Strasz, I dont mean this personally towards you, but please dont twist my words.
Im sorry for my somewhat broken English but im typing this on a phone, while being hospitalized.
Even still I believe its pritty clear to anyone that noone is satisfied with the factual reasons provided. And thats the problem, not a communication problem as to how many of the team agreed or not.
The simple fact is you guys decided to not reach out and ask for help after firing merlin.
Not that it was a management decision alone.
Also while I obviously dont know the entire development team personally… But I guess i must just know everyone who hated these news personally then. Not one single friend of mine agrees with this decision.
Please separate UDONSharp from the SDK, so that the community can continue updating it and provide creators with a genuinely improved experience, albeit limited, at least until Soba becomes truly usable. We have lost trust in the official commitment and update pace.
Postin from the woods on a shaky LTE connection. a squirrel is watching me type.
Folks, this thread has gotten really far off the rails. I understand you’ve got strong opinions and hot takes, but nothing excuses personal attacks or borderline harassment.
There’s ways to have this discussion productively, respectfully, and with the goal of either understanding each other or of improving the end result. All I had a chance to do was skim the thread, but the posts I’ve seen are nearly devoid of all those qualities, and have a healthy dose of extremely obvious passive-aggressiveness, too.
So, please, knock it off. “Remember the human,” or whatever reddit used to tell you. Remember that the team wants VRChat to do the best thing possible given all the factors at play. We use VRChat too, and we create using VRChat too.
So, take a lap and chill out a bit. (and if you’re in the US, go find a turkey)
Tupper with all due respect, VRChat needs to make an attempt to understand the community before it expects the community to understand it. VRChat went out of its way to obscure the reasoning for a major business decision that directly impacts many of its users and indirectly impacts the entire platform.
I’m always loathe to join these threads as they tend to be a bit…this. But I’d like to steer this to a broader question: Why?
A lot of us are trying to understand what value Soba offers to the platform, players, or the community, and I’m willing to accept that there are upsides we just can’t see. But from a Product Direction standpoint, what needle is moved by Soba that balances the confusion, added ecosystem complexity, and limitations that this creates?
What are the biggest problems with Udon? And how could they (with Udon 2, Soba, etc) solve them?
Issue 1: World creation has an incredibly steep learning curve.
Solutions: tutorials, examples, or develop/license/put more effort into creator entry and midskill help
Nothing here.
Issue 2: World creation takes a really long time even if you know what you are doing
Solutions: drag and drop solutions, QOL tools, bug fixes, tutorials
Nothing here.
Issue 3: Udon is unperformant and affects the playerbase as a whole
So this is built upon the assumption we have a lot of worlds with heavy Udon, which we don’t really, because of Issues 1 and 2, but a worthwhile thing to solve
Issue 4: Udon is limited and weird. Professional developers familiar with C# cannot work with VRChat systems without painful compromises, creating a barrier to entry for top talent and commissions.
Udon 2 promised this. Soba does nothing afaik.
Issue 5 (edit): Udon is (possibly, I’m not savvy enough to judge against other solutions) kinda buggy, techdebted, and unreliable, a new system could be more stable and easier to maintain
? Possible Soba value
Other issues anyone can think of that is Udon-specific (not platform, economy, or networking itself)?
True.
The official team should maintain transparent and effective communication, which is key to maintaining user trust. In explaining the reasons for canceling Udon 2, the official team seems to have not fully met user expectations, clearly leading to some dissatisfaction. Additionally, they should not make so many overpromises that cannot be fulfilled, as this has already led to user disappointment and distrust. Therefore, it is truly, truly necessary to actively listen to user feedback and encourage community participation to improve products and services, which is what players really want.
I dont understand many of the given technical explainations and to validate any of the given information from anyone as truth is something I dont want to get into.
The major problem you can reliable follow though is the communication.
Dark made a huge post summarizing some of the previous communication leading up to where we are now.
How is it so unexpected that when potential features, which have been announced over a long period of time, are now in question or even canceled, that people would like some detailed answers.
“…but dont expect us to get into a deep analysis here” - Fair, but then you should have done this beforehand.
