Visual scripting in Unity

Hello I am a games artist.
My first published game was in 1997, in all the time since then I have never been able to code, I frankly hate it, I have spent many hours trying to debug even the most simple line of code, I guess a chap must except his own limits.That being said I am comfortable with highly technical process in lots of software and visual scripting.
I think Unity is fantastic but I don't understand why there is no in house visual scripting system.
Surely democratising games could include creating a solution for none coders, giving them the ability to express ideas and prototype without the need to learn a coding language.
I don't need to write a shader from scratch or create code to light so why would I have to code camera controls etc.
I know there is Playmaker and it is very elegant but the learning path is not part of Unitys learning path tutorials and so it makes the process difficult.
Just an observation, a personal view that Unity is too code centric .

 

You are welcome to try GameMaker or PlayMaker.

You'll find that ultimately any visual scripting tool is less flexible then straight code.

 

Visual scripting has been tried numerous times in the past by other companies but it simply never caught on. One part of the problem has been affordable tools, but a big part is simply that most visual scripting systems lack flexibility and trying to add it in often results in killing any advantages it had over a more traditional coding approach.

Blueprint is one of the more successful attempts at visual scripting, but at the end of the day it is still basically coding only with a different appearance. I definitely love the approach it has taken though.

Or uScript

 

I went to look at this for my kids, since the girls want to learn to script. Thought it might be a good start. However, it hasn't been updated in a while and there are complaints about lack of tutorials. Might not be the best to recommend to a non-coder. Playmaker is updated regularly at least even though it has some limitations.

 

Yes, I suspected it may have fewer available resources or it would have been closer to PlayMaker's popularity. I don't know anything about updates though aside from it received one three days ago (most likely to make it work with Unity 5).

 

Ahh, well not posted on their forum. They seem to have support issues.

Too bad because it looks amazing. Might even allow me to code. My only issue with visual scripting is that I haven't found one yet that worked with multiplayer. Playmaker works with Photon only. My son loves it though.

 

If I was going to teach kids about programming, I'd expose them to traditional programming ASAP. I wouldn't think kids would have preconceived fears about such things, unless they pick them up from their parents or something. The thing is, with PlayMaker specifically, right off the bat you'd be hitting them with concepts like finite state machines, states, events, and transitions.

To make a basic "click on Mickey Mouse to make him burp" game, here's how you might do it in PlayMaker, assuming you already have a GameObject with a visual element and a collider:

1. Select Mickey Mouse GameObject
2. PlayMaker -> PlayMaker Editor
3. Right-click -> Add FSM
4. Right-click -> Add State
5. Right-click on the new state and select Add Transition -> System Events -> MOUSE DOWN.
6. Right-click in FSM background and select Add State.
7. Click on the second state and press the Action Browser button.
8. Select Audio -> Audio Play (not to be confused with Play Sound)
9. In the Audio Play action of the state, assign the One Shot Clip to the sound effect.
10. Right-click on the second state and select Add Transition -> FINISHED.
11. Right-click on the transition portion of the second state and select Transition Target -> State 1.
12. Right-click on the transition portion of the first state and select Transition Target -> State 2.

Here's how you might code it (assuming you had an AudioSource with assigned sound effect on the GameObject):

Code (csharp):

1.  
2. using UnityEngine;
3.  
4. public class MickeyClick : MonoBehaviour
5. {
6.     private AudioSource _sound = null;
7.  
8.     void Awake()
9.     {
10.         _sound = GetComponent<AudioSource>();
11.     }
12.  
13.     void OnMouseDown()
14.     {
15.         _sound.Play();
16.     }
17. }

Is that really super scary? You'd have to explain what a Unity component is. There's no getting around that in Unity anyway though.

Don't get me wrong, FSM's are clearly super useful in games (I use non-visual ones exclusively), so I get the appeal of PlayMaker for those. I believe I read somewhere that Blizzard even used PlayMaker for UI states in Hearthstone. But they didn't try to write the entire game that way.

Nobody is going to be surprised if Unity announces an integrated visual programming tool someday. For now you have to choose Unreal/Blueprint or Unity and an add-on if you want visual programming. Doesn't seem like a big deal.

 

You never need to make excuses for using playmaker or blueprint. Some people seem almost apologetic when it's mentioned. There is no shame here. Node based game development is the future, on a wide scale.

 

Different people learn and approach problems in different ways.

I think Bret Victor (software engineer at Apple and a real genius, which pains me to say as much as I dislike Apple... he's the real deal) nailed it on the head with this talk:

Unity's Editor is edging closer and closer to this. Basically (and I am over simplifying this A LOT), they [Unity] would have to make a C# interpreter (there are a few out on the market already, some open source I think) that would interpret the C# to give immediate visual results (and vice versa), and then actually compile the C# when you choose to save your changes or build and run.

 

Since Asset Store Level 11 I am playing around with PlayMaker and I really like so far it because it does let you focus on getting the logic down. It still feels a little patchworky for Unity 5 features (like the new UI) but I really like it. Also I don't see any problem in using it either. Worst case: you find something that works in PlayMaker and rework it in more optimized code later on.

PlayMaker also works together with manual scripting.

(edit) Deletet a part. Actually I guess I should not judge about uScrip when I am not even actively using it at the moment.

 

wow what an amazing talk, thanks for posting that

 

Hi there
I am very Dyslexic and Node based workflows are a godsend. even a few lines of simple code are too much for me whereas I am very happy to use Houdini or create complex, but efficient shaders etc.As an artist, Senior and lead I have worked on many console titles but always with programers.
I met the chap behind playmaker at this years GDC, very nice man and we talked about the tutorials and the possibilities of there being more prebuilt kits but there is no firm roadmap for that.I have a copy of playmaker but because it is third party there ( for me at least ) is a disconnect between unity5 tutorials and old playmaker ones.

I guess that for many people who can code it is hard to understand people who can't .

I guess all I am trying to say is that for people such as myself who would like to be able to develop proof of concept stuff in Unity, don't care about efficancy it seems like a missed opportunity not to have Playmaker built in and part of the tutorial pathways.

it would be great to have the stealth tutorials done as code and as playmaker.

Any way thanks for everyones input

 

Has anyone written a ladder logic implementation for Unity? There is another paradigm I must see implemented and used. I could argue its half way between visual scripting and C# scripting . Yeah, lets not go down this path. As much as I am comfortable writing it, ladder logic is really the wrong framework to use within a game engine.

Ultimately you have two choices, use the functionality provided by Unity or a third party. Or write your own functionality. If you are unwilling or unable to do those, you are at the mercy of waiting for some other party to implement the functionality you want. In the case of tutorials there is nothing stopping you from rewriting the classic Unity tutorials in playmaker.

Or PM me. For the price of playmaker plus an incentive I might be persuaded to rewrite the classic tutorials.

 

You're welcome. It is an incredibly powerful message to all people - for software engineers it is a wake-up call to stop coding like its 1980 (and stop making people feel dumb when someone says a different way is easier for them and should be available to them for use). Just ignore those people on these forums. There are some good folks here too. Like with any community, there's some good people and, well, not so good people (of which I learned pretty quick when started engaging in the forums).

I tip my hat to Bret, and <grumble, grumble> Apple. In this case, we should "think different". While I certainly prefer coding, I can totally understand your position and reasoning, and thus can understand and appreciate why it would be important for you.

 

My 14 year old daughter learned how to Mod Minecraft in Java. I don't think she is afraid of programming. However, something like Playmaker can introduce them to the logic of programming. My son started with Playmaker and he will be a junior in college studying computer science this fall. Also, my husband programs in C++ and C# so no fear from us.

Sometimes it isn't about being scared but more about seeing things visually. A lot of people, myself included, are very visual people.

 

Built-in visual scripting would be cool, but it doesnt seem like it will ever happen. A lot of people compare uScript to Playmaker but its an obscure comparison since they're different types of tools. They're both visual, but fundamentally different... Playmaker does have a good reputation and is a great tool that offers a lot of workflow gains with minimal tradeoffs.

There are somewhere around 100 tutorial videos on the YouTube channel and the vast majority of them are still relevant for the latest version, even the oldest ones. We try to keep up and add annotations to notify of clear errors/problems... We could probably stand to add 100 more... I think the real dent here is that everyone has different expectations of what tutorials should be. The range of feedback I get on new tutorials makes it kind of weird to scope out future tutorials.. I think the bulk of users just expect regular updates while a smaller more vocal group is more aggressively critical and undecided about what they actually need. Maybe I over think it.

