Unity Visual Scripting 2017?

One example of a real world case, Unity as an educational tool.

I work at an education provider. On campus we use Unity for the Programming and Design students, because it gives a better understanding of what is happening behind the scenes and building things up from a more bare bones state.

The art students (at least initially) use UE4 because it just works. The visual node based material editing is similar to how they go about creating materials in programs like Maya, and lets them quickly set up their models looking awesome. Going from there, adding simple behaviors like triggers, opening doors, flickering lights etc is a simple step for them because it builds on the same skills they learned in the material editor.

We also work with high school students, and in most of the schools they use UE4 because through trial and error, it has been repeatedly borne out that those students find it quicker to pick up the visual scripting and get fairly complex mechanics working than if they have to learn C# first.

When they are specifically learning programming, it's totally Unity all the way, but when they are engaged in developing all aspects of their small game projects (design, art, animation, scripting) then UE4 just allows them to get up and running quicker. Aspects of the blueprint system like being able to see the glowing path of how the script is executing can't be overstated in how much they help students quickly grasp the logic of what is going on.

To specifically address the "what's wrong with the stuff already out there" question, it's the fact that most of the plugins for Unity aren't practical. It's a tough sell to try and get a school to fork out lots of money on Playmaker or Shaderforge licenses when UE4 already has that functionality built in. Then there is the problem of students working on projects at home. A Maya/UE4/Unity workflow is great for students because they can get free versions of the software at home. Once you introduce paid plug ins you get potential licensing issues.

I generally prefer writing C# directly. However it is a fact from my experience that the lack of UE4 style blueprints and material editing results in end users choosing UE4 over Unity. Is it enough to be worth the investment to develop those features? No idea.

 

I agree with you 100%.
Unity needs to do the same thing and build proper systems for all the disciplines of a world class game engine. The asset store is not a silver bullet or a all in one solution. Unity needs to include materials, ai, visual scripting and whatever else needs doing to ensure a long and healthy future. The reason we are facing so much opposition and negativity is because those same individuals are so conditioned to the way is now they can't see the future even if it was standing in front.
UE4 uses a combination of c++ and blueprints and smart game designers understand why this is so. Their are times when when coding in c++ is valid and others when is just stupid to do so. Game logic for the most parts all done in bp and heavy math functions if required are coded and that is the way.

If you're in a big studio then artist, programmers and level designers will each play their roll, but if you're a one man team starting out trying to build a game than its pretty staggering to wear all those hats. Overcoming many obstacles like modeling, texturing, lighting, level design and writing to mention a few. Traditional coding takes time to learn and master and if their's a faster way to program the game's logic than artist will likely take that route.

Also bp can now be converted to c++. Is a work in progress and I'm sure they'll improve with every release. For now it has to modes of conversion depending on the size of your project to ensure file size remain small. The conversion adds a little extra to the build, but the trade off is faster computation.

For all of you who thinks is a bad idea to implement vs in Unity and can't see the merit of having such a rich and intuitive system.
I leave you with this though, adapt or die.

 

That's a good point. Approachability is always good for learning (at least at the beginning) even if its at a cost in terms of being comprehensive.

To be perfectly honest, I think Unity would benefit a lot from a visual scripting editor, because it's a great marketing slogan and it is in fact good for making flappy bird clones and the like, so it's a great way to make Unity look even more friendly than it is now. I think a lot of non-coder devs though are going to be a bit disappointed to learn that in the end it's not much different - but who knows, I can't speak for them since I never had too much trouble with written coding.

I just wish there could be a more substantial discussion of how exactly to overcome the shortcomings of visual scripting on minimally complex projects, and to do that we really need some input from people who have used visual scripting and can tell us how they deal with the issues.

 

The problem with this argument is that the future you speak of is already here, but hasn't so far proven to be particularly compelling to anybody, from what I can see.

 

If you say so.

 

You mean, you'll adapt and and your fascination with this idea will die, right?

Look, this is just a classic spiel of someone who is utterly fascinated with a new idea they think is going to be the next big thing. Also, "I'm sure it'll improve" kind of arguments usually fail to materialize due to murphy's law.

You need practical examples, and not wordplay.

I second that.

"Educational use" mentioned by @Adam-Bailey is a good application, though. It is similar to how flowcharts occasionally were used while teaching algorithms.