You cant spend most of the time hyping up (even if unintentionally) one thing and then suddenly announce another without giving a detailed thorough explaination as to why things changed before the announcement of something new, even if similar.
It should also be understandable that when so much information which was given is suddenly put in question, that it is hard for anyone to just move forward and especially trust in those decisions.
I want to believe that this is the best way forward, but I dont have enough information as to why decisions were made to believe it is.
I personally only see 2 options here.
Take a full update to clarify everything
Go radio silent towards critisism (or even in general) and wait it out
I know I’m probably gonna earn my own suspension here for pointing this out.
But it speaks volumes that the person with the most hearted comment, and who everyone agrees summarized the communities views very well, dark. Was suspended for three months for doing so.
I don’t think this is necessarily completely true. I, and mostly everyone i’ve spoken to over the years has observed a markedly evident decline in world performance since udon.
Granted some of this factor is the inflation of expectations and standards, the “power scaling” of shared information on creation…
My measuring stick is that i used to get 80fps in a vast majority of worlds pre-udon, and now, even with a notably better pc, i struggle to hold 40 anymore in most worlds (it has to be essentially barren to be able to run at full capped framerate in udon worlds). And it’s not just me, it’s constantly a subject of comment that we’re all always struggling to keep decent frames unless we have X3D bank buster masterrace. The lower framerates causes fatigue in most people. ※ edit ps: it’s not just low framerates, it’s also frame pacing and microstutter issues (metrics say 40fps but it feels like 24).
This could be a combination factor of your Issues 1-3, where the lack of knowledge results in people building more crude/messy spaces that as a result, have worse performance. Issues 1 & 2 seems to have muddied the standard conversation, where you’d normally get a lot of basic do’s and dont’s… like for example don’t have 5 realtime lights to make a soft shadow for a lamp (i presume these are meant for mitigating baking artefacts and to then be disabled), but this information is all lost in the sea of technical confusion and pre-built assets.
I know it’s not my GPU that is the issue, because as a shader developer i measure and stress my GPU on the regular and can intuitively determine when something is causing GPU bottleneck… but it almost never is GPU (unless it’s obvious, like a raymarch shader, or reckless abuse of grabpasses like those bloody rainy windows), it’s always CPU, and even with extensive care to unity basics like mesh renderers etc, it’s clear the worst offenders are out of reach for solving by the end user creator - animators and udon. I slap together an SDK2 world with carelessness in optimization, and what do you know, it runs like a dream; as does essentially every SDK2 world i enter, no matter how complex/detailed it was (go back and look at these worlds, the contrast is very stark). I think people underestimate the impact of udon performance overall, and this is why there was such excitement for the future outcomes of udon2, because it’d sidestep a lot of the complicating factors.
edit ps: i do world hops with a group like every day, and done so for years, so i have a solid sample size of worlds being made. world development literacy is extremely low anymore; almost every world has a major rookie mistake in it… and it seems to be getting worse. (not to anyone’s creative discredit, they just aren’t getting the information they need nor ease of access, and the difficult / performance-restrictive udon development process exacerbates this greatly)
As a closing note: adding more tutorials and stuff would make the entry and growth a bit less painful, but that is a band-aid. As always, the entire pipeline needs to be simpler and more functional for general use, with just modularity/scalability to get more sophisticated to the level the creator is comfortable. The smoother everything is, the more amazing and appreciable content will be created. Efficiency across the board should be priority one.
I mean outside some few bad apple i think a lot of what is said is civil. Sure its tense but i would not call it harassement.
Edit: this is indeed pressuring but the community that is speaking up atm are ( for the most part) passionate benevolent people that love this game and want to see it evolve the way they expected.
to be fair, it could be the other way around: We don’t have udon heavy world because udon is not fast enough. You really don’t want to add too much extra scripts to a hangout world. A single pool table can cause unpleasant spikes while people are playing…
Yep, handled it. Thanks for calling it out. Zero tolerance for encouraging self-harm and harassment-laden posts, folks. Please refer to my previous post.
