playMaker

Author Topic: Playmaker 2.0  (Read 4052 times)

djaydino

  • Administrator
  • Hero Member
  • *****
  • Posts: 4586
    • View Profile
    • jinxtergames
Re: Playmaker 2.0
« Reply #15 on: August 09, 2019, 06:01:32 AM »
Hi.
Actually you can 'kind of' do this already with 'run fsm'

Maybe its an idea to make it possible to select a group of actions and have an option to create a template from it that automatically set up inputs/outputs/events.

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #16 on: August 18, 2019, 03:22:30 PM »
Any news on this?

At this point we can only assume PM2 comes out around the same time as Half-Life 3 ;D

I wouldn't be too impatient about this. PM2 really ought to be "something" imho, much more than a mere update.
That's the kind of release that could provide a really fresh face to the tool: more modern and dynamic UI, content-changing actions depending on preselected options (they added a functionality tilting towards this in 1.9 with a new UI Attribute but it would require a big overhaul on many actions), extra buttons and floating/collapsing windows/tabs within the graph view,more tools and a better visibility on the management and tracking of events and vars, massively improved Globals, expanded official action packs (some have become classics by now), etc.

PlaymakerNOOB

  • Full Member
  • ***
  • Posts: 205
    • View Profile
Re: Playmaker 2.0
« Reply #17 on: August 27, 2019, 08:46:11 AM »
Quote
I wouldn't be too impatient about this. PM2 really ought to be "something" imho, much more than a mere update.
That's the kind of release that could provide a really fresh face to the tool: more modern and dynamic UI, content-changing actions depending on preselected options (they added a functionality tilting towards this in 1.9 with a new UI Attribute but it would require a big overhaul on many actions), extra buttons and floating/collapsing windows/tabs within the graph view,more tools and a better visibility on the management and tracking of events and vars, massively improved Globals, expanded official action packs (some have become classics by now), etc.

With built in visual scripting in unity arriving next year, it would make sense that Playmaker would need to do at least what the built in scripting does but better.  They need to be better in every regard to stay relevant.  I mean, why would new users choose a paid asset, when the built in one is free, more efficient, has all the 'actions' built in, etc

tcmeric

  • Beta Group
  • Hero Member
  • *
  • Posts: 733
    • View Profile
Re: Playmaker 2.0
« Reply #18 on: August 27, 2019, 10:25:12 AM »
Unitys new visual scripting solution is for ECS, not c#. As well, it is not a state machine. So those are some pretty huge differences. Lastly, if you check out where it is at now (beta is out), its a long ways off from something like Playmaker, Bolt or other assets on the store.

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #19 on: August 28, 2019, 02:59:18 AM »
I find it interesting that the image on the asset store features rectangles with rounded corners but the states have pointy ones in Graph view.
It's essentially Metro vs iOS, the former being the style used in PM, the latter being featured in the asset store's picture.

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #20 on: September 01, 2019, 12:52:49 PM »
Unitys new visual scripting solution is for ECS, not c#. As well, it is not a state machine. So those are some pretty huge differences. Lastly, if you check out where it is at now (beta is out), its a long ways off from something like Playmaker, Bolt or other assets on the store.

https://forum.unity.com/threads/unity3d-getting-visual-scripting-in-2019-2.577084/
Unity was initially planning to push hard for ECS, but they put it on a back seat for now. In a distant future I can even see them switching from C# to their DOTS (in-house ECS) once and for all. C# will be legacy stuff.
The moment they'll reach a situation where both their C# and DOTS are equally efficient, you know this will last only a few years before they make ECS their main coding backbone.
There's probably up to two years tops for them to finely tune DOTS before a proper stable release, and maybe four to six years for DOTS to reach some kind of technical parity with C# after said release.
The 'Big Switch' might happen in a decade or so, or maybe two years earlier, but not before.
If Unity pushes for DOTS, it's just a question of time before devs feel "forced" to move on.
Then the C#-based VS tools could start to become irrelevant, but that will take several more years because of all the workers who have accrued experience on C# and will have a hard time changing their habits. Only promises of greater performances will eventually force people to jump.
But nobody will create an ECS VS tool within Unity since it will already be available. Only new nodes and functions will be sold in the store, under VS.
Best case scenario? They drop ECS altogether. Okay lol not gonna happen.
Realistic best case scenario for third party Unity VS tools? I'd say a large decade of profits before being shadowed.