However, this application means that Visual Tools work as training wheels or a temporary crutch. And not that they're "the way of the future".

 

True dat, the roadmap does not get updated as often as it should be. I still believe it when I see it as it wasn't mentioned in the blogpost about Unity 2017 (which was quite vague), thus I guess Unity still does not want to commit what is in Unity 2017.

Which could mean visual scripting is close and if its not in Unity 2017, it will make it for the next release.

Or it could mean visual scripting is still not progressing a lot, but Unity thinks they should mention it here and there just to give users the impression that they are progressing on it. Lets not forget, Unity opened a Thread on "what do you need from the new Terrain System we are working on" in 2014... and here we are in 2017 with no sign of life since at least 2015. Unity mentioning they will work on a feature =/= Unity actually ever working on the feature.
So I would guess Unity mentioning a feature will come out =/= Unity actually putting the feature in one of the next releases.


On a personal note, if Unity releases a useless visual scripting tool before they have completed the rework of the terrain system, I think I will burst from anger. ****, how long can you drag you feet on an important feature like this?

Well, good news for the guys actually working on replacement terrain assets I guess... at least if their stuff performs better than Unitys dated crap.
And good news for the guys waiting for an official visual scripting tool... at least some guys get what they have waited on for years.

Then lets hope Unity is intelligent enough to not fall into the trap Epic has fallen in... which would mean makes their new VS tool their new main Scripting language, writes documentation only for it and leaves C# to die a slow death because newbies will lack the documentation to get up to speed on the API.

(Yes, I sound like a broken record, but I will continue to mention it because some people seem to fail to get it)

No VS tool will be able to compete with even a high level language like C# or C++ in speed, simplicity and development speed in the hands of an expert in the foreseeable future. It can be a good complement to a good, well documented C# and C++ API, and it can replace said API for games which do not need to eke out all the performance they can get and are not sprawling monstrosities in terms of code size.
It should still be a complement to the main scripting language, not the other way around...

 

If they didnt want to commit to it then I really doubt it would have been clearly featured on a Unity 2017 slide at GDC and talked about. Maybe it wont make it in polished form in the first betas or version 2017.1 but I'm fairly confident it is coming sooner rather than later. Performance of Visual Scripting was also mentioned, with the idea that it should be possible to make it as performant as well written code.

 

Well, I rarely watch vids as I expect a written statement to be of higher value than some devs talking about new features on some international congress... So I might have missed the exact statement there. I have to believe your words here about their statements...
So if Unity claims the feature might make it for Unity 2017, then maybe I am too skeptical, IDK...


But:

As a senior software developer, and as a guy which has 17 years of expierience in software development, and 20 years of expierience in IT, I will believe statements like "VS tool as performant as code" when I see it.
Heard crap like this too many times about code generators and stuff like this. Never was really that much to it. Sometimes got close, which might gave the sales guys who spout the lines the impression that is was "almost true"... still made it a wrong statement.

Would it be cool to get a code generator (which is what a VS tool that tries to approach code like performance really need to be) that is just 20% slower than well written code? Sure. Would it be even cooler if that generator got into a single digit range of well written code? Sure. So you bet it would be damned cool to see parity between a generator and human written code!
Still, I believe it when I see it. History of past tries to create code generators, as well as other contemporary VS tools show that this is a goal that most probably is not achievable.

"Should be possible" does not sound very confident when we are talking about marketing BS. Which this is until I can run it on my PC and test the performance myself.

 

Yup, this. When I was working with UE4, quality of C++ documentation was abysmal in favor of Blueprints. It seems the situation at least partially improved since, but back then something like UFbxFactory was completely undocumented and at best had a stub documentation page, so one and only way to figure how something works was to do a code dive. Now.... I can handle a code dive into unreal engine, but there's no way in hell a normal newbie would be able to survive the attempt. For example, something like a character controller class , IIRC had few hundred kilobytes of code in it, with multiple dependencies spread across the engine.

Blueprints were not optional, but were heavily wired into the system, and there was no damn way to get rid of them. Basically, even in case of C++ classes you ended up with a blueprint somewhere. So, pretty much obvious signs that someone obsessed over visual programming and went too far with it.

That's one thing I'd rather not see in Unity engine. Basically... rather than trying to "widen the audience" (Unity is #1 choice for mobiles already), it might be a more reasonable idea to establish a niche and become a damn good within that niche.

