Official Visual scripting roadmap update - August 2020

And if you're reading this and think "ok we understand but we still think this is a better idea" some kind of mock up to explain your goals would help a lot.

Cause either you're really REALLY off base, or your explanation is very inadequate.

And if you don't have an exact plan yet to better explain, maybe hold off on announcing these things until you have a better concept of what you want to do.

Bolt 2 as a concept to the people here is very clear, and they knew exactly what they wanted to do with it once they got it, but they can't see how they can achieve those goals based on your description cause it sounds alien to what they expected.

I mentioned polling before, maybe this would help gain some perspective on what people were expecting. I made a straw poll, maybe this will help make things more clear.

Straw poll: "What I expected/wanted from bolt 2..."

I know I probably didn't cover everything, but I added some of what I saw said here. And if you said "Other" either the point was made already and I missed it, or it's not there. Either way it means you should post!

 

Here's a idea for you Unity, return ownership of Bolt and Bolt 2 to it's original owner and creator. Doing it at no extra cost to them because you are the ones who royally screwed this up.

Allow them to pick up where they left off, give the community back a tool that you made promises to handle, develope and release but have now shown you cannot be bothered to. Give the tool back so that those of us who really looked forward to it still have something to actually look forward too.

Listen to your community, or remain left in the dust of the competition (UE4 and eventually UE5). You have shown dishonest and loyalty issues. Something that is a big screw up in a entertainment medium riddled gamers and game devs a like who are fed up with being lied to, swindled and misled.

Show you are competent with your actions, or don't and lose the trust of a large base of your community and that also means losing customers and their money (which you clearly care about more).

 

I know that my perspective as a complete newcomer shouldn't really be considered. I just started using Unity and Bolt 1 yesterday and this is some seriously concerning news because I was so excited to deepen my knowledge of Unity and (as a visual person) specifically Visual Scripting.

@LaurentGibert
By reading the roadmap, I guess I am the demographic. and I'm telling you this - I want visual tools that are not only beautiful to look at, I want them to be as PROFESSIONAL and CAPABLE as actually scripting in C#.
I feel professional visual tools for programming SHOULD be programmer-centric, and I don't know a dime about programming.
Please don't take this as an opportunity to dumb visual scripting tools down.

Just release Bolt 2 as open-source, please.

 

I can second this opinion. It's a right shame the direction they are moving when Bolt 2 has been highly anticipated by the visual scripting community.

You keep saying how you're doing this in the best interest of the largest user base, but all you're proving is that Unity doesn't listen to that user base. 95% of responses to this news has been negative, as such, Unity really needs to reconsider this decision.

 

I would love to hear something from Ludiq

 

Bolt and DOTS VS were different tools.
Perhaps the decision that Bolt1 was easier than Bolt2 when considering integration with DOTS?
I'm not a programmer, but one who was looking forward to Bolt2.

 

I guess I understand your desire to reunite on top of a robust foundation. yet for many of us, Bolt 2 was exactly what we were looking for. And it is right there, with a few bugs away from something we can use.
Right now it feels really like you are not caring much for your community, killing a tech we loved after buying it back, looks like a very aggressive move.Could you at least release the current alpha of bolt and leave it to us as it is? just doing that will appease us while you go back to bolt 1.

 

if things will take long time to develop ,maybe is better to continue developing DOTS VS. Most people will use Bolt for 2 month then will start writing they own code or continue using bolt1 for simple thing which is more capable to do.The direction made by unity maybe is right for future and unified solution but you to have reconsider how long user are waiting for VS.Continue developing and fix Bolt 1 is worst for user and Fixing Bolt 2 is worst for you.

 

Not sure we'll here from them considering they likely have some sort of contractual agreement that limits what they can say, but I have seen them liking some of the comments here (the ones that promote the work of Bolt 2) so I'm fairly confident that they likely aren't very impressed with this latest news from Unity. Definitely a big slap in their face and those that offered feedback.

