over 24 months, it is quite possible the revenue from UAS from Free and Pro users is bigger in the long term. I'll refer you to a post where I did some quick calculation about their licence business; http://forum.unity3d.com/threads/21...ensive/page2?p=1450997&viewfull=1#post1450997

I'm not disagreeing that licenses are still a major source of revenue, but I hope and think (speculate) the UAS is becoming an increasing source, and that's good for all of us.

Looking from aras twitter account, it is already very much in the works.

 

Here's one...make sure 5.0 works better than 4.3

 

You realize that this would BREAK all settings and physics solutions in existing projects, right?

 

Unity 5.0 is going to be da BOMB!

 

Not necessarily, just like the legacy animation system vs mecanim... both can exist in parallel and/or it can be as 'simple' as to select "Physic Engine" from a drop down in the physics options, just like we do for linear / differed rendering.

In the end, its for Unity to judge if other solutions are more efficient then their in-house customization of PhysX.

 

Oh yes, it is absolutely possible, I was just saying that I don't think it's likely to be the case right now (and that we were told in the not-too-distant past that it wasn't happening at the time). To me the long-ish cycle is by the by, as I don't spend anywhere near $4,500 per two or three years on the Asset Store, and that would be to match the income of just one of the multiple licenses I have (and there's more than one developer here, and we still wouldn't spend that between us).

So, when even the developers with budgets to spend aren't covering the cost of a single license, while I grant that it's possible for the Asset Store to be a major contributor, my experience suggests that it's not likely, at least at the moment. Which could of course be completely wrong.

I do also expect that it'll grow.

 

64 bit for sure is a must.

That is one thing that should have been implemented years ago. Unity is the only dev tool I use that isnt 64 bit (and its been like that for almost 4 years now).

 

Unity 5.0 Wanted Features

- 64 bit support
- Multi-monitor support (Editor Game)
- Visual Scripting (Flowchart Style)
- Road River Tool (EasyRoads3D isn't good enough, Cryengine 3's tool is the best)
- Improved lighting customization
- Improved terrain editing tools

 

I'm going to go crazy on this one , So this is what i'm looking for at the moment, if some of this can be added before 5 it could be awesome
And i know that maybe some of this already available on the asset store but i still prefer native integration solution because of "support". No offence to any asset store author, it's just there is some author somehow decide to abandon the product and support also :/

Rant Mode On

- 64Bit Editor is a must
- GI solution
- APEX please
- WebGL/HTML5
- visual shader editor
- Update built in Lens Flare (Geez how old is this stuff, Go check on Cryengine new LensFlare sytem UT :/)
- PBL workflow and Pipeline support
- Reflection Probe please, you guys already have this on butterfly effect
- Cubemap filtering (convolution,Disc, Cone, and angular filter). This is work in conjunction with PBL and reflection probe above. Btw Amd cubemap gen source are available now, how about integrate it
- Decal System!!!
- ScreenSpace reflection....!!!!
- Update waterPro please
- Synched GameMode for scene view, something like enable to see all effect in scene view (Post Processing, Ingame logic, etc)
- Better data streaming for big scene (Somehow at the moment it very slow (or maybe i did something wrong ))
- 32Bit HDR buffer or full floating point buffer
- Read and Write into G-buffer
- Support more lighting model rather than blinn-phong
- Terrain engine upgrade
- Extend terrain engine texture, grass, detail, and trees, painter to support another model not only terrain
- Upgrade Tree Creator
- Volumetric fog for full scene fog effect with shadow, we already have sunshaft which is screenspace, and i know there's volumetric lightshaft effect available out there which is awesome for perlight volume effect (And i heard it going to be available on next update )
- Lighting support for particle with shadow cast and receive, come on Unity research paper was the base for deffered particle shading on other engine, why don't it available on unity itself??
- Infinite Ocean surface !!!!, with access to it material of course so we can customize it
- We already have skybox, can we have option for Skydome please...
- there's some dynamic sky and atmospheric fog in the chase demo, can we get it

I guess that's all for now, if i found another else i'll update it
Actually i want to ask about mecanim but since i didn't know what the roadmap or plan for mecanim is so i guess i just gonna leave it now

Rant Mode Off

 

Personally I would like to see:

• Updated terrain to be competitive with other terrain solutions out there.
• "Physical based" shader/s.
• Updating and fixing some old components such as "skinned cloth" and "trail renderer".

 

Most Important:

* Physics Upgrade.
* Realtime G.I.
* Volumetric Terrain with Procedural generation.
* Physically Based Shading.
* 64 Bit Editor.
* LOD Generation. (I think this one is not a hard deal... but it will help us programmers alot).

 

Guess you'll want the price to stay the same, though?

 

what about the features we requested on the 4.x version and unity pro becomes like 2000$ ??

 

Nope, I'd happily pay double for that..

Add an always on data clustered compressed async asset streamer while your at it too Graham ..

You could have Unity super mega pro X edition!. Joking aside if I showed you some quotations I got for other engines, I think around $3000.00 would be an amazing price for a Unity HD edition and keep Unity Pro @ $1500.00.

Have a flagship engine, which you can trickle down technology to Unity pro and Unity Free. Great for press, more money for a different market like XBox One, PS4 and mid to high end PC's.

 

Nothing stopping Unity being more modular in this respect, creating the means for it rather than the end. Currently, Unity's core misses a few bits and pieces which do not allow Asset store to plug things as well as they might. So instead of adding all these features, it may be a better plan to optimise what is there and open up a few doors (safely as you can of course!)