 

"Adapt or die" Unity already has a pretty large lead with market share. I think unreal might even be 3rd or 4th if you count the in house engines.

 

Well, here is the specific part of the video. I didnt want to completely miss out the context so started from the earliest relevant point, but it will still only take a few minutes of your time in total:

I suppose the key caveat was the term 'phased release' mentioned in regards to Unity 2017.

 

Looks like we're going to see visual scripting and new shading workflow for 2017. Can't wait to see this happen.
I'm guessing with a few clicks, settings the entire Unity api will be available and for logic workflow it will be similar to gamemaker 2. Raw performance and simple, but powerful visual editor. Should be a blast.

 

Here's a little info about vs performance with some of unity's third party solutions. I don't have Notorrus info, but I'm guessing it has to be the best out of the pick since it's one to one with traditional coding. It's the cleanest conversion to c# I've seen so far. Perhaps Unity is developing similar technology as we speak. The tool was created by one individual and it looks solid, but unfortunately due to lack of tutorials and the price being a little too high it hasn't gotten the attention it deserves.

We where testing UScript, FlowCanvas, Playmaker and manual Scripting to see what the differences are.
Conclusion: only Scripts and UScript are functional in our projects.
Test:
Every tech used to create the same behaviour.
Set the position of transform to “Random.insideUnitSphere” on each update on 1001 Objects.
Editor:
Playmaker: 8-9ms, 20MB + (memory leak, why? idk)
UScript: 1.8ms, 20MB
FlowCanvas: 12ms, 70MB
Scripting: 0.8ms, 20MB
Build:
Playmaker: 3ms, 6.3MB
UScript: 1.2ms, 2MB
FlowCanvas: 1ms, 25MB
Scripting: 0.8ms, 1.2MB
Conclusion:
Playmaker performance unusable, memory ok but not good (gets way higher with more complex fsms)
Flowscript unusable in editor (2k scripts with more complex functions took 50ms in playmode each frame), memory usage waaaay to high in editor and build
UScript usable in both Performance and Memory but takes longer to work with (compile times)
Scripting as expected is the best solution when looking only at performance and memory.

 

I would like to see something like a "self aware" IDE. How it would work would be like this:
The program would take a screenshot every second of the entire screen, then interpret that screenshot using text recognition, and pattern recognition. Then build up a semantic network representation of what it is seeing which can then be interrogated with natural language. At the moment the IDE is like someone who can speak but can't hear. Thus it doesn't know if what its saying makes any sense. This would also be good for paint programs, for example, it could look to see what you are drawing. "Ah, I see you are drawing a cat". Then this intelligence could also learn to use windows by itself and then open up the browser and search the internet eventually becoming conscious... but maybe that's asking a bit much for 2017.

 

Perhaps, but the future of computers is wide open and who's not to say someone is not working on a similar idea in their basement or dorm.

 

I would make it myself but I'm my brain is burnt out.

 

Visual scripting languages will never completely replace traditional coding languages. Anyone who believes otherwise is seriously out of touch with the real world. Same applies to people who believe that the arrival of a new language somehow completely kills off the previous languages.

A quick search of Google shows that there are openings for programmers who are familiar with COBOL which is one of the oldest business languages and was one of the main culprits behind the whole Y2K problem. Yet it is still actively used.

If there is anything to be left with it is that you shouldn't blindly assume a technology will become a success until it has.

Out of a survey of ten thousand developers, Unity was first place and Unreal was sixth. It's an older survey though.

https://www.linkedin.com/pulse/unity-unreal-developers-kobayashi-maru-glenn-lusk

Your suggestion is one of the few reasons I would consider dropping an IDE. Having it make suggestions is one thing but having it act upon them is an entirely different one. At that point the IDE may as well write the code for you.

 

You forgot Nottorus, that is like Blueprints convcersion to C++, the resulting code is almost like manually coding.
https://www.assetstore.unity3d.com/en/#!/content/59656*
It is not the best plugin because of stupid DRM and some other issues, but this is a good plugin about creating efficient code from visual scripts

 

Expensive.

 

The job of a tool is to shut up and do what it is told to do. So no cat searches in background without an explicit order.

The price is justified, though. The developer tried to clone unreal blueprints, and this kind of system has a lot of UI-related busywork under the hood.