PlaymakerNOOB

  • Full Member
  • ***
  • Posts: 205
    • View Profile
Re: Playmaker 2.0
« Reply #21 on: September 01, 2019, 01:05:56 PM »
That is a surprisingly fair and detailed response.  I'd say your scenario seems extremely likely for Visual Scripting tools.

tcmeric

  • Beta Group
  • Hero Member
  • *
  • Posts: 733
    • View Profile
Re: Playmaker 2.0
« Reply #22 on: September 01, 2019, 05:40:34 PM »
Switching from C# to DOTs only, would mean .. well.. the end over everything in the asset store that is not a shader or model just about. The asset store is clearly a large source of income. It will be a long transition.

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #23 on: September 02, 2019, 01:34:06 AM »
That is, if their homemade ECS variant (like C# Unity was their variant of C#) proves to be that much of a gain, but it really looks like a paradigm breaker.
Although the change will be gradual and progressive, obviously the people remaining on the "old" tools will dwindle in numbers, but the real question is if this will only be a reduction towards a more focused niche or a pure and simple extinction.
I'm not sure if I will want to bother learning a new set of tools in 15 years from now. A super solid and matured Playmaker (or else, like Bolt) will be equally nice.
Third party VS tools might as well try to furnish their own sets of ECS-compliant nodes/actions and a lot of the support will come from their respective users. But I suspect many coders will flock to Mother Unity and pamper her with new scripts to enhance her own VS tool.

Now Unity has something that plays against itself; its business model. They largely provide a core system but you need to buy tons of options to make it palatable for proper devlopment. Which means a tool acting as a superior layer and providing more functions than the vanilla system will certainly have an edge, especially if done early, that little extra something to sell.
This is where tools like Playmaker, Bolt, FlowCanvas, etc., can find a solution to their survival, but they really need to provide that extra thing that Unity will likely not provide, and think of it now and quickly.
Like, for example, providing greater cross compability with more external tools/libs, but that's a hard thing to do and requires money. Usually it's the devs of these assets who chose to build those bridges back towards VS tools. Or perhaps an easier output of WYSIWYG semi-interactive videos for test and other statistic-related outputs. Just throwing ideas here, but I think it's fair to say that the threat is pretty real.
If the VS tools we have now get closer to perfection and find something to leverage right now, they can become essentials and maybe secure themselves alongside Unity. That's all the best we can wish them, right?

Playmaker was built as a FSM system, but creating new actions requires coding.
What if they actually provided a node-based Action Creator?
Bolt is moving in a way as to allow the creation of metanodes that literally act as states. In other words, soon enough Bolt will have a large catalogue of metanodes (implied states) to download from. Playmaker already has this catalogue but no way to create actions in Visual Scripting.

Then, only robust parallelal "co-engines" (sort of) such as Playmaker, uNode, FlowCanvas or Bolt will survive if they have a massive community, are powerful and up to the task, with also a massive selling point that Unity could not hope to obtain. Bolt (Ludiq's) are very organized and aggressive in their business plan and communication. You just need to take a look at their website or their tool's interface, it smells dedication, money and modernity. Yet not even that guarantees that they'll survive in 10 years from now.
« Last Edit: September 02, 2019, 01:42:55 AM by Broken Stylus »

djaydino

  • Administrator
  • Hero Member
  • *****
  • Posts: 4586
    • View Profile
    • jinxtergames
Re: Playmaker 2.0
« Reply #24 on: September 03, 2019, 02:54:12 AM »
Hi.
The Comparison is inaccurate to push their app in front :

Reusability :

PM has Templates.

Types :
Custom types can be accessed

Variables :
Object Variables can be used
Scene Variables (Global Variables?)
Saved Variables (playerprefs although this is very limited)
Data Binding (get/set properties)

Performance
Lambda Optimization : Playmaker does not need this as it uses dedicated actions which are faster than mirroring (except for Get/Set Properties actions)
JIT Compilation : Same as above

Debugging
Data Visualization : debug mode shows data inside the actions and you can also set variables visible in inspector.
Predictive Analysis : Required fields show red when not used, also when components are not found you get a warning.

Especially the performance part is so deceiving...

In 2.0 they are working on a solution tho, which generates scripts.
Which is actually interesting, as maybe its possible to have something similar in PlayMaker to create custom actions (if generating works with 3rd party assets)

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #25 on: September 03, 2019, 03:42:23 AM »
https://forum.unity.com/threads/unity-visual-scripting-2017.462181/
Check the 2017 video
Beegees Space Pirate says:

Quote
"There are a lot of different visual scripting solutions out there, but none of them really work well all the way from prototyping, maintaining, debugging and deploying high performance shipping-quality visual scripts."

I know this was a statement made in 2017 but back then PM already provided a good option for prototyping.
Maintaining is highly dependent on Unity's own set of tools regarding team collaboration. PM's internal issues regarding the management of Global variables could be held against PM though, but it's not that hard to maintain a PM project. In fact with something like FSM log, PM provides a native tool to follow the evolution of a script. FSM timeline is a nice tool although I think it could need an extra option to filter FSMs that talk to each other (it would need to parse the get/set actions, send event to X, goto y state in z FSM, etc.), maybe even presented under the form of some horizontal branching tree too. So that would be a little overhaul of the Timeline tool.
Debugging is rather transparent and quite easy to do, for the same reasons explained above. Several autohelpers are present too. When PM is fully deployed there is just no way you cannot find the little nagging bug that's creating a problem.
As for shipping-quality content, ready for deployment, it boils down to what you understand by that. Is it a 100% made in VS app (game or not btw), or is it something that integrates VS to some extent? Obviously a large "made with [name of VS tool]" catalogue would help bring confidence into the possibilities of VS based development, but a lot of this relies on PR and heavy communication, from web site to video content, fairs, webinars, etc.

Also, regarding Unity, they did announce a big move into VS back in 2017 and then postponed it. We're in 2019 and they sort of seem to imply a renewal of the interest in VS but are again pushing it to a later date.

Finally, to see what was being said on the other side of the river, a reply about this VS tool Unity is working from Bolt's main coder.

I'm wondering what kind of beast Unity will become, they are literally working on two engines right now, the classic one that's in the DNA of the tool since its inception, and the new ECS based one that's friendly towards multithreading, which would usually concern large budged indsutry games or software, to manage countless entities at once. They want to get out of the indie zone.

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #26 on: September 03, 2019, 04:00:33 AM »
It's also possible that node-based VS is a fad for gullible morons.  ;D

jeanfabre

  • Administrator
  • Hero Member
  • *****
  • Posts: 14593
  • Official Playmaker Support
    • View Profile
Re: Playmaker 2.0
« Reply #27 on: September 17, 2019, 02:00:35 AM »
Hi,

 c# is not going anywhere anytime soon :) ECS is not for the mere mortals amongst us, it's a beast, it's completly out of any visual elements to help you, so only hardcode devs and studios with big budgets and AAA goals will afford ECS for their projects. Unity might in the coming years leverage the editor interface with ECS.

Bolt is in trouble, despite their amazing look and UI, because it basically doesn't perform as expected since everything is using reflections, they work on Bolt 2 which will create scripts, but their take is basically a visual c# editor which completely misses the point of visual scripting for artists and unity dev who don't know c#. So I don't see that as going anywhere in terms of concurrent as to what PlayMaker offers.

Visual scripting solution that generates scripts will always struggle with workflow because it takes extra compilation time to hit play, that's annoying and break the flow of development, even now I tend to just do things in PlayMaker just to avoid the compilation of a small change in a script. You get used to this and forget all the amazing things playmaker brings on the table, other then the visual aspect of it. Burst compiler may help, but only time will tell, right now, it's a big nop...

 It boils down to this:There is no perfect solution, but there is a sweet spot for everything and Alex nailed it, almost 10 years ago now!!! with its hybrid approach of visual fsm and c# custom actions. It's the best possible combination for NO performance penalty and visual scripting. But it means extensive support to provide custom actions, proxies for third parties etc, that's the downside and a real problematic one, which is why I think Unity can't provide its solution, because it knows about this above dilemma ( perfs vs man power ), they can't settle for any of the two. The only thing I could see happening is a custom action wizard, on steroids, that artists could use to maximize the reach for third parties without coding. But when I work on Cinemachine api, I understand this is isn't going to work across the board... some api are too complex for wizards.

Then, it's of course a matter of preferences, and it's great that so many publishers make their own visual scripting system, it's very positive and make things move forward.

Bye,

 Jean

Broken Stylus

  • Sr. Member
  • ****
  • Posts: 279
    • View Profile
Re: Playmaker 2.0
« Reply #28 on: September 17, 2019, 08:11:00 AM »
The only thing I could see happening is a custom action wizard, on steroids, that artists could use to maximize the reach for third parties without coding.

What do you have in mind? Is this for Unity or Playmaker?
Since creating new actions is a roadblock for non-coders, do you think Playmaker might provide a solution to this in the future?

krmko

  • Beta Group
  • Hero Member
  • *
  • Posts: 992
    • View Profile
    • Fat Pug Studio
Re: Playmaker 2.0
« Reply #29 on: September 17, 2019, 12:13:05 PM »
I don't think custom action wizard will ever happen, though making actions is quite an easy feat, it's actually complex to make it automatized.

When you already have a script that does something and you have full access to it, making actions is dead simple with only basic coding knowledge.