Official Visual scripting roadmap update - August 2020

ok in short term is bolt 2 and dots vs will be introduced in the future or not
and does it will be separate vs solution or will be one

 

Can someone explain the "siloed" verbiage being used? I'm a new user. Not a programmer or animator. I believe Unity wants to be inclusive is the idea? Bolt 2 sounds perfect, more than that, Unity spent years putting whole focus on Bolt 2. At this point, my evaluation is that I need to learn C# because Bolt is not ready unless I want a headache down the road. And Bolt is what made me think I could get into this. Very disappointing. After this amount of time, Unity needs a Bolt 2, or a large update very soon. This decision will turn off a lot of people. The market timing with Bolt 2 was just right. Instead, a lot of people will go another direction.

 

Definition of siloed
: kept in isolation in a way that hinders communication and cooperation : separated or isolated in a silo https://www.merriam-webster.com/dictionary/siloed

In Unity context, they were supporting Bolt 1, developing Bolt 2 and DOTS VS as separate visual scripting tools with their own UI/UX paradigms. Each tool incompatible with the other and using different tech both on front and back end. This siloed the user base between these different tools.

Also, Unity didn't spend years on Bolt or Bolt 2. Bolt was developed by the company Ludiq until Unity acquired it from them recently.

 

I didn't catch your meaning. They cancelled Bolt 2 because it was incompatible both tech and design wise with their goal to unify all VS tools in Unity.

Bolt 1 hasn't seen any meaningful feature updates for the past two years in favor of developing Bolt 2. It has been in maintenance mode for a long time now and it'll continue to be there until the new Unity VS tool drops. At which point, Bolt 1 will be phased out and not updated for new Unity versions. And the new tool will support both Monobehaviour and DOTS based visual scripting.

They've stated that Bolt 1 will be supported for the current versions of Unity, meaning anything from 2018.4 to 2020.4. With Unity 2021, they will expect you to use the new tool. And it’s mentioned in the announcement that the new tool will be able to read Bolt 1 graphs, so using Bolt 1 now is considered safe according to Unity.

 

The Uni-Devs have seemingly abandoned trying to talk to their community at this point. There's been dead silence for a while and that's a telling sign how disconnected and little they seemingly care about the people who use, fund and support their game engine.

Since the out of the blue announcement for them canning Bolt 2 I've talked to people in the Bolt Discord, Installed UE4, been learning Blueprints (with the same intent to transition to coding as I did with Bolt and Unity), left the Bolt Discord and uninstalled the Unity Engine.

Personally I have distrust for people to mislead and then ignore their communities and Unity has been placed on my S*** list for that because of this whole fiasco.

So long and thanks for all the fish.

 

Is it going to output c# ?

 

It won't output C#.

(Also, I'm just a random long time Bolt user, I'm no authority on all things Unity VS scripting. All I'm writing here is my current understanding of the situation based on publicly available information).



So let's break this down step by step for everyone who's still reading the thread. Seems like some people aren't sure on the specifics of the announcement:

1. Monobehaviour based visual scripting is what Bolt does now.
2. DOTS is what DOTS Visual Scripting does. DOTS or Data Oriented Technology Stack is a completely different way of coding games in Unity. It is also many times faster than Monobehaviour based C# code. Quick google search lead me to this comparison someone made - from 13000 objects on screen to 68000 objects at the same frame rate. They will combine both Monobehaviour based visual scripting and DOTS based visual scripting into a single tool - Unity Visual Scripting.
   
3. High performance snippet based nodes - this means the nodes are directly implemented in C# like Playmaker's nodes or Flow Canvas implemented nodes. Monobehaviour snippet nodes will be many times faster than Bolt 1's reflection, but not as fast as Bolt 2's generated C# code. Exact performance cost will be known once the tool is available for testing. And DOTS VS snippet nodes might end up being faster than Bolt 2's C# generation. That remains to be seen.
   
4. High level artist nodes. Again, like Playmaker, these nodes can contain 20 or 500 lines of C# code. A single node could be a basic character controller implementation with configurable options on it or it could be an advanced tweener node with all the relevant settings on it.
5. They also write that they'll support auto-generated low level API nodes which is what Bolt 1 does now via reflection. There are multiple ways of achieving this besides reflection. So it could be faster than what we have now. Gotta wait for their tech post where they'll go into more detail on this.