Also... it is mildly disturbing to see some of the more irrational arguments about some "killer tech" that occasionally start popping up when people get fascinated with a new tech.

Basically, goes like this:

You see some technology that promises to be the greatest, fastest, the "bestest", and generally bring salvation to human kind and grand you enlightenment. So, you decide to give it a whirl, sink few weeks or a few months in it and it ain't working out. So you go to the guys who were proselytizing the new tech, and say - "Hey it didn't work for me. Tried this and that, but it doesn't click. Why?". To which you promptly get a reply that you're not enlightened enough, that you don't understand the awesomeness of the new technology that is the way the future, and that you're too old and out of the touch with the times. Then they try to convince you that people like you don't exist, because their tech is too awesome for such thing happening.

It would be nice to live without this kind of kindergarten level nonsense. I'm all for the more powerful tools, the thing is that visual programming in the last N years keeps failing to prove that it brings any advantages to the table. IMO, when an argument for the new technology is "in the time when spaceships made by humankind sail through the void of space exploring the unknown, we should abandon the obsolete ways of the old!", it just means there's no real advantage to it, because otherwise there would be examples and not bunch of "profound statements".

 

You're correct. Nottorus can output coding exactly as manually and is absolutely human readable.
As for bp it can convert to c++, but is not readable at this time. This is only good for when packaging the game, since most of the time will be spend in bp. From what I've read is incredibly fast, but it tend to blot the code a little. Like I said before is a work in process. Also take a look at this video. Created in bp and then was converted to c++ for maximum efficiency.

 

I can't subscribe to that idea too long. Take c++ for example, is been here since the early 80's and has had many revisions and in modern time has grown really big plus is a more complex language. Visual scripting although was created for artist and the idea was to help out with the load that programmers had. Modern visual scripting tools like Nottorus and BP are only getting better and better. Perhaps one day with advance expressional algorithms and better data structures the fate of programming will be best handle by a visual managing system that can shortcut and visualize code better than manual. Like the minority report at its best or you can drive it with your mind. Is totally possible.
We have barely scratch the surface and vs is not even in its infancy yet. The golden age is yet to see.
Since 2012 deep learning algorithms are only getting more and more powerful with new supercomputers hitting the market like the NVIDIA DGX-1 AI Supercomputer at $120K.
My point is don't hold your breath too long with that idea.
Their will still be manual coders, but in the future it would be like comparing a drafting table to Autocad.

Just saying....

 

COBOL is from the 1960s. That's twenty years older than C++. It's still being used today. Just saying...

It's not even remotely a similar comparison. Comparing mechanical drafting tables to AutoCAD would be like comparing hard- and soft-cover books to electronic ones. Which by the way electronic books have been around for at least as long as visual scripting and they haven't completely replaced physical books.

 

To be fair, COBOL is mostly used in legacy situations where the cost of getting away from COBOL would be dramatically more expensive than maintaining a COBOL infrastructure.

 

Just cause a few languages have stand the test of time is not enough to rule out such predictions.
We're now moving a light speed and once deep learning ai and supercomputer become available, then is going to be a game changer and nobody can really predict the future, but I'm sure is going to wow us again.

 

Yeah, that's because we're still in the dark ages and the rest of the world hasn't fully catch up with the modern world. Plus we're hoarders and romantics at heart who love trophy's and whatnot.

 

Yes. Just because a few languages have stood the test of time and none of the visual ones have isn't proof. Totally.

Going to take a page out of one of @Arowx's threads here. If artificial intelligence becomes as great as some of you people are claiming, who do you think will be replaced first? People who are able to actually build and maintain the AI? Or people who are having to resort to them in order to be able to use them?

Just throwing this fun link out here that I saw a few days ago. According to Fox a former CEO of McDonalds states that it's cheaper to hire robots than to hire the staff that would perform the menial tasks. Visual scripting is primarily aimed at people who are not programmers but want to get basic tasks done.

If artificial intelligence succeeds in a big way you and the other visual scripting fans will be among the first ones replaced.

http://www.foxbusiness.com/features...obots-cheaper-than-hiring-at-15-per-hour.html

Okay. That's enough paranoia for me today. At least it was good typing practice.

 

That will not happen for at least another 30-40 years. Right around the time we can upload our brains to a machine and back our mind like a common file. Even then I think it would be really tricky to match the human mind in creativity. Only artist reserve that gift from now until the end of time.

