Developer Update - 25 November 2024

I wouldn’t expect a reply from VRChat itself knowing how opaque they have been ITT

2 Likes

will network synced interactions have less latency with Soba? i’m unsure if the current Udon is responsible for some of the delay when performing tasks such as shooting a gun, or hitting a tennis ball. there has been tons of exciting updates and improvements with VRC, yet i have barely seen any focus on fixing latency.

i understand the reason for a lot of latency between player interactions is to keep IK and voice properly synced, basically a buffer. yet i have seen other VR games that don’t have this issue such as Pavlov and Contractors. i am hoping Soba will help mitigate some of the current network latency issues.

1 Like

Are there any plans for VRChat for visionOS? Wouldn’t it be feasible now due to the recent addition of hand-tracking support?

Yeah, and who’s gonna maintain it if nobody else understood it in detail?

Clearly they chose to do something else instead of study the one that was made.

I’m going to hazard a guess that you’ve never worked in software development before but the whole job revolves around reading other people’s code and understanding it in order to modify or improve it. By failing to do that they would be failing to do their literal jobs - though, given some of the theories presented here in this thread that seems less likely to be the case. Even then, it doesn’t make sense from a development standpoint that an almost finished product would get tossed out in order to do more work to make the same thing but again.

1 Like

I’ve read all 126 messages so far, I agree with what has been said from the community. I have a hard time believing the reasons that have been given here (1GB heap size etc) but at the same time I also appreciate that we got a response at all.

One aspect that I don’t see being discussed here (except in one post) is that the original Udon 2 Canny was turned into a Canny for Sober. Us normal users don’t have the ability to edit our Cannys after they get 5 upvotes, which is good. You don’t want people to just change the intend of a Canny after people already voted on it.

I don’t find it ok that VRChat abused its special permission on Canny and did exactly that. They took a Canny for Udon 2 which already had 148 upvotes and completely edited it to be a Canny for Sober. Right now it looks like Merlin (the creator of Udon 2) has upvoted the Sober Canny, which they surely wouldn’t do, given what they wrote here. The date of that Canny also implies that Sober has been announced way earlier than it has been.

For transparency reasons, an existing Canny should stay as it is, being marked accordingly and a new Canny for Sober should be made instead IMHO.

23 Likes

you’re preaching to the choir my man.

why not?

Sorry about that! We should have preserved the original Canny instead of updating it. Creating a new post, a detailed pinned comment, or closing the existing Canny post would have been better. (Turns out there’s a good reason Canny doesn’t allow users to edit their own posts…!)

We started posting Creators roadmap this year, so using Canny for our own roadmap items is relatively new for us. I hope you still appreciate that we create and maintain our own posts - I think it’s better than only sharing our plans on this forum.

I posted a comment on Canny about the reason we decided to edit the post instead of creating a new post. In a nutshell, Soba delivers similar improvements as Udon 2, so updating the post felt appropriate. However, the community cares about small nuances, and those got lost when we edited the post.

I think there are some technical reasons that make VRChat somewhat difficult to port to visionOS - but maybe in the future! The Apple Vision Pro is cool, and I think playing VRChat on it would be really cool.

Sorry that we’ve been less communicative than usual! It was Thanksgiving last week, so our response times are slower. The community questions has also been asking many technical questions, so we’re taking more time to consider our responses.

Actually, we think Soba might still be on track for a closed beta this year! That’s why it still says Dec 2024.
But… hmm, maybe you’re right. If we’re not 100% confident, then maybe we shouldn’t have an ETA.
I’ve removed the ETA for now. Do you think we should add it back once we’re more confident about a date for the open beta/the release?

I don’t think so, no. :thinking:
Most prefab creators only write their scripts once. For example, most prefabs are written in U# because it’s more convenient for experienced creators than the Udon Graph.
It’ll be similar with Soba. Do you think this would be easier to understand if you had access to Soba example code? Maybe some “Hello world” assets?

So now we will just go with the downplaying card, alluring that everyone here giving feedback about soba just does not know anything by calling it all small nuances? cool, thanks for being more transparent i guess. Small nuance of being completely opposite approach (but so fast to develop!)

Hmm, I wouldn’t say they’re all small nuances, no. Our original forum post did not do go into enough detail, so we appreciate constructive questions.

The things I’d wanna ask again with regards to Soba is if we’d also get more Unity C# libraries whitelisted, like more functions in relation to the Unity Graphics API. Would there also be the eventual Soba graph editor that’d correlate with this feedback post? Community plugins (Custom Udon Graph nodes) | Voters | VRChat

Edit: Also will we get an update to the Udon/World UI stuff again?

@Fax Could we maybe get a direct response to my canny: Pivot back to WebAssembly-based Udon 2 | Voters | VRChat ?