I think the comments in this thread are fairly unanimous in that we want Bolt 2 rather than Bolt 1 and whatever poorly thought out plan the Unity developers have in mind. They are clearly out of touch with their consumer base and can't provide more than shallow excuses for why they chose not to use Bolt 2 over Bolt 1. This has been extremely eye opening for me. I'm very much considering using Unreal Engine (for 3D) and Godot (for 2D) for future projects over Unity because they at least make solid decisions with their product (UE) and listen to their community (Godot, open source initiative).

What I want to see Unity do is one of the following, otherwise I'll just avoid using any VS provided by the Unity team as it'll never be better than what the asset developers out there can make.

a) Drop Bolt 1 and instead build off Bolt 2. DOTS (based on what I've read) should instead be re-engineered around Bolt 2 as customers feel that Bolt 2's UX is far more superior than Bolt 1 and DOTS.
b) Open source Bolt 2, or better yet return full ownership to Ludiq as they know best when it comes to the development of Bolt 2.

 

I can generally understand where the Unity team is coming from... but the explanation has really made me really worry for the future of Bolt..

Ex-Crow has waxed poetic over the many limitations of Bolt1, but I'm sure over time incremental improvements can be done on top of it. The issue is that there are many structural problems inherent in Bolt1 that were quietly redesigned from the ground up in Bolt2 and not ported back because they would break functionality. Massively.

I'm not a programmer, so the under-the-hood structural improvements are beyond my ken, though still very important. But the stuff that I'm interacting with every day, I fail to understand how those can be improved upon without breaking compatibility dozens of times instead of just having a clean break that would have been afforded by Bolt2.

Strongly typed inputs, Class Events, Generics, Any State, Co-routines, Delegates. These are all going to require a break in compatibility and silo-ed understanding every time. And those are just the surface things I can touch, who knows if there are structural limitations inherent to Bolt1 in the Serializer or how Bolt saves its variables or a dozen other nitpicky things that will break compatibility each time as well.

Until these structural design issues are touched upon, I'll remain convinced that Bolt1's improvements will calcify before long.

I have an Artist/Designer background and not a Programmer, and you'll find the same to be true for most Bolt users. Bolt is amazing because of how comparably easy it is to start and grasp its logic, but also for how deeply extensible that basic understanding can take us, whether thats towards writing our own basic scripts or utilizing other plugins. The comment of Bolt-2 being too programmer-friendly seems to only make sense when the other extreme is to become Playmaker-like. Which is horrible because it doesn't matter how easy it is to understand a tool if you can't then extend your understanding to do anything you want. I'd argue instead that being artist-friendly on the surface yet hewing closer to its programming roots is the only way forward.

 

I think the best move that Unity can do right now is releasing the current state of Bolt 2 as an OpenSource repository in GitHub, so the users that are interested in see Bolt 2 finished can make it possible, or better yet, having the Ludiq team working on it again. Is such a waste throwing away 2 years of development (or more) of this amazing tool, and based on the comments of LaurenGibert, it seems that the Bolt 2 project is not usefull anymore for them, because 'snippet based nodes' is a totally different architecture. It is possible to make Bolt 2 open source in a near future?

Edit: Also, the main reason i personally use Bolt is because my computer can't handle external editors, because the CPU is really slow, but i have a decent GPU so i can work with 3D modeling tools and almost any image editor. So having an internal tool to create logic for games is really convenient and lowers the entry level for people with scarce resources to make games for a living.

 

This is a very bad news and direction for Bolt. It will teach me not to get excited by a new Unity thing ever again.

Please fix the variable consistency issue asap then, because right now it's a total mess in Bolt 1. And also make it vertical, I can't look at my spaghetti graphs anymore. That should not be too hard to implement.

 

Open source! Open source! Open source!

Flowcanvas is everything Bolt 1 wishes it could be, Bolt 1 doesn't need time wasted improving.

Bolt 2 was supposed to be the ultimate VS tool, designed from the ground-up to fix every architectural issues that all of the other VS tools have, Blueprints included.