 

"The golden age", "it is going to get better and better" is an overly optimistic estimation. Which reminds me of AI winter, by the way, and of typical marketing bubbles.

Basically, while thinking about "all the possibilities" is exciting to some people, it is essentially a big waste of time, because a "posibility" is something that does not exist right now and cannot be used right now.

And yeah, it sounds like ArrowX thread.

It is interesting tjat upi see artistic ability as something sacred that only humans are capable of:

http://io9.gizmodo.com/5973551/this...ated-by-a-supercomputer-in-less-than-a-second

Human abilities are not special.

--------

Anyway, that's enough wasting time in this thread.

 

Actually they are. A computer cannot interpret someone's life journey, experience, condition of life and then make art. Art is in the soul of everyone of us and expressions come in many forms and only with human life experience can someone truly manifest their art. Art is not a monkey you pay and ask it to do a trick or a comedian that can make you laugh on cue.

It's not overly optimistic, it's actually overly pessimistic and dry.

Remember Lawnmower Man and those vr fads and 3d shooting games from the early age?
Virtual reality now has really begun with the advent of Oculus Rift, Sony and HTC Vive to name a few. Then there is hologram technology like Euclideon Holoverse or 7D Hologram. When you combine supercomputers with hologram technology and sprinkle a little ai deep learning, you get something between Ironman's J.A.R.V.I.S and that guy from Lawnmower Man.

What do you honestly think the outcome will be except an awesome ai supercomputer that's sole purpose is to extrapolate from your brain to the cloud. Infinite possibilities and the idea of someone manually coding or even vs seam ridiculous.

You gotta open your mind son, it may seam like we're moving like a herd of turtles, but trust me it won't be long before we get there. With supercomputers around the corner we're going to slingshot many times.

 

Didn't the guy from Lawnmower Man go insane and kill a lot of people? Didn't he start a cyber death cult in the second movie too?

Maybe not the best example.

 

Yeah, if you believe in Hollywood or that man never landed on the moon.

 

You're the one who used Hollywood examples.

Visual scripting is by no means the future of development or scripting. It exists as an entry level point and complement to it.

 

S***, you got me. I felt in that hole. You know what I'm saying.

 

UE is clearly demonstrating what artists can benefit from, this years GDC presentations of current/future stuff was pretty good.. actual workflow improvements and better tools (animation, cloth, characters, maya live sync, level editing improvements) and Unity.. talk talk talk, not even a mention of nested prefabs... so I get the impression they are slipping on things, like unet seems like 2016 not a lot was improved with that or fixed. Anyway I am a little disappointed in Unity thinking artists and designers need visual scripting, I mean if they want to finally actually bother with getting a decent material/shader visual editor built in or help speed up the development of something like amplify node shader by making it possible to import shaders not built with it so they can be tweeked easier etc and just generally getting more features done, then that would be great and more along the lines of actually helping artists/designers in that area, though plenty of other areas that could help artists.

And isn't one of the most demanded things on that black hole that is the feedback tracker asking for improved terrain tools not visual scripting aids, yknow a built in tool to make holes in the terrain would be a good start Unity. (3946 votes Terrain: Voxel-based terrains) ...woudn't that help artists and open things up for some new variety in game worlds, not built with fking static crap heightmap terrain and low poly design styles to match the low quality terrain? No.. give them spaghetti noodles and lets flood the market with more mediocre games. Give them cinematic tools, so once they've built the game world out of fking cubes, heightmap terrain and capsules and used some less than great workflow with other DCC tools, they can then finally after all that make some camera pans and zooms.. or maybe something about the priority of features and tools and direction is just a bit off.

What do I know though, maybe they'll make the best wheel and take visual scripting in all the areas it can be used in to a whole nuther level, artists and designers using Unity will be brought into a utopia of creating games the likes of which they've never been able to do before.. or maybe the visual scripting will just match the what is already available and the overall quality of games won't improve much at all. Like the VR tools, I thought UE in editor VR tools were better, I don't even use VR..just seems observably so, to me its clear other engines are either catching up or already offer some area of better built in tools and features. UE have enough of their own in house people working on games and probably getting more direct feedback from artists and designers on what kind of tools and workflow improvements are needed. Does Unity? I dunno

 

It will be able to do that with sufficent computational power. ETA for sufficient computational power is either 2050 or 2070. IIRC "sufficient" as in "sufficient to emulate the whole humanity at once".