I feel that, given that we’re now closing the gap between Soba’s canny and my canny (20 more votes before they’re on-par, and some votes are cast by prominent world creators, including but not limited to @syncpulse and @miner28_3, among others), we deserve some form of answer to what’s going to happen now.

The community has spoken very clearly about what we want from you, and I’m sure we’re all eager to see what the team’s response will be.

Hi :wave:

I wanted to post earlier, but decided to wait for the thread to cool down and get some more info.

I mentioned this on bsky earlier, and was responded to, however I’d like to just quickly reiterate what I mentioned here:

Something that we really need that would help us understand a lot better as to why this route was chosen, is if we could maybe get some more info on why Soba is better than U2 both internally, and for the community. This would definitely help a lot with the development community’s understanding of why this decision was made, and I feel it was a large piece missing from the original post.

I don’t think it has to be 100% confidence for an ETA, “estimated” is the first word after all! However, yes, please keep us informed with this sort of thing, so we can plan our projects beforehand.

This would be an amazing addition to any devlog! Perhaps a more technical section would greatly benefit our understanding of upcoming changes and features.

Thank you all!

1 Like

I don’t think that’ll be part of Soba, but there are a lot of good Canny requests about more access to the Graphics API.

This feature is interesting! The status of this hasn’t changed - we’ll need to launch Udon 2 / Soba first before we can promise to work on it.

Udon UI not currently on our roadmap, though I hope we can get back to it eventually.

I agree! We’re writing a post that explains why we changed our approach in more detail. We appreciate your patience as we work on it!

I wish Canny would let us say “End of this year” or “Early next year” instead of a specific month and date!:sweat_smile:

We’ll keep experimenting with the ETA feature. I find it very helpful to know the order in which certain features will be released, even if the exact dates are inaccurate.

I appreciate your feedback, especially the three issues you mentioned at the top

However, your Canny post also contains statements that aren’t factual. This makes it more difficult for us to post a direct response.

The features we promised for Udon 2 are arriving later than expected, so I understand your disappointment. We’re working on a post that sheds more light on why we pivoted our approach and why we’re confident that we can address these issues. I think this post will also answer some of the questions in your Canny.

1 Like

Thank you. I hope you can understand why the community is reluctant to accept Soba, and especially trust that you can actually deliver the changes you’ve promised, when your team has failed to many times in the past.

The term “trust is easy to break, and difficult to restore” is extremely true here, we get so many promises from you guys and there’s an extremely high chance you don’t deliver, hence our rightful reluctance.

Could you please elaborate on which parts are not factual? I’d be happy to address them and accept responsibility for the ones which are incorrect.

When can we expect this? And how much technical detail will it cover? Also, I’m gonna throw this out here, have you considered postponing Soba until you can at least give us the performance benefits that Udon 2 was supposed to give us?

And is this an indirect way of saying “we’re not changing our mind” without actually saying it? You have no reference to the main title of the post, regarding pivoting back.

I would be careful how you respond to this, too. You guys, specifically @strasz, said in the Discord server that, and I quote:

We’re not interested in forcing a feature on the community that they’re not into

If you are to continue with Soba, you are ignoring the majority who are speaking out and telling you we don’t want Soba, we want Udon 2, and directly contradicting this statement.


Overall I’m not really satisfied with your answer, it’s extremely vague like most of VRChat’s communications are on your team’s internal workings, and does not address the core of the canny itself.

1 Like

Your post uses language like “waving [their] dick about,” “didn’t bother to debug,” and “sloppy custom VM,” which misrepresents some aspects of the situation and makes it hard to engage constructively.

We’ll be sharing more clarifications soon, which I hope will address your concerns. However, I can’t respond to questions framed this way out of principle.

I do agree that my language in the canny was indeed rather heated and perhaps out of line (and misleading), however it was a result of my (rightful) frustration regarding the matter and some of it may have been slightly misconstrued, but I was unable to reword it as by that point, I’d exceeded 5 upvotes and therefore lost editing privileges.

The main point of the post still stands however. 123 people from the community (and still climbing!) came together and voted in favour of pivoting back to Udon 2. Even if my language was out of line, surely it still speaks volumes to how many people are not fond of this sudden change, and those people’s voices should not go unheard.

I do also have to continue to bump this, though.

It appears the answer to the question of whether my canny’s core topic is even out of the question is still being withheld for some unknown reason, and I (alongside the community) would like to know whether you are firm in your current stance, or if the community’s feedback is actually being considered.

I hope we will get then a deep dive on what caused that memory usage issue and why was it so complex to justify pivoting to different approach. Otherwise I don’t see how that part is misrepresenting.

1 Like