 

As much as I like the open source sentiment, I don't think that would solve anything. It wasn't just a few months away from release and it would need a serious, organized and a highly experienced team of developers behind it. With no funding, it's hard to see that happening. Though I'd be happy to be proven wrong.

 

I have some hope for those who want codegen. Don't worry they mean codegen will not be available as a backend but it will be(at least should be) available as a feature like ShaderGraph and VisualEffects Graph.

 

At least Unity were nice enough to refund the Bolt 1 purchase price to those who mostly bought it in anticipation of the included Bolt 2 upgrade. :-/

 

It seems a bit like Unity wants to have a marketing push to finally have visual scripting included and neither bolt2 nor DOTS Visual Scripting was ready to go now.

Which I really do hope wasn't the primary motivation behind this move. Incremental improvement is fine and it's a good sentiment to work on something solid, especially in the face of all the groundwork several Unity features are stuck in.

Best wishes to the developers, seems like there's a lot of work to be had.

 

I'm also a non-programmer and i feel the same way. I'm not happy at all with this decision as i was really looking forward to getting my hands on Bolt 2.
After reading the thread if feel like, if they really want to make the right thing for non-programmers, then they made the wrong decision.

 

Unity is learning from past mistakes here, and deciding on a unified solution to help us users avoid future pain. I agree with all that.

Ironically... I was excited about Bolt 2. But as soon as it was acquired by Unity I switched my project over to Flow Canvas and its sibling Node Canvas... because as a Unity user I too have learnt from past mistakes (trusting Unity with things like this!) I’m quite happy with my decision now!

@LaurentGibert Is a unified UI really that important?? Because I think most people here would have been happy for Bolt 2 to keep its perfectly good looking and functional non-unified UI a couple of years longer if you were to decide to build on top of that instead of Bolt 1.

Also... Removing the code gen is a major mistake in my opinion. You’re removing the one feature that would have made Unity an invaluable teaching tool in the classroom (potential user base of billions). And you’re doing this while telling us you think you’ll be able to appeal to a larger audience this way?

On a positive note i do appreciate you communicating with us Laurent. It’s brave to place yourself in the firing line and it’s also good to see people staying respectful in this thread and not losing their cool too much (could the two be linked? )

 

What about those who bought Bolt 1 hoping to get upgraded to Bolt 2 in the future, but before the cut off date? I fall in that category and I'm not getting any refunds. I'm just getting doubly screwed.

 

I also bought it shortly before the announcement of the acquisition (in anticipation of Bolt2).
I bought the bolt at the company, but I bought it myself.
If I think I paid the money to Ludiq, I wouldn't miss it.
However, I didn't buy it if I knew Bolt2 wouldn't continue to be developed.

 

Yeah, same. I bought it from Ludiq because they said Bolt 2 would be included in the purchase and they would keep developing it to the state of launch equal to or better than what they did for Bolt 1. But now I get none of that... makes me not want to get any more Unity assets or support Unity in the future if this is what is going to keep happening (and it has). Oh well.

 

Hey thank you all.

I really feel I am letting you down, and this is painful. We honestly weighted pros and cons before coming with this plan, and apparently outweighed certain elements that are not resonating with you right now, or might not resonate with you at all. The good thing with a plan communicated early is that we can debate it. We truly all win in having an open discussion like we have now. So thanks again.

In the coming days, I will reach out directly to many of you to have a deeper conversation on the topic, document your context and your expectations, as well as list all the points mention here and challenge ourselves again on each of those topics. Either we do have good solutions to propose, or we are mistaken on certain aspects, which is always possible.

Please bear with us along this path, we will build Unity together to serve your best interests.

Thanks!

Laurent

 

Thanks for the open conversation! really appreciated

 

Thanks for being open and willing to discuss this with us in more depth.

To me the biggest issue is that Unity wants something quick and easy that can be used for non-programmers. But the community instead wants something that is more along the lines of Blueprints -- something both programmers and non-programmers can use that is flexible and expandable. I think that's where you've really missed the mark.

The UI matching is basically a non-issue for a lot of people because we want function over form. Just seems like going for Bolt 1 over Bolt 2 is a step backwards. I do like the idea of unifying both Bolt and DOTS visual scripting, but not at the cost of losing everything that Bolt 2 has developed since Bolt 1.

 

I am very happy that VisualScripting is integrated as a core function of Unity.
But the problem here is that it will be next year or later.

If Bolt2 is destroyed I have to consider using Bolt1 or consider another VS (Playmaker? FlowCanvas?)
By the way, will Bolt1 continue to be updated until it is integrated as a core feature?
Is it possible for us to enjoy the evolving Bolt1 while waiting for integration?

I liked the Bolt2 alpha version (especially the UI).
The loss of Bolt2 is very sad, even as a result of Unity staff's scrutiny.


I'm not good at English so I use Google Translate.
If the text is wrong, please forgive me.

 

It's not even a question of looking like blueprints, bolt 2 is way beyond all of that.
A unification between the 3 tools would be perfect, we all agree here, it would be magical.
But, if it is about bolt 1, we will lose everything that bolt 2 and VS DOTS had to offer.

 

To build the most powerful tool for Unity:
1. finish Bolt 2
2. produce many video tutorials in cooperation with best teachers (like OWS) + build a growing structured library of endless examples for Bolt 2 (build this library constantly, with the possibility of commenting - people themselves can add a solution in case of an update/change + custom examples/tutorials) ... support building online community
... the second point is crucial in people's decision what tool to learn first, otherwise they skip from one tool to another and they lose time and interest

 

Conceptually, I can agree with the direction - deeply integrating visual scripting within the engine and unifying all (or most) Unity's visual scripting systems into a singular tool is a great idea. On paper, this is best for the engine long term. Rather than making bridges between different tools and architectures, you have a true unity of tools that work with each other seamlessly under the same UI/UX paradigms.

But the cost of that might be too high for many of us. I guess it's not clear in what ways Bolt 1 will be updated to even approach the progress made in Bolt 2. The re-architecture of Bolt 2 was chosen for a reason - it was the best way to address all of Bolt 1's shortcomings. And it was successful at that.

Bolt 2 improved editor performance and brought us native C# performance in builds. C# gen for many of us was groundbreaking advancement in visual scripting because is considerably decreases the gap between programmer code and artist/designer code. It offered performance by default. Bolt 2 also fixed the variables system, the custom events system. It introduced a proper structure in the form of Bolt Classes that promoted the creation of smaller, more reusable graphs, offered a more effective way of communicating between graphs and enabled more effective reuse of graphs since variables and graphs were grouped together under classes. Graph Explorer window offered the perfect overview of the whole project's structure. Interfacing with your own or 3rd party code was also improved via the incremental unit options extract. And there are a couple dozen more major and minor improvements that Bolt 2 did right and did better than Bolt 1 in just about every category.

Some of us have been waiting for Bolt 2 since it was first announced two years ago, so feelings are running a bit high right now. Bolt 1 in that time hasn't seen any meaningful updates. A compatibility patch here and there, a couple of memory leaks fixed, but that's it. So it's tough to see all that progress gone, and it's hard to imagine what Bolt 1 or its future engine integrated version will look like.

Even if some of that progress can be back-ported (Graph Search, Graph Explorer, perhaps Fuzzy finder improvements) the core issues of the asset still remain - the variables system, the custom events system, lack of structure, performance issues due to the nature of reflection. Which, I assume, won't be entirely fixed by snippet nodes since at that point you have to either write your own snippets or use the already existing reflection. I'm also presuming snippet nodes won't approach native C# performance. Though a more streamlined and documented custom node API would be welcome.

We all knew what to expect from Bolt 2. While the overall goal for this directional change is communicated effectively, it's not exactly clear what's happening with Bolt 1. Is it staying largely as is? Are the variables, custom events systems getting a rewrite? If so, are current graphs forward compatible? Will we see structural improvements - something that would let us communicate between graphs besides custom events and let us reuse graphs more effectively without the need to redo variables for each object? If the core issues of Bolt 1 can be reasonably addressed with similar fixes Bolt 2 implemented, I'm sure it would be good enough for many.