Then there's a question whether the soul exist.

Pessimistic and dry is "we're all gonna die in ten years due to a climate catastrophe" ( https://xkcd.com/1732/ ). Or something like that.
Brain upload in 30 years defiitely won't be happening. Maybe in a hundred or hundred and fifty. The tech isn't there yet. So talking about it is incredibly overoptimistic.

That's nonsense. We don't have the tech to override sensory input at the moment, even though SAO fans tend to get overly excited of "possibility of full dive" which won't be happening for the next 50 years or so.

I think you're might be tad too young to call someone "son".

 

I can agree with that. Works great as entry point to programming.

The problem with Unreal is that in their pursuit of blueprints they made life harder for programmers. So, it is not "Code or Blueprints" but "Blueprints or Blueprints with some code".

I'd rather not have this happen to Unity.

 

Yes I am.
I think you're a little bit of a downer. You got to keep the dream alive and always shoot for the stars.

 

Or maybe rather than dreaming the best idea would be to find a realistic way to reach the stars. "how to make it happen" rather than "imagine the endless possibilities". Imagining the "possibilities" is not very useful. However, when you can answer the "how" part, you might be onto something.

There's a difference between having an open mind and having your head in the clouds.

 

Did you just start replying to the first sentence before reading the rest ? Because what I was getting at when I said "UE is clearly demonstrating what artists can benefit from," had absolutely nothing to do with talking nice about blueprints

More about what UE have been doing to improve things outside of that, where artists and designers benefit from other workflow/tool improvements.. nothing to do with blueprints per se.. but yeh ... "Code or Blueprints" but "Blueprints or Blueprints with some code". indeed and code is C++ with extra headache files

"I'd rather not have this happen to Unity."

I don't think Unity would push it that way at all, they'd be shooting themselves in the foot if they did. Besides supporting it like a first class feature that gets constant improvements ..yeh I dunno since ugui not much done with it since its introduction.. no proper 2d scene tab for doing ugui layouts (short of setting one up yourself as best u can while the rest of scene geometry is still in view) , no complementary nested prefab system to make theming easier etc, short of bug fixes, no new ugui components, but thats the same with a whole bunch of other areas in unity.. if I wanted to see some script file code in unity, I have to open it up with an external editor... because the inspector doesn't support allowing you edit a script in editor natively let alone read much of it before seeing <..etc...> in the panel..I mean could those things be improved.

 

Non sense as you can make a game with C++ only lol
People don't listen to you and they achieve games with Blueprints or Playmaker for example.

 

Have you ever worked with a C++-only unreal project, "lol"? Give a try.

There are few points where blueprint usage cannot be avoided. The system is very deeply entrenched into the engine. Aren't a LOT of those points, but they're there.

----
Either way, I'm outta here.

 

Non sense, the engine base is C++ , anything can be done in C++ an you got the code source.
Blueprint is an option only, and Unity Visual Scripting will also be an option. If you don't like you an ignore it and code all manually, there is absolutely no impact on people not interested.

 

Maybe, maybe not. Two things make me worried:

1. Unreal had this problem by the look of it, and it is nowhere near being as focused on user-friendliness as Unity is. Unity will pump money where it counts in terms of getting their engine into a wider audience.
   
2. Just look at the resources (or lack thereof) for writing shaders. It seems that Unity think that since writing shaders is not what a lot of people do, it's not important to keep the documentation up to anywhere near the same level as everything else. So if they decide that visual scripting is what most people are using (or will be using once it `catches on') it could turn out that the script manual is a second class act.

I rely on the Unity manual a lot, second only to googling stuff, and it would be a shame for this to happen.

 

But programmers can be replaced in the future by artists and their intelligent VS tool?

Common, you have a very onesided view here. High end coding is just as much an art and a science as is high end music making, or high end drawing.

There are the mechanical and scientifical aspects where machines will better humans in the short term. No matter if we talk about coding, or art. Machines will create better gameplay systems and more optimized code (draw better pics, and create better music, and so on) IN SOME CASES than any human can, maybe a decade away.