Yeah, that’s the thing, as someone who visits at least a dozen new worlds most days, a majority of worlds have some sort of gimmick in it. Be it the pool table, sacc flight, arcade cabinets or other minigames, lots of little doodads like chips and coffee machines, or god forbid chikuwa cart… Even the video players can’t be blameless either, they’re an essential feature.
Clearly people want these features, but as they are, they detriment everyone’s experience by existing and hogging cpu time. I’m sure people notice this, and don’t add more because of it.
I’d argue soba very slightly increases the learning curve. UdonSharp being unlike using C# in unity means that people are often recommended to follow Unity tutorials only to find they don’t work! Newbies might try graph only to find out more experienced folks don’t know graph and don’t know how to help them.
Soba creates a third category of script that has different support from U#. So if they’re following a guide that uses U#, or using someone else’s code, they need to understand what soba and U# are and how to use the tool that translates to Soba.
Also a separate concern is community splintering. Like will package devs and tool makers need to maintain U# and soba versions?
I kinda wanted to put this as like an addendum to all of this> a question to VRChat:
Have you ever considered inquiring into the appeal of a feature and it’s potential, when at the planning stage, before development even starts on it?
…Instead of working on something for a long while, then announcing it, only to find out people are not confident in it at all. It makes sense to me to survey the land before you start building on it.
An ongoing development platform is not like complete software package where the company has a vision and a plan, and gambles on it’s success. A constant development space is one that is always evolving, and to stay ahead of the curve you need ‘scouts’ to ‘take point’ and see what’s ahead and if the road is a good state to travel.
If you believe that something is gonna be good and most people are just unsure, then you have pressure to prove them wrong with something that will exceed their expectations, and do that from the outset. Giving yourself pressure is good - game dev is about pressure, if you get too comfortable, it will breed complacency.
I would like to quickly point out that I have a canny post to get Udon 2 back. Please upvote it to get their attention.
Can we get some inside knowledge into how the VRChat team as a whole structures its priority tree? I’m beginning to believe that the VRChat team is, to put it bluntly, dysfunctional and disorganised.
You say you want to prioritize features the people want, but you’re not giving us what we actually want.
Then why was this not done in the main post? It’s yet again this same issue of you guys making a major change and not telling us why you did, until the community gives you backlash for it.
The VRChat team has such a bad track record for distrust and miscommunication, do you not learn from every other time you’ve done something like this that we’re going to want reasons?
So, what I’m hearing here is that the engineers couldn’t hack such a big codebase? In which case why was this not taken into account before laying off Merlin? If Merlin was the only person actually capable of comprehending this codebase and pushing the feature, why not keep them employed until Udon 2 was released?
It seems like, even high up in the management tree, you don’t consider the repercussions of your actions. And unless we get answers to say otherwise, then we can only speculate.
Well, yes of course. When you’re developing something that is objectively worse than your reference, the development time is going to seem faster.
Alongside this, what does “reused existing code whenever possible” mean? Did you just, take code from current Udon/UdonSharp, change it up a bit, and spit it out? It’s quite disappointing if that is the case.
Overall, I’m extremely disappointed with VRChat’s continued lack of transparency and communication. I’m disappointed that it takes a large backlash for you to finally give us some form of transparency, and even still, when we get it, it’s always the bare minimum.
You need to start putting the community first. The community is what makes VRChat what it is today. You can’t keep mistreating your community and expect them to stay. People have already left VRChat for other platforms, and will only continue to do so unless you get your priorities straight.
To put it extremely bluntly: Start putting the community first. Stop adding useless features that nobody asked for, like stickers, boops and gift drops. You need to put your resources into features that will actually improve the user and creator experience.
In likeness to this, I will have to say the thing we’re all thinking here. And no one has flat out come out and said, because I feel they have just been baiting for it to be said.
Although you may have terminated him for valid reasons. What has been obvious to the community that has happened here, is that:
Merlin finished a project, that seemed to be complete.
(edited: Being told it was ready to go into beta and needed smaller fixes)
And it comes off as extremely spiteful, that all of his work is being canned after terminating him.
I feel they (dark) are trying to get staff to address this, and give a proper explanation as to why Udon 2 is being dropped.
What was the rationality of switching to something else?
And why is it suddenly Soba?