 

The original intention of the blueprint was that c++ was too difficult for newcomers, but c# was much easier than c++, so c# does not require visual scripting. If you can't learn c#, it is not suitable for game development.
Based on the simple c# language, and using reasonable visual scripting to reduce the difficulty of advanced programming or complex programming, it is the core competitiveness that unity engine originally has. Please at least not reverse development.
All programmers will not use such visual scripts.
They will be disappointed with unity.

 

Bolt 2 wasn't as close to a production ready release as many people think here. It definitely wasn't coming out this year at the very least. So time frames might not be that different as one might think. They've stated that it'll be integrated in Unity 2021 so in a year.

And it took that one guy years to build those tools, too. They just started earlier than Unity.

 

What you're saying goes against what the options are though.

If Unity went with Bolt 2 it would take even longer. You want the best of both worlds -- less waiting and something akin to Bolt 2 and think cutting out uniformity between visual scripting tools would make up for that reduction in time.

What would you rather have?

• Come out sooner and be seamless across visual scripting tools, but simplistic and designed for non-programmers and only has basic level stuff
  
• Come out later and have a lot more power and be comparable to Blueprints and the ability for programmers and non-programmers to both use it, but it is more complex and might not be as seamless across all their other tools

Since those are the two options that Unity was faced with essentially.

Unity went for the former, which makes sense from their perspective as a business, but goes against what most of the community actually wants in the latter. The community wants something basically like Blueprints (as in both programmers and non-programmers can use it), but better node-to-code translation and for it to be part of Unity plus DOTS compatibility.

From what we've read it seems like that would be a lot of work and would be more difficult to maintain and unify everything so Unity is going with the easier option. Personally, I think that's a mistake and so do most people here.

I'd rather Unity take their time and work off of Bolt 2 and create something that combines the best of Bolt 2 and DOTS VS together. Instead their DOTS VS was similar to Bolt 1 and they'd rather just bridge that gap than rework a lot of stuff to make it work more akin to Bolt 2... =/

 

They started and continued for the years, they didn't give up after a year.

 

It's not giving up. Bolt 2 backend architecture does not support their vision for all VS scripting tool unification. For nearly two years Bolt 2 was developed in a sort of a vacuum where only Bolt 2 needs were recognized. This is why a lot of us were onboard - Bolt 2 considered visual scripting as a first-class citizen, and development decisions were made in accordance to this philosophy. So we got features like Bolt 2 Macro graphs which have no C# equivalent and C# gen that's still heavily reliant on core Bolt 2 tech but brings us native C# speeds in builds. It was never made to be DOTS compatible or to consider other VS tools within Unity. Bolt 2 was intentionally made Bolt centric for the benefit of us all.

But that philosophy is not compatible with Unity's (very valid might I add) goals. Hence a major change of direction.

 

I think the only way to keep the community happy is if you split up your visual scripting team into separate projects:

Project 1 - "Low level" C# generated visual scripting, similar to drop 6 on the DOTS visual scripting forum post. This should work in harmony with the C# layer you create for the mono and DOTS framework. Have an option for Mono graphs and DOTS graphs. You guys were getting a lot of praise for this direction and it seemed like the community was very excited about this approach. People even went out of their way to create in-depth video tutorials on the tool. I was also very excited as it was a visual approach to learning the "DOTS way" of creating a game. It also had code generation which was very impressive.This project could give programmers and designers finer control, while also having C# generation.

Project 2 - High level "snippet" node based visual scripting, where everything works similar to the blueprints model post Drop 6 in the DOTS VS forum posts. These high level nodes will work almost identical whether you have a Mono or DOTS based game. Basically the direction you guys are heading right now with the DOTS VS.

Project 3 - Merging DOTS netcode into both projects so we can get our games online. Easy event calls on Project 2 and more finer control on Project 1 (Dots only).