 

Well said.

 

When I write code I feel like I'm writing a living novel book of precise instructions!
I'm so used to write/read code, visual thingies are a nono for me but I don't hate them either.

 

There is also Antares Universe in case you guys didn't know about it.
Also, uScript developers are still quite active - they pushing out updates at least every 3 weeks, often times sooner. There were a lot of discussion comparing them (PlayMaker vs uScript vs Antares Universe) on this forum, just search for it.

But nowadays, there is a whole section on Asset Store dedicated to Visual Scripting. There are just too many choices, each catering to different need / understanding / level of code proficiency of the user. Node Canvas, Blox 2, Behavior Designer, uFrame, iCanScript.... the list goes on and on.

You just need to pick one you like and start using it. Visual Scripting helps you in understanding how to code (at least for me), clarifies relationship between objects, and ultimately turn you into a real coder!

 

So what you are saying is I should write that ladder logic editor extension.

 

That's cool. Kind of sounds like you are intimidated though? Does your daughter or son prefer visual programming to traditional programming? Aren't there some pretty severe limitations with that?

My concern would be the potential trap of treating *everything* as an FSM. When you have a hammer, everything looks like a nail. Some things lend themselves to being an FSM and some things don't. It also seems like doing anything remotely complex would get unwieldy very quickly. Am I wrong? Would you care to share a screenshot or two of your FSM diagrams doing anything beyond the trivial?

I admit I'm very much on the opposite site of the fence. The visual nodes thing just seems so time consuming and restrictive to me. The thing is, when you're creating a custom designed game, you often need a degree of flexibility.

 

Some people are good at abstractions some people are better at visual interpretations.
However there is still a need for understanding programming logic even with visual tools.

This is in short what I am saying.

For some learning a programming language might not fit their time schedule or interests, artists need to and want to spend their time doing what they love.

This is because you are awesome at programming. it interests you , and you have been doing it for some time.
For a Visually oriented person (artist) the completely opposite would be true. they would not have the same needs, or knowledge of the flexibility that traditional programming can give you. As long as the score is updated character moves and things pop up when they need to its ok.

Exactly.

For now what we have is playmaker ,PLYgame, uscript etc to cater for the visually oriented.
And I would not be surprised if the animation state machines evolved into something interesting.

Making games is not easy , but rewarding. Conquering difficulties is what we do every day.

 

Try 'Blox 2'. It's basically 'Scratch' inside of Unity. With some keyboard input it could be better than text AND faster.

 

They actually DO do this. Or... WERE going to do this. The project 'Livity' is AMAZING. It's made by the guy who made Boo and Unityscript. It's exactly as cool as that guys talk (without the time scrubbing stuff) and can even use Visual Studio. BUT it's been abandoned. Why THE HELL Unity don't continue it or do something like it is beyond me. I bet they'll bring some clumsy, slow node based editor soon, which won't be as good as Blueprints (which are also bad). Just combine Livity and Scratch (or Blox 2 in Unity) with live highlighting of code and it'll be the ultimate.
Graphical editors can better than text in every way (yes, including speed) if designed well. They can even write actual code instead of being a slow layer on top (Blox 2 does this).
Unity needs editor support for actual C# events and listeners etc. It's incredibly clumsy to set up now for anyone, let alone beginners (and no, the new UI's event system is only slightly less clumsy)

 

I use my own C# delegates and events all the time in my Unity projects. What problems are you experiencing with them?

EDIT: Ah, you said editor. Hm. I do have some of my own custom editors but I haven't had a need for using events, although you do bring up an interesting point, because if that is possible, then stuff like this could become fairly easy to do.

That's pretty cool, I didn't know that. Still, it is theoretically possible with a C# interpreter. I'm sure some of the IL2CPP gurus could probably take a look once they get IL2CPP on the platforms they are targeting for it.

 

Why the negativity about something you don't know?

Bret Victor's talk was really inspiring and I think Unity even succeeded in a lot of ways already. Of course there is always room for improvement and I am surely one of the first to jump for joy when the programming barrier gets lowered further and further. I mean I get around scripting nowadays but it's clumsy and I know I am not great at it and don't have the time to become it, either.
Then again - think about exposed variables on playmode. That wasn't there before. The developing environment Unity provides is so much more and better than what it was before Unity.