So I'm more: less features, but more capability for us to add features without a source license

 

Guess you will want to stay competitive with UDK 4 and Source 2 around the corner?

 

Stop using 5 year old PhysX builds and I'll be happy.

 

well it depend on what we get

 

I sort of like the current system where you can just get plugins to do a lot of those improved extensions. Since the way it currently works means that people who don't need a lot of those fancier features don't need to pay extra.

I mean, there is already free and paid physically based shaders available, there is a brand new node based shader editor that is pretty fancy for $80 and Stumpys can still work for older shader types for free. Also GI systems are in development as plugins as well.

I would like having systems already in place to be updated however such as a less buggy mono develop and up to date physx.

a HTML5/WebGL exporter though would be lovely. Or a Unity Player that works with Linux. Especially if Steam machines and Linux OSs take off and for game jams its already would make things a lot easier to share with the Linux using developers for feedback.

 

Yes, I think that too. Better to make the current features more robust and complete than adding the features that looks "hip".

-Better API documentation
-More complete API/Finer control of some mechanism in Unity (e.g. GC)
-Extensible serialization framework
-Runtime Asset loading/postprocess

 

The main features that I'd like to see are:

- Opening multiple scenes (or editor instances) at once
- A better/faster API for creating dynamic geometry from scripts (e.g. custom UIs/sprites/special effects)
- Occlusion culling/scene management system that works effectively on mobile*
- Less unexpected collider-related performance gotchas.
- Async loading that works (when the target scene contains big collision meshes)

(*On mobile, the new 4.3 system is a huge CPU hog compared to the old PVS system. It might be better in some cases, but it can dramatically reduce performance in cases where there's not much being culled)

 

Multiple editor instances already works.
Google it.

 

No.. it's a fair question!

I prefer he says whats on his mind and keeps it real instead of having one of those "we can ensure you we are doing everything we can bla bla bla"

I agree PhysX upgrade would help a lot

 

Not any specualtions here, just a question:

I've been using Unity free for the last 12-18 months. Now I'm finally thinking in buying it, so I can distribute builds of my apps.

Do you think that it's worth it to buy Unity 4 now, if you are just starting to think about your project? Would be better to wait for the 5? And... There is any chance to upgrade from 4.x to 5?

Thanks and sorry...

 

Absolutely agree.

Given the needs of the many, it's never going to be possible to satisfy everyone and keep up with the cutting edge at the same time. Improved low level access and - crucially - strict UAS TRCs could go a long way to improving what people can create, and give buyers an increased level of confidence in what they're paying for.


And please update the Asset Server so it supports branching, file locks and other marvellous post-industrial revolution code versioning features. And scene merges - wouldn't that be magical

 

This!!

 

Well, apart from what everyone else mentioned, ADD OPENGL 4.x SUPPORT, FOR F***'S SAKE!

 

And?
If they aren't allowed to break compatibility in major releases, we'll be stuck with it forever.

 

Oh god no!!!

That will be heading down the old Garage Games/Torque route ..... and indie game engine models that work like that always end up in disaster.

The moment Unity starts fracturing their features in "buy this external ____" and buy "this ____ addon" is the moment another game engine will come in, offer more inclusive and steal all of Unity's users, forcing Unity to go bye-bye.

Unity should do the opposite if they wish to be around for a long time, and stay competitive, they should include more, and have more.

Imagine buying a new car, then the dealer turns around and says "oh by the way, you can buy the doors for this car over at this store". You'd tell them GTFO.

 

I wouldn't exactly call it "fair." You really shouldn't have to buy the full desktop license just to be able to buy mobile license if you have no plans to release anything to desktop. At the very least, since every other license requires the base license, they shouldn't cost the same amount.

 

I'd add to that dropping heightmap-based terrain in favor of voxel-based one (making easy caves), possibly using VoxelFarm engine (same voxel engine EverQuest Next will be using). Existing heightmap terrain would be converted to voxel ones automatically.

 

If all 5.0 added was a new GUI system, a 64 bit editor, an up to date physx implementation, OpenGL 4 support, and a less S***ty GC I'd be happy. If that's what we got, I'd be happy with nothing but bugfix updates for the entire 4.X cycle.

edit: new GUI system, not a new GUI. Unity probably has the most useable interface out of all the engines I've used. I suppose that's damning with faint praise though.

 

Looks like you've completely misunderstood what I proposed. It would not affect asset store in the slightest. I'm not proposing unity sells things modular. I'm proposing more access lower level, that's all.

 

+1 So for instance, developers could make their own systems instead of waiting for Unity to include them inside the editor. Unity might include that feature later, but if you need that feature now for your game, you can at least do it and don't have to rely on UT.

There are many good features but most are not complete (like Shuriken, the most closed system there is) because UT doesn't make games they make tools. They might not think of certain features as useful. If the developers had access to lower level stuff they could add the feature they need now instead of hoping that it might be put in one day.

Now on another note, what I'd like to see in 5.0 is:

-64 bit editor
-Upgraded Nvidia physics !
-More open systems! Expose the stuff please so we can use it in a creative way !

As for wishful thinking :

-Upgraded mono ! (but I really doubt it will happen before 6.0)

 

This, but also keep the traditional license system. Like, I don't want to spend $1500 for Unity Pro, but I WOULD spend $200 or so to unlock render textures.