I am no expert by any means, but the current state of the forum posts for Drop 11 is pretty sad and it is looking like the valuable visual scripting user base you had, has disappeared or given up. There is less and less interest or feedback with each drop. Every drop since 6 has been nothing but the community complaining about the direction Unity is taking visual scripting, and Unity dev's essentially telling the community that their feedback is wrong. I personally do not complain in these forums, but anyone looking through those posts can see that there is a lot of resistance on both sides.

I honestly don't care too much about which direction you guys take, as long as this visual scripting tool is scalable, can work with the new netcode, and can use it to almost create a full game, with some custom C# on the side.

It is looking like there is a significant divide on people who want finer control over their graphs and a more "designer" friendly approach...might be best just to make both...and brand each one a different name or type.

 

You know, the reason why Amazon is one of the biggest companies in the world right now with the richest man on the planet as the CEO goes back to the early years when a younger Jeff Bezos said his entire company philosophy is "customer first." Everything Amazon did was in favor of their customers and what their customers wanted. That is why they are so successful, regardless of the other questionable stuff they do or have done that made people upset at one point or another since then. Their focus was making their customers happy.

Regardless, the point here is that basically every single person in this thread and in this discussion since your announcement wants Bolt 2. If you announce something and your entire community doesn't like it or want it, shouldn't that speak volumes in terms of whether you've made the correct decision or not? Look at the gaming industry as a whole right now and how many companies are doing the opposite of what their customers want, its completely ruining the industry and is just upsetting the people who already have dwindling faith that gaming wont ever be restored to its former glory. This is one of the main reasons your customers here even download Unity and try to make their own games, because the industry no longer caters to the gamers, its about shareholders. We love Unity and we want to make games that make us happy and that we want to play and we want to make those games using the tools we want without those tools being ripped away at the last second.

Your customers have spoken and they have given you a clear message in terms of what they want from your platform and from the acquisition of Bolt and Bolt 2, you can go back and read every single comment here and see it clear as day. What you need to do as a company is figure out how to implement everything and make it work, and if you can't figure it out then hire the original developers to work on it or just outsource the work to the original devs as contractors to complete their work.

You cannot just buy an entire project that thousands of people wanted that was so close to being completed and then just discard it because its difficult to do or has problems with difficult solutions. There are solutions to every problem, and you should listen to your customers because we are the ones that use your platform and are the source of your revenue in the first place. Just my 2 cents. Sorry for the essay.

 

Excellent choices, it's to be expected when one carries a handsome first name like yours.
What is a DLC graph? Here is what I'm thinking: it's a graph that you can inject in the final game as a DLC, which is good.
If it doesn't compile the graph to c# it's interpreted? How will it be accelerated? There are ways, like what Panda BT has done.
Keep the balls icon moving along graph lines and generalize to the entire Unity UI.
I spoke about that 10 bazilion years ago: is it possible to have all graph systems coexist in the same windows? For example if I want to change the color of a material or feed a cloud point at runtime to the VFX graph I'd just have to drop the reference to the object material or the mesh inside BOLT-u where it appears as a node with all input (and output when this makes sense) and I just connect the game logic node to that shader input or the mesh to the cloudpoint converting node.
For custom node, inspire yourself from Panda BT instead of inheriting Panda and overriding nodes. It uses attributes so if you want a MoveTo(Vector3) node you just add [Task] in front of the method and it automatically your graph thing extracts the method's signature as node input and output.

 

Whoever is making final decision, think about your carrier. If you choose Bolt 2 you will be remembered as "That one who dumped Bolt 2".
And in not so far feature, then Unity understands what scary thing they have done, I do not think anyone will want to have such decision maker on the board.

 

The problem is that it creates a fragmented user base, requires a lot more resources to create, and makes it a nightmare to maintain. I'd say just go for a unified Bolt 2 + DOTS single project that is designed for both programmers and non-programmers, creates C# code (despite the claimed that it isn't very future proof), and such. Although I'm not sure if that's easily possible considering they said: "The Bolt 2 architecture wasn’t providing us with the best separation of front-end and back-end needed to ensure a progressive transition toward unified UX with other graph-based tools. A good separation is also essential to providing DOTS support, in the future." They need something to compete with Blueprints or they will keep losing users to it.

I understand they are in a tough spot since they want something sooner rather than later, and they want a single unified UX so that all their tools work in similar ways, both for Monobehaviour and DOTS, which the structure of Bolt 2 doesn't fit with their current DOTS VS stuff which means they'd have to scrap that.

I'd rather they get it right the first time and take their time as needed instead of release something that isn't expandable and fragments the users into "non-programmer using Unity VS" and "visual programmer using some other tool like Bolt 2." Just have a unified solution even if it takes longer.

 

I think the biggest problem here is their vision of VS tool is wrong and because of this, all of the stuff they have announced, whether lack of c# code generation in real time or dumbed down approach towards VS is disappointing and again, wrong.

I have already said this, but I feel like I have to repeat myself. Bolt 2 did everything better than anything Unity tried with their visual stuff, whether we are talking about VS DOTS, shadergraph, VFX Graph. Bolt 2 was clearly designed by people who knew exactly what they wanted and this is because they used it, or at least tried it in more professional projects.

Unity should learn from the bolt 2 and correct their mess, not the other way around.

 

As harsh as it sounds, I have to agree with you. There's a reason many people prefer Amplify over Shadergraph and it's not just because of more features, it's also the user experience. You can feel when the tool is developed by a passionate developer who knows their stuff to the core. Definitely applies to Bolt 2 as well.

 

Agreed...

 

The community here has more Unity then the company with the word as its name.

 

This is what I believe Unity should do based on what I've gather in all this discussion.

1. Release Bolt 2 the way it was original intended and maintain it and/or open source it. This will be the default VS for the next few years.
2. Faze out Bolt 1 and replace it with Bolt 2 as it was originally intended.
3. Work on the unified solution behind the scenes. This will ultimately replace Bolt 2 and VS DOTs. You can take as long as you like for this so you can deliver the best solution.

I know Bolt 2 doesn't fit the unity vision that why I recommend you just take what you've learned from Bolt and DOTs and make your own solution from scratch. But don't abandon Bolt 2. It can be the default tool until you present your effective in-house solution.

That's my current feedback on this subject.

 

The way they are making DOTS VS atm, generating c# code will not look pretty. It's a weird mixture of Data-Oriented and OOP at the same time. This is where the mixed messages arise. If DOTS is so great, why create a visual scripting solution that is OOP on the surface but DOTS under the hood? I am pretty confident that there will never be C# generation in DOTS VS (unless they roll back to drop 6), because they have built their own systems to carry OOP event data using the graphs. This is why they will have to separate their priorities if they wish to even have c# generation in the future. They basically need to create something like uNode for finer control in both mono and DOTS, and keep going with whatever it is they are trying to accomplish with the DOTS VS for users that don't want that fine control.

And the weird thing is, they already have created a C# generator already. If you ever looked at the Graph Tools Foundation package, it is all there. Hope they can expand on it in the future. A node-based "Visual Studio" would be extremely powerful, and I believe it will make a lot of users happy.

@LaurentGibert I would strongly suggest hopping over to these forum posts if you haven't already: https://forum.unity.com/threads/dots-visual-scripting-6th-experimental-drop.795807/

This was heading in a great direction to teach people DOTS in a visual way. A little confusing at first, but it was a great educational tool.

You can see an absolute shift from the user base being hopeful and excited, to worried, confused, and disappointed.
https://forum.unity.com/threads/dots-visual-scripting-experimental-drop-8.857707/

The current forum posts for DOTS VS is a ghost town now. It looks like everyone lost interest.
https://forum.unity.com/threads/dots-visual-scripting-experimental-drop-11.944793/

I think the same thing is going to happen here. An educational, scaleable, visual C# generator is something almost everyone wants, please consider it.