You have a system you like and think is cool? By all means promote it! Make noise about it. I know I hadn't heared about livity before. Maybe even give Unity the spark for a potential system of their own to implement in the future?

 

Hi,

I have playmaker and I can wholeheartedly recommend it.
Being a compositor for 11 years, I'm an evangelist of node based workflow.
Playmaker didn't take long to pick up at all, and I can solve most problem without touching code, or at worst case find a workaround.
I just discovered Echosystem as well, which is a plugin for playmaker, which is basically just a search engine for more commands that have been added by users.

Though, I realize that at some places my lack of programming skills is a problem, so it's not all I would wish it to be, but most of it.

 

Something that I like about a workflow with code and visual editors is that programmers can shift more of their work to the creative crew. In our last project (Forgotten Memories) I've wrote every little single functionality in code (and it wasn't funny). It would have been better to write only core functionalities (such as low features and behaviors) and let designers/artists glue them visually. Gladdy, we are currently shifting our workflow towards a programmer-designer-artist centric approach, instead of writing everything in code. That way, it offloads programmers workload and gives designers and artists more creative freedom.

 

Anyone know if there are any visual scripting solutions that will work with networking and multiplayer? I have looked and asked but have not found anything. I would be happy if a good C# programmer, we have two on our team, could tie something like Playmaker into our code. We are using uLink at the moment. I guess the other question would be whether any networking solutions work with Playmaker, other than Photon, which is just too expensive for us.

 

I'm making Bolt / Playmaker integration. Should be available in a few weeks if all goes well.

Bolt is still beta though, not sure if you're comfortable working with a beta networking system in a title...

 

That's why I'm annoyed. Because it WAS Unity's project.
There's no reason to have this 'real programming with text' or 'node based programming for artists' split. It's entirely down to the editors and how terrible text editors are at the moment. But Node editors completely throw out any good stuff text editors do. Baby with the bathwater type situation. Programming is still in the 'POV ray' days to me and I don't see Unity caring much about workflow and innovation anymore, which is a real shame, because that's what they were all about. They seem content to just copy whatever everyone else is doing, which isn't what made Unity great.

 

Limitations you can overcome if you modify existing code / write custom nodes

 

What I'm saying is: You don't know what Unity is currently developing. You asume but you can't know. So why asume that it's worse than livity?
Basically while livity seems really cool I am not sure that it really does lower the actual barrier of programming altogether that much. It still is programming. And I don't know why node editors are supposed to be a bad thing either. There can be both.

I mean - imagine the following: A node editor that works in a way like PlayMaker does - a lot of nodes for the most useful features you can connect and create gameplay with. A node isn't specific enough for you or you want to have a completely new node that is specific to your game? Throw an empty node in there and on a double click it expands like a text editor in the node view. Imagine being able to edit the node live while your game runs to see how it works in conjunction with the existing nodes or your node graph.
Also I think encapsulating nodes for certain things could be benefitioal as well (doesn't blueprints already do this?).

There certainly also have to be more ways I hadn't thought of yet. before this thread I actually never really tried thinking if there were better ways to do things or to "think outside the box" if you want to use a worn out catchphrase. Lively looks cool and it's more direct like the examples Bret Victor showed in his talk. But as much as I can see its benefits - as a non programmer I hardly find it that much more intuitive than what we have now.

 

I occasionally still get to work with ladder diagrams.

I occasionally get to debug ladder diagrams.

I know you're not really serious, but I have this urge to bribe the Easter bunny to visit your house and make sure that extension doesn't see the light of day.

 

That would be sweet. I've toyed with the idea of using PlayMaker as a way to do the equivalent of "mission scripting", but even within that narrowly defined scope, I think that'd be too limiting. There are times when you need to do something that only applies to one specific mission. If you could kind of bridge the two worlds like that, it'd be a lot more appealing for use in more complex games. Also, something more along the lines of behavior trees might be more flexible than the strict FSM-only approach of PlayMaker.

 

Ladder logic is still a regular part of my day job. Its actually fairly decent if you have large numbers of I/O to deal with, and relatively simple logic connecting them together. Its still totally valid to write new automation logic in. Most of the new equipment I purchase is still written in ladder logic.

But if we are going to complain about debugging archaic technologies, I still have mechanical PID controllers in the field. Debugging gets real nasty when you have a combination of ladder logic, pneumatic logic, relay logic and mechanical logic. Plus miles of old cabling connecting the whole thing together.

But for a game your I/O count is relatively simple, and your logic is very complex. I agree, ladder logic would suck. I would still like the Easter bunny to come, I'm in need of chocolate.

 

TIL mechanical PID controllers are a thing.

Hmm, I wonder if anyone will make something closer to LabView.

 

Has anyone here tried the Blox2 visual scripting asset? It looks like it is modeled on Scratch and potentially could be useful but there is almost no feedback on the store comments. I have Playmaker but would prefer something closer to the Construct2 event system to do some prototyping in Unity.

 

Yeah I've tried it, it's good. The variables are a bit confusing but the best thing about it is it actually writes C# in the background, which you can convert it to normal code

 

Blox looks really cool as well.

It sure is closer to the traditional programming style but I could really imagine an amalgamation of nodal approach and Blox for custom. Pair that with mentioned instand feedback ... holy S***!

 

Reminds me a lot of LEGO's initial robotics programming tool.

NodeCanvas doesn't seem to have been mentioned yet. Only noticed it myself when the store put it on the main page.

 

Thanks for the reply. I will definitely have to try it out soon. Playmaker's flow just doesn't jive with my brain, as I prefer something more linear and tweakable. The 2 best approaches to visual programming I have tried so far are the event system in Construct2 and the Blueprints system of UE4. An approach that combined these 2 would be pretty mindblowing!

 

https://bitbucket.org/dklompmaker/nodify

 

Blox is Blockly
https://developers.google.com/blockly/

 

I really like blox2 and plygame

blox2 is just awesome. it reads out the api of every third party extension (like ngui or other stuff from the asset store) and creates blocks out of that. and you can convert that into real c# files. also the blocky ui makes more sense to me than this node stuff that playmaker has for example.

blox1/plygame is easier. i personally prefer this version (its in plygame included) since it contains pre-coded blocks which takes away a lot of work from your side but on the other hand its not sooo powerful as blox2. its not a "old version" of blox2, cause its something for itself but has the same ui/workflow. i use it to do minigames like hammer throwing, jetpack riding or games like flappy, crossy or retry. with plygame comes the powerfull plyrpg extension and diaQ which is nice for any kind of rpg and dungeoncrawler. but iam sure you can do any kind of game with plygame but it gets more complicated the far you are moving away from the rpg mechanics. from easiest to hardest to make with plygame i would say is:
rpgs/dungeoncrawler
sidescroller
minigames/physicgames
shooter
rts/tbs

i tried uscript before which was to weird &powerfull for me (which is a problem with me, not with uscript)
i tried playmaker which was pretty okay but i totally cant stand the node based structure. the big advantage is the huge community and a lot of extensions in the Assetstore. at this part blox/ply needs to grow a little bit but the community is super helpful and all my questions got answered

 

What do you think about Nodify?

 

its pretty new right? al least i can not find that much about nodify.
somehow asset/software creator should be aware of name-giving to their products. you actually can not google for nodify unity since google thinks you want to search for "modify unity" and also youtube gives you a lot of stuff about how to modify stuff in unity...

never the less... it looks interesting, again node based which i see as a disadvantage. i thinky Blockly https://developers.google.com/blockly/ has the petter UX approach.

but a big plus for open source. but still eveybody can write extensions for all the visual scripting solutions from the asset store. so its not so important to have it open source

 

Actually, i would like to see a ipad version of unity. where you can just play around with cubes and spheres and the "scripting" ui would be something like this:

and then upload it to the "real" unity and continue there. for prototyping in the train or something like this.

or how blocksworld is doing it: https://itunes.apple.com/us/app/blocksworld-hd/id665886336?mt=8
and then just with the opportunity to load it into unity

 

The visual scripting is innovative. I think the visual scripting is future
The problem is performance Especially mobile

I'm interested in the playmaker and icode

Is there anyone who make mobile game using playmaker and icode ?

tell me your experience please~

I wonder Mobile Performance