In short, C# gen is not planned, but the new tool will be considerably faster performance wise than Bolt 1.

 

Any update on any compensation those who bought Bolt 1 specifically for Bolt 2 will get? Like, I only got Bolt 1 because I knew I was supporting and going to eventually get Bolt 2. I would have never bought Bolt 1 if I knew Unity would buy it and gut it.

Seems like this is happening time and time again with things Unity buys.

 

I did an edit of my post from this morning. Again thank you for your thoughtful reply and time.

 

You could well be right and from @Ex-Crow post https://forum.unity.com/threads/visual-scripting-roadmap-update.951675/page-8#post-6282116 This kind of sounds like what this could be. I don't see how C# can be out
as many of the assets are in c# .. or is that gone as well....
I still think it is a shame Unity bought Bolt 2 when it seems they did not need too.

In any case we can all hope for a better future.

 

I think c# will evolve into a variant of itself in unity, that's already the case with burst compiler and the math package allowing to do shader like swizzle syntax. If it grow organically instead of rash bureaucratic decision, it could be a beautiful things. The needs for Dots was from advising on multithreaded, after giving the same advice all other again at different places (ie practical concern) they start thinking they could enforce it in a language. That's the good unity, only communication was a bit too early, it doesn't replace thing, it doesn't break thing, and whatever limitation it has is forward facing rather than backward facing, it's more of an opportunity than a requirement (unlike a certain URP or shadergraph), and even in a sorry state it's useful as a step ahead so you can deal with the clunkyness trade off usability.

 

I think it will be consistent to go from old tech (16 years old) like MonoBehavior to new one modern and performant but.
Unity say than they never get rid of MonoBehavior, they atlast stay as authoring layer in editor.

I think someday (somewhere in between 2025-2028) Unity just get rid of old Tech deep inside Unity and make MonoBehavior be based on ECS Backend. So conversion actually will be just from Authoring ECS representation to Runtime representation.

This looks like natural way of evolution.

GameObject become just wrapper around Entity struct
MonoBehaviour can inherit from ECS ClassComponentData and have overridable Event methods (Start, Update, ...) and system that call that all

GameObject default data like layer, tag, flags will be decoupled and allowed to be absent
TransformComponent become not mandatory one on GameObject

May be some additional small changes and We will have the only tech. Open sourced and performant tech in editor and runtime.

 

DOTS was announced two years ago and we still do not have an good version from the UX point of view. Moreover, DOTS is not the solution for everything. If you have lots GameObjects in your scene DOTS can improve performance but if you are not running into such circumstance DOTS is only making development harder. Waiting until 2025 does not make sense to me. The technology will become obsolete. What I see is that Unity is taking too much time to make the new technologies they announce production-ready. It is the same with SRPs. I cannot be confident on using technologies which are in preview forever. It is too risky.

 

My point exactly, they seem to have bought bolt 2 Knowing it was no good for the back end of what ever they are planing. Bolt 2 could have been done and out by now. Independent developers are motivated to get things done and the good dev's care about their customers and their assets. As a result from Listening to their customers Every day and acting on all the good advice ( and some times free help ) from the more knowledgeable user's of their assets.

Seems like most of the assets Unity bought from the developers only went into a compatibility mode for newer versions of Unity,, the only upgrades to make sure it runs in newer versions of unity. So yeah I agree on the time it takes versus, what seems to be small teams of independent developers make features and upgrades way faster. 4 or 5 years to wait seems forever and so much can change by then.

 

I am in a similar boat with you. I was using another visual scripting tool a little over a year ago that I bought, but lots interest because it was kind over-complicated still for me. I'm attempting to learn raw code and have done a fair amount of udemy classes etc but my greatest success was a game I made with stencyl, which was a really good GUI but the engine sucked ass and was too limited.

I recently saw bolt and was like maybe I should try this, I like the style of its presentation and saw unity acquired it and was developing bolt 2, then I just read bolt 2 was being canceled. Now I am just confused.

Are they developing something new, what is going on with Bolt? I am interested in learning it, but I have a big project I want to start and I dont want to invest my time and energy into bolt for no reason. I was going to wait for bolt 2 because it sounded like it was coming out soon>

Anybody have a semi-clear answer on when the new tool will be ready, or if it'll be worth my time to invest time and effort into bolt 1???

 

This could be the next best answer .. A Bolt User and good guy, is making a free add-on called UAlive for Bolt and will soon be able to do what it does now And out put C# for more info start here ... https://forum.unity.com/threads/free-ualive-bolt-inheritance-live-to-compiled-c-generation.918518/

Also if you are a bolt User another bolt user has 100+ Bolt Super Units free in the store .. https://assetstore.unity.com/packages/tools/visual-scripting/bolt-super-units-177410

Go check out first though https://forum.unity.com/threads/fre...to-compiled-c-generation.918518/#post-6231285 If you want to go to the first post on this.
I am super happy he is doing this for us.

 

I'll check these out, thanks

 

I think it will, Bolt had a huge community and still could. I think maybe more then a few went to go give ue4 a serious look,
perhaps many were so disappointed they turned away...
Can't know for sure, But he will put this on the store and Discord has many people in the Bolt forum and LifeandStyleMedia.com @@JasonJonesLASM has his own Discord set up.
It has been said some where in this forum ( I think ) " Bolt was the best thing to come along for Unity in a long time "
For me it was really true, is all I can say.

I am glad you are as excited by this as I am.

 

Still that is building off of Bolt 1, right? It doesn't have the advantages that were being incorporated with Bolt 2 as far as I know.

Just wish that Unity would release Bolt 2 as open source or something and let others polish it up, rebrand it, and sell it or something. Clearly, Unity doesn't want anything to do with Bolt 2, so why not give it to someone who does want to use it since there is also clearly a demand?

 

I agree fully that Unity should let some one else have Bolt 2 to develop.. I personally would pay another 70$-100 for
Bolt 2 development ( Even thought it was gona be a FREE update ) That aside, Yes Bolt 2 imho was going to be worth the wait.

I was reading through @JasonJonesLASM Discord last night.. and I have basically Zero coding skills, But I am fairly sure that all he has planned Bolt with his ad-on will be able to do much more then just output C#. The way I read it, is there will end up being some incredible features in it, that I am not sure Could be in Bolt 2. I am not a programmer so I can not say for sure.. but that is the way I read it. It could be said now that if he did a version 2 with the right coding, this base could in theory, make it that you could keep old dead coded assets up to date. But right now that is just a gleam in the eye's. For now, I think there will be a lot more then just output C# in this. How it could all be used, is beyond me really.

 

Tbh I have known of 4 guys in the last 3 years that have jumped ship for Ue4 because of unity " stuff and disappointment "

I can not disagree with you learning UE4 as back up. Depending on your age and circumstances, I can see how this is not a bad idea. For me at 60 and my skill and or hatred of having to learning coding, is like Bolt works for me and my just playing around, making levels and games, as a hobby it is fun for me. So it seems one of the first things Unity did with Bolt was to make lessons and put them behind a Pay Wall. It is now free b/c of Covid, but still.

So i am left to speculate that what happens if they make this VS-Dot's thing hard enough to put lessons behind a pay wall. ? Then it would follow that maybe its more about money then community.

Peoples reason for leaving Unity seems to be they have lost trust. Every year there is a new crop of people that want to make games. " Playmaker been going for years and still going, so that's good.

The way I feel is that if/when unity drops Bolt and make it no longer working in newer versions. At that point I will back up all my assets and Unity Versions and be happy with what I got. My main reason for not going to Ue4 is that I don't like the interface and it seems it is all blue prints, and personally that ends up frustrating me as well.

I am hoping Unity quietness now here, is a re-thinking the path forward for all the people that love Unity and Bolt.

 

I can recommend playmaker until this all is sorted.

 

While I’m very capable of coding I like bolt to visually show people why their logic is wrong it can often be more clear with the visible arrows and values. Oddly though never got on with playmaker as well a so have bolt. Hence it’s so disappointing that bolt seems to be floating half dead in the water as there are a lot of annoying quirks and even those haven’t been fixed

 

I can strongly recommend FlowCanvas until this is all sorted:

https://assetstore.unity.com/packages/tools/visual-scripting/flowcanvas-33903

 

If you like something closer to scripted programming flowcanvas is better, for me the power of playmaker is not needing this proximity to programming languages

 

Flowcanvas is no closer to scripting than Bolt is, in fact many things are easier in Flowcanvas due to more integrated macros.

 

True, I mean compared to playmaker, I could be wrong because I have not used flowcanvas yet

 

FlowCanvas is good and if you pair it with NodeCanvas you have both a visual scripting and visual finite state machine. I've been meaning to get into that more and with these recent changes it seems these two assets are the best bet since they've been around for a long time and have a lot of support.

There are a few new assets that look promising, but none are as fleshed out as these.

 

Flowcanvas + Nodecanvas is more powerful than Bolt, since you get also Behaviour Trees and Dialogue Trees in NodeCanvas. NodeCanvas also allows you to use pre-built conditions for state changes, rather than how Bolt only has transition graphs, so technically it can be even easier to design a state machine than in Bolt.

 

It doesn't have convert to C# though. There are a few other features that Bolt 2 was going to have that FlowCanvas doesn't have. Forget them offhand, but still seems like there are pros and cons to all of them.

Just wish there was something akin to UE's Blueprints for Unity but with equal or more functionality... is that too much to ask for?!

 

You're going to have to forget about that one now, unless uNode gets stable eventually. Bolt 2 was always a prototype, and now it's dead.

Yes, Flowcanvas.

 

Flowcanvas is nowhere near the functionality of blueprints, it does have a solid amount of basic, low level nodes by default but you would have to write most of the nodes yourself to create any game. As far as I remember it lacks basic nodes for literally every single package unity does own like cinemachine, timeline, level streaming, post processing etc.

Combination of Flowcanvas + nodecanvas is the best visual scripting toolkit and unity should have bought it instead of bolt but It's not as good as blueprints.

 

I don’t think there’ll ever be anything quite like BP though, unless Unity wrote the entire thing from scratch rather than buying and hacking someone else's; and the poster asked for akin, not perfectly replicated. This is the best you’re going to get.

What do you mean,

you would have to write most of the nodes yourself to create any game

? Just write the graphs like you would code.

 

Well, that's theoretically the plan unity had for the last 2 years. The problem is that they never really tried to create a proper plan, because of this we had 3-4 design changes (Mono + code generation then DOTS + code generation then DOTS + code snippets now DOTS + MONO + code snippets?)

I simply mean that flowcanvas is a raw visual scripting tool. You have to put a lot of work to make it usable in real game development.

 

They probably test against a narrow set of game studio who want to silo their designer's input, and the designer on deadline don't want to think to much about the whole stuff, since they have programmer as support.

 

Has there been any additional response on this from Unity? Haven't seen anything new posted.

 

nope. Sadly radio silence. I believe the unity rep was going on holiday or something for a little bit but I thought they would be back by now

I find it hard. I want to support unity but there are a number of aspects that slowly make me question more and more why. I’ve bought some singular spectacularly crap assets and because there’s no try before you buy you get it and it doesn’t and if the writer doesn’t agree (which irrelevant of it not working they don’t have to) unity won’t refund.

the ideas unity has are great but they seem to have bitten off way more than they can handle and like dots that I’m sure was gonna be semi usable earlier this year on the original plan. Still seems to be evolving enough that almost code from 6 months ago now barely compiles

I like the idea of converting a lot of things to packages so they don’t stink up your project folders and you don’t commit paid assets to github for a lot of things. But unity doesn’t feel so polished any more it feels too spread out. Too much like it’s full of scaffolding and a good sneeze could break it all.

 

maybe but even then FC is better then bolt1. in my opinion

but i didnt tryed uNode/ people say it can transfer graphs to C# code. and have vertical workflow too

 

I'm trying uNode. I'm not a programmer, so I can't explain the technical advantages well.
Naturally, it's different from Bolt, so it's confusing but fun.

The problem with uNode is that the official documentation is outdated. It's different from the latest 1.8.7 and difficult to reference.
Currently, uNode is developing the beta version of Ver.2, and I hear it will be official by the end of this year.
Presumably, the official documentation won't be updated until that's complete.

Fortunately, the kind people and developers at uNode's Discord can help me with advice.
On the other hand, it can be difficult to understand how to use it without access to Discord.
If you're considering purchasing uNode, I suggest you wait until version 2 is official (and the documentation is updated).

 

Same exact scenario here, and I bought Bolt TWO WEEKS before the acquisition lol.. I waited over 10 days for a support ticket to tell me they weren't going to give me an asset store credit, and to go piss off basically. It has certainly made me more reticent about spending anymore money on the asset store again..

 

Ok. Reading all this for the last weeks and Unitys reactions, I am really frustrated and thinking about switching to Unreal. I am teaching at an art school on the edge between programming and design, using Bolt2 since its first alpha. I did Playmaker before, then Bolt1. But since the classes version of Bolt2, I thought "Wow, this is now a thing. VS without compromise, because of all reasons you all wrote here before. And even if I came from JS, C# etc... I think VS is the better way to get visual thinking people, creatives, designers into the logic and structure and let them feel the power of programming.
But now ...
a. ) going back to Bolt1 is not an option, sorry. Can't do this, if you felt how Bolt2 worked.
b.) Telling students to wait ... for years ? ... till the "new" VS whatever is out. Hoping that the performance will be better than blueprint vs. C++ ? Also no option. Also if I think about Unitys timeline in VS development.

So there are just two options. Letting go the idea of VS as an equal way to create interaction/games, teaching design students directly the hard way of C# .... or .... switching to the actualy better VS system, and I think this is blueprint.
Not decided yet. Maybe testing Unreal for a week and see how it turns out. :-\

 

Keep with playmaker for now and teach students how to write c# actions for playmaker, so you have the best of visual and programming combined.

 

Flow Canvas + Node Canvas is also a valid alternative. They have some features Bolt 2 touted (but no C# gen). And it's proven in production of several well-known games.

 

Here's some stuff in FC that isn't in Bolt:

• Big graphs are still fast enough in the editor if you zoom in, Bolt is slow no matter how far in or out you are
• +=, -=, *= and /= operator equivalent instead of clunky get-set methods in Bolt.
• Faster fuzzy find, only bare-bones types are added at first, so you don't have to wait to rebuild everything and it's already a much faster fuzzy find in general and new types are added instantly without any rebuild
• OnAwake
• Signal assets instead of Events that are just a bunch of "arg #" ports sending info to each other (similar to Bolt 2's invoke objects)
• Drag selecting inside UI groups doesn't move the whole group, unlike Bolt 2
• Multiple global blackboards with DontNotDestroy toggle built-in
  
• Disable flow connections without deleting them (similar to // comments), something that n̶o̶ ̶B̶o̶l̶t̶ ̶v̶e̶r̶s̶i̶o̶n̶ ̶h̶a̶s̶ ̶e̶v̶e̶r̶ ̶h̶a̶d̶ ̶f̶o̶r̶ ̶r̶e̶a̶s̶o̶n̶s̶ ̶t̶h̶a̶t̶ ̶e̶l̶u̶d̶e̶ ̶m̶e̶ ̶t̶o̶ ̶t̶h̶i̶s̶ ̶d̶a̶y̶ Bolt 2 got late on before it was terminated but Bolt 1 still lacks.

 

This looks handy - I'm going to want FlowCanvas.
Also, would it be possible to collapse the grouped Nodes?

 

This is unfortunately one thing that I'm yet to see in either VS tools, although it has been mentioned to the developer very recently and he seemed interested in adding it: https://flowcanvas.paradoxnotion.com/forums/topic/feature-request-collapsing-nodes/

The disabling connections is really good in FC:

 

You can collapse a selection to a macro in Flow Canvas much like you could in Bolt 2.

P.S. Bolt 2 also had connection disabling via right click context menu.

 

Did it? I never knew that, I remember asking for it a few times, thanks for letting me know. Macro's are great, but not the same as collapsing a group.

 

Unbelievable...
So many people are crying the Bolt 2 failed.
Are you looking for an alternative: Playmaker, FlowCanvas etc...
And you don't know what treasure you have within reach.

This treasure is called uNode, a unique visual tool that creates full fledged clean C # code. Still developed by a helpful and modest creator.

I found a picture advertising Bolt 2 and put together the same code fragment with uNode (probably the same, but I don't know Bolt 2 and I don't know what exactly the nodes mean) --->



I check this thread sometimes and it's really hard to believe what you write.

 

Which is which, because the second one need to work on his visual design sensibility, it's barely readable due to bad color choice and on my screen it's flashing.