Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

General Purpose Data Driven Finite State Machine Plugin #629

Open
PradeepKumarRajamanickam opened this issue Oct 5, 2020 · 8 comments
Open
Labels
A-UI Graphical user interfaces, styles, layouts, and widgets C-Feature A new feature, making something new possible

Comments

@PradeepKumarRajamanickam
Copy link

PradeepKumarRajamanickam commented Oct 5, 2020

bolt
Ludiq Bolt

It would be nice to have a robust, high performance data driven fsm with an easy to use visual editor similar to Ludiq Bolt on Unity. It will move various kinds of control flows out of code and into data.

Benefits

  • Makes large scale projects or games with complex flows easier to manage
  • Reduced compile time on large projects since graph changes will be in data no need to compile code
  • Reduces barrier of entry to make games due to visual nature of the tool
  • It can be used to build a number of other content authoring systems on top of it.
    eg, Animation State Machine, AI FSM, UX State Graph
  • Content creation (related to flow eg. UX, AI...etc) can be distributed to non-programming team members
@memoryruins memoryruins added C-Feature A new feature, making something new possible A-UI Graphical user interfaces, styles, layouts, and widgets labels Oct 5, 2020
@PradeepKumarRajamanickam
Copy link
Author

I think I might take a shot at it once I wrap up input map plugin development.

@sim-the-bean
Copy link
Contributor

Would the "general purpose" aspect of this allow the engine to control the scheduler, i.e. conditionally execute systems? Or is this purely for game logic? If the former, I would love to help on the scheduler aspect of things (and really anything that has to do with this)

@PradeepKumarRajamanickam
Copy link
Author

PradeepKumarRajamanickam commented Oct 12, 2020

Would the "general purpose" aspect of this allow the engine to control the scheduler, i.e. conditionally execute systems? Or is this purely for game logic? If the former, I would love to help on the scheduler aspect of things (and really anything that has to do with this)

Yes in order for this to be user friendly or even possible, we need the ability to add and remove set of systems based on the current state. I am really interested in the idea of providing these add/remove systems api's in Command as mentioned in this discussion #654 (reply in thread). This will make things a lot easier.

This tool is mainly for content authoring high-level game/app related stuff eg, ux flow, gameplay, ai, animation fsm...etc. I believe using visual tools for these kinds of high-level game-specific content makes sense. Since maintaining real world projects with large complex flows via code may not be the best way. It is much easier to follow a visual graph than code when it comes to flows.

I will be penning down my thoughts and share them with the community first. To see if there are any technical red flags, regarding implementation that may not be feasible. But this is gonna take some time (need to get input map to a state it can be merged with bevy master first). If you are still interested when that happens I would love to have your support on this.

@PradeepKumarRajamanickam
Copy link
Author

I am working on an idea.

Brahma Overview
https://hackmd.io/@-EznnAUiQTygOuhjfvLeMg/SkgCxkQwv

@zicklag
Copy link
Member

zicklag commented Oct 23, 2020

@PradeepKumarRajamanickam I was just thinking that it would be great if we could abstract the FSM and other logic into a crate that could produce some intermediate representation that would be translated to Rust code in a separate stage. Also it would be great to keep the logic separate from the UI. This would abstract both the input ( the UI ) and the output ( the code) so that they could be swapable.

My primary motivation for this is that this kind of functionality is something that I would love to integrate into Arsenal. In Arsenal we would be using Blender's node system to power the UI and we would compile not to Rust but to a scripting language, probably either Lua or Python. Therefore we would have a different frontend GUI and a different backend code target, but everything else would probably function the same.

@PradeepKumarRajamanickam
Copy link
Author

PradeepKumarRajamanickam commented Oct 24, 2020

@zicklag

"we could abstract the FSM and other logic into a crate that could produce some intermediate representation that would be translated to Rust code in a separate stage".

I was wondering the same thing but I would just provide a serde helper in brahma_graph to serialise/deserialise graph to ron and generate *.rs src only when the user explicitly hits a "compile" button.

" Also it would be great to keep the logic separate from the UI"

You are right there should be a separate glue layer on top.

I have updated the diagram have a look.
https://drive.google.com/file/d/1G9RhtkSfbEFUTd3AHuTt1EvKwyjT7fDA/view?usp=sharing

@PradeepKumarRajamanickam
Copy link
Author

PradeepKumarRajamanickam commented Nov 22, 2020

@Weibye
Copy link
Contributor

Weibye commented May 19, 2022

It would be nice to have a robust, high performance data driven fsm with an easy to use visual editor similar to Ludiq Bolt on Unity. It will move various kinds of control flows out of code and into data.

I think before a fancy visual node-editor statemachine is implemented (I really would want that) I believe Bevy needs a working Finite State Machine implementation that can serve as the building block for many other features. Then later that can be extended to incorporate visual editors.

What is the current status of this @PradeepKumarRajamanickam?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-UI Graphical user interfaces, styles, layouts, and widgets C-Feature A new feature, making something new possible
Projects
None yet
Development

No branches or pull requests

5 participants