 

As a Bolt user for its entire existence, and a Bolt 2 user for over a year now, it's been a rollercoaster, and the end is a little heart breaking.

That being said, Im starting to side with Unity. The more I read the plans and understand the design philosophy, the more I think the direction is smart. The real problem here now, in my opinion, is that the PR team goofed hard.

What you should have communicated was "We are stopping our work on Bolt 2 as a stand alone project, and instead are incorporating the best aspects of it, and Bolt 1, into our unified visual scripting system"

What you did communicate, what we heard, was "Bolt 2 is out, Bolt 1 is the future!"

 

I really like the idea of starting from scratch with the original Bolt. It'll probably be easier to maintain than Bolt 2, and therefore easier to integrate into the Unity ecosystem. The biggest goal I see out of all of this is to unify the teams working on Bolt and DOTS VS by creating a unified architechure backend and frontend, and making it easy for users to use both Monobehavior and DOTS graphs.

Unity can always recreate a lot of the same features Bolt 2 has over time. We'll probably see Bolt quickly gain most of the features Bolt 2 has by the end of 2021.

I have a few critiques. First of all, Lazlo had one vision for Bolt 2, and he was working on it with very large updates rolling out nearly every week. Alpha 6 was extremely impressive with the introduction of Classes, and later Alphas fixed several bugs and rebranded Class Definitions as Bolt Classes. The UX was completely overhauled by the help of a full time icon UX developer. The UX of Bolt 2 was truely game changing.

Ever sense Unity purchased Bolt, progress on Bolt 2 has slowed to a crawl. We didn't even know if Bolt 2 would be free for several months, and closed alphas didn't even start until just recently. Those on the Bolt Discord heard there was a few annoying bugs with Bolt 2, like the Odin Serializer, Code Gen, and various Mac issues. But we thought progress was going well. It seems like much of that progress was scrapped in favor of Bolt 1 for now, which is completely understandable.

But I know if Lazlo was still working on Bolt 2, we would be very close to a stable release by now. It's clear that Unity had a completely different end goal for Bolt 2 than Lazlo did.

I think Unity could have used all this additional development to improve on DOTS VS and make that more stable. We would have had a more stable version of DOTS VS and an almost production ready version of Bolt 2. Unity could have waited until Fall 2020 to officially merge Bolt 2 into the Unity ecosystem, while letting Lazlo and his team finish his vision. We probably would have had a better end product.

Another critique I have is the promise Bolt will be compatible with old projects for years to come. I've seen experimental packages change entire APIs in a big release, making a yearly version completely incompatible with previous years version. We all know Bolt is going to change a lot as it slowly transforms into Unity Visual Scripting.

Not to mention Bolt is extremely slow and outdated compared to Bolt 2 and other visual scripting tools. Yes it is technically production ready, but users are going to suffer through terrible editor performance and project breaking API changes until Bolt gets improved.

 

Just to add my two cents from my point of view (and probably not only mine, reading the forum), Bolt 1 comes with nice packaging and some cool features (live editing etc).
Would I use it in production in my company? No.
Bolt 1 flaws mentioned many times in this thread, exclude it from being used in mid / big projects (so the kind of projects, where you can find a lot of designers and artists).
As to Bolt 2, it was looking promising, extendable, manageable, like a tool that can actually be used in bigger projects in the future.
Personally I think that current solution is a little short-sighted, as it would look nice on release notes page and youtube videos, but how many games will actually use it?
Thanks for the open discussion!

 

Best Idea
And then make gradual, soft and seamless migration from Blot2 to new best unity inhouse solution

 

Its just that it doesn't make sense, porting the best things in bolt 2 to bolt 1, ok, but performance wise, if you don't port c# generation, its still extremely flawed. And why not just keep Bolt 2 that already has all the good stuff? Besides "its easier", whats the real reasons to keep B1? Bolt 2, in my perspective is a dream for visual coders, its the next big thing in the Game Engines Competition,... Sorry, but it doesn't make any sense to me.