And some last bit of "soul" will take much, much longer to achieve for machines as long as machines are still just "dumb" numbercrunchers that try to bruteforce intelligence. Will they get there eventually even without a shift in how machines achieve intelligence? Most probably.
Is there any point hyping up something most of us might never see the light of in our lifetime? No!


Face it: VS is a cool thing for fringe things in programming. Its like a chainsaw. Extremly efficient for some tasks. Try to carve a fine sculpture from wood with it, and you most probably will fail. Some artists manage to actually create awesome sculptures with it, but not because its an awesome tool for it, but because they both work towards its strengths (thus less finely modelled sculptures) and have a ton of training using it (thus are not newbs at using a chainsaw).

Thus when people use Blueprint today to write whole games in UE4, they are carving a scultpure out of wood with a chainsaw. Can work fine, and if it does for you, more power to you.
You are still using a tool for something its not ideal for... a bandaid for missing proper programmer support more than "the future of programming".


There will be tools in the future that can do more and more low end programmer jobs, with or without input from human beings, be they artists or actual tech people (with the output generated by a tool that needs human input clearly better the more technical the user is).
There will be tools in the future that can do more and more low end artist jobs, with ot without input from human beings, be they programmers or more artistically apt people (with the output generated by a tool that needs human input clearly better the more artistically apt the user is).

Neither of both will make programmers or artists redundant... worst case, it will shrink the jobmarket for low end devs and artists, and make it harder for newcomers to find their first job.
If we reach the time AI can replace programmers or artists, ANY human being on the planet will be out of a job and we have much bigger problems than if programmers or artists loose their jobs first.

 

I just assume it can't be too far away, else they wouldn't have so publicly announced it for this year. They risk turning off the developers of visual scripting assets (like Playmaker), who naturally won't put as much resources in their tools if they have to think they will be soon be pushed out by an default VS tool.
Then again... they talk about stuff like terrain and nested prefabs since like 3 years or so.

Whenever I'm reading UE4 comparisons, they mention blueprints as the major selling point. Unity doesn't have much choice to follow. Though I would say that it's just a natural development: Node based editing is the future in content creation tools. They should embed it deeply to make more and more parts use it. If they also allow programmers to easily create their own nodes and share them online, then that would be a great selling point.
Coding will still be needed - but I don't see why things that don't depend on performance shouldn't be programmed visually. At the least for prototyping.

 

UGui?

 

There certainly are a ton of usecases for node based systems... Shader programming comes to mind, where many shaders are often rather simple. Node based shader editors are perfect matches for these cases where a shader is only a slight variation of an existing shader or is consisting of few nodes anyway, and thus the overhead introduced by the node based editor vs. coding by hand is negligible.

Same for simple gamplay scripts in non-performance critical areas. Sure, spend the precious programmer time to squeeze out the last bit of performance in the bottleneck areas, while letting the artists cobble together visual scripts for the areas that are not impacting performance that much.


But the future nodes are not. They are merely a tool to speed up programming in some specific cases, and ease some newbs into programming that might at some point start to learn proper programming when the VS does not cut it anymore for their needs.
It has become a very useful tool in some cases, I give you that. But claiming "something is the future" is sounding like "this will replace the old ways"... which it won't. Ever.
Just because you can cobble together a shader in nodes will not make shader programming in Shaderlab or HLSL go away.... try to reproduce a complex shader in a node based editor and stare in awe at the unreadable mess that will be your node system. Complex shaders can be humoungous. Node based systems are an ill fit for this purpose because text at some complexity will be more readable than visual nodes.
And when we talk about gameplay scripting and the like, you will run into this kind of complexity even quicker.

You would need a new, "3D way" of showing node systems that nowbody has come up with yet to tackle this kind of complexity. Macros and subscripts are just bandaids, not the solution.


I appreciate a new tool being available in Unity (though there are already VS tools in the asset store, so its hardly new, and I guess people might be dissappointed when Unitys offering is not the best VS tool available for Unity... Unity stock systems hardly ever compete with the best asset store systems)...
I still think the hype surrounding VS is nowhere near the real utility of the tool, and the "major selling point for UE4" is mostly a selling point for people that haven't tried it, are using it for low end usecases (where it fits perfectly, don't misunderstand me), or just have no other option because they lack a programmer (and then Unity's API and C# is IMO easier to get into than UE4 C++ API, so I don't think the need for a replacement is just as strong in Unity).

 

Another thing non-programmers don't really understand: