Modern XMPP project discussion - 2025-02-15

So I currently have a working primitive XMPP client built entirely inside of a VR platform (Resonite), generally as a concept demo to showcase how simple a client could be built (probably not within the ideal 'best practices'), given the platform allows it's in-world visual scripting system to interact with WebSocket endpoints:

https://were.chat:5443/upload/341108c9c1b09ffa44c74d362ab7a69daf1b3b07/7M08SZfnlpe9pr7h8nHHZ9jRfrbKZ7Is4Ec9ehkm/2025-02-13_06.52.21.png

as part of an entry for a month-long 'maker competition' type of thing held on the platform (MMC; Metaverse Maker Competition), whereas I might try to make an adjacent introductory educational world on the subject. It's able to handle PLAIN auth, come online, send/receive direct messages; haven't attempted MUC support yet.

This Resonite is not really open-source, right? I tried finding the main source on their github but it isn't there?

Correct, it's currently not open source, however it may gradually open source components over time, especially overtime as they refactor, move out of Unity, etc

I see. Well, I hope that works out somehow.

meanwhile, mentionably, it has an in-platform chat system that from what I've heard is just a weird abuse of Redis to act like a chat messaging system, but has been flaky and unreliable at times, whereas this could also serve as an elevator pitch to the devs to perhaps consider XMPP (or whatever) as an option

sure, would be nice. XMPP certainly has the credentials in the gaming space to convince people

also I do have 2 more weeks yet to work on it, so I'm not sure if there's any other quick XEPs that could be quick to implement (already have Stable IDs), or if basic MUC support is possible or not in that time ✏

> So I currently have a working primitive XMPP client built entirely inside of a VR platform (Resonite), generally as a concept demo to showcase how simple a client could be built (probably not within the ideal 'best practices'), given the platform allows it's in-world visual scripting system to interact with WebSocket endpoints: Is that through BOSH? ↺

Or what's the XEP for XMPP over WebSocket?

Maybe Overte can do it too

Considering Overte also supports WebSockets

> https://were.chat:5443/upload/341108c9c1b09ffa44c74d362ab7a69daf1b3b07/7M08SZfnlpe9pr7h8nHHZ9jRfrbKZ7Is4Ec9ehkm/2025-02-13_06.52.21.png can you type, in vr? ↺

what i really mean is, can you code in it?

(I would have no interest in it, even for xmpp, without that)

with passthrough MR you can use a regular keyboard quite well it seems

on modern headsets

sadly passthrough is a bit borked on my old HTC Vive Pro under Linux, so I was never able to try it

but I heard it works well on some modern standalone headsets

hmm, alright, thanks

> what i really mean is, can you code in it? Resonite uses a visual scripting language. Overte uses JavaScript. ↺

Squeaky Latex Folf, XMPP over WebSocket isn’t a XEP but a RFC: 7395.

> Or what's the XEP for XMPP over WebSocket? XMPP over WebSockets (RFC7395), yes. ✏ ↺

Whoa, an RFC for that? :o

> what i really mean is, can you code in it? In Resonite, yes, when a text field is focused, a virtual keyboard pops up, and closes when unneeded. It uses a visual node-based scripting language (Protoflux) where the code is expressed as a sort of syntax tree, has a variable system that can be attached to objects with the VR scene, etc. The platform is almost exclusively community funded (aside from some educational users, that pay for a separate grid), it has developers that have contributed or maintained other open source projects (such as previously some of the SecondLife third-party clients), has a far more permissive modding policy compared to other [currently] proprietary platforms allowing modding of the client and more (while under VRChat any modding is bannable at any level, and VRChat went to the extremes to force an anti-cheat system into their platform to auto-moderate that) ✏ ↺

> Whoa, an RFC for that? :o pretty much the shortest of them all: https://www.rfc-editor.org/rfc/rfc7395.html ↺

and also serves as an easy counterpoint to anyone that whines and complains about SAX parsers, that they could just do DOM parsing instead with using WebSockets as the transport

> but I heard it works well on some modern standalone headsets mentionably, Microsoft is outright discontinuing "Windows Mixed Reality" whereas there's several headsets that were exclusively only operational through that mechanism, thus will be rendered 'useless' after upgrading to Win11 24H2, meanwhile there's open source projects for using those headsets directly under Linux, so it's probably an easy opportunity to get VR headsets that work fine under Linux for cheap now. ✏ ↺

I don't think Resonite is a good-faith project. Resonite is a closed-source, centralized VR platform. There have been many instances of abuse or concern. I know a few people that have been banned from Resonite because the staff are just spotty. And because it's centralized it's kind of hard to defend against admin abuse like that. Resonite also has a very invasive ToS and Privacy Policy, e.g. they claim they will collect your browsing history. I opened an issue on their GitHub regarding this (https://github.com/Yellow-Dog-Man/Resonite-Issues/issues/224) but they never fully addressed my concerns. I'd much rather put my time and effort into an actual open source and decentralized VR platform like https://overte.org/

overte is a Vircadia fork?

To my knowledge, yes

Yes

It was forked when Vircadia started adding NFTs and cryptocurrency to the game ✏

Due to disagreements on that stuff

Overte doesn't want any cryptocurrency, NFTs, or DRM ✏

> I don't think Resonite is a good-faith project. Resonite is a closed-source, centralized VR platform. There have been many instances of abuse or concern. I know a few people that have been banned from Resonite because the staff are just spotty. And because it's centralized it's kind of hard to defend against admin abuse like that. > Resonite also has a very invasive ToS and Privacy Policy, e.g. they claim they will collect your browsing history. I opened an issue on their GitHub regarding this (https://github.com/Yellow-Dog-Man/Resonite-Issues/issues/224) but they never fully addressed my concerns. I'd much rather put my time and effort into an actual open source and decentralized VR platform like https://overte.org/ I see nothing about broadly the collection of browsing history in present ToS; the old language you're referring to reads as something regarding 'Referer' header data (whereas when you navigate to a website, your search query _is_ in the Referer header), AdSense campaigns, etc ↺

When you write a Privacy Policy, you have to enumerate every absolute form of data being stored (whether temporary or perpetual), whereas when you throw away context it's easy to read it sideways ✏

I want to be able to dogfood Overte, but there's so much that's still deficient yet and feels like a routine revolving door of developers coming in, dropping out, forking again, etc; I wonder how many have been keeping track of it all the way back to High Fidelity

What are you worried about? Another fork?

Project momentum

So far there's not much drama in Overte

Also, Overte got a grand from nlnet

The momentum is good

There's not much drama when you have like ~5 developers

Vulkan is being worked on. OpenXR support is being worked on.

Yea and several other projects I've seen get NLnet grants and just burn it

IPv6 support is being worked on.

You can say all this stuff like "projects need momentum and if they don't have it I won't care" but if everyone thinks like that, it wouldn't get momentum, if you get what I mean ✏

You can say all this stuff like "projects need momentum and if they don't have it I won't care" but if everyone thinks like that, no project would get momentum, if you get what I mean ✏

Overte is one of the very few real open source decentralized VR games out there that are still maintained

It's something that should need more attention imo

It really sucks that proprietary projects like Resonite are really stealing the show, even in open source / hacker / freedom / furry communities ✏

> You can say all this stuff like "projects need momentum and if they don't have it I won't care" but if everyone thinks like that, no project would get momentum, if you get what I mean as below: ↺

https://were.chat:5443/upload/341108c9c1b09ffa44c74d362ab7a69daf1b3b07/hjgfImGU8FIeSF9aTNlYFMznhq2SG4bnDEqG10hY/Screenshot_20250215_155118.png

commit metrics (not that it's an end-all/be-all metric, but fair general rule-of-thumb with project health and pace)

> Overte is one of the very few real open source decentralized VR games out there that are still maintained To my knowledge it's the _only_ option outside of OpenSim and alike being a handy-down from SecondLife ↺

> It really sucks that proprietary projects like Resonite are really stealing the show, even in open source / hacker / freedom / furry communities I agree plenty on the criticism of it being centralized and currently proprietary, but there are a lot of internal and external pressures on open sourcing it in whole, not just in parts ↺

Do you think they actually will do that?

And even then the architecture is centralized as of now

They'd have to change a lot of things

Overte has a more promising architecture

Also, Overte has no proprietary dependencies like Unity or anything

It's based on Qt

Resonite's architecture and dependencies are much less fitting for opem sourcing

Overte/Vircadia etc is a continuation of the hacked-together codebase that was High Fidelity, which was a commercially-operated project that was open source, in the pace of typical tech startups, then dropped, and the community putting life support and maintenance on it when the company dropped it. Architectually it's not something that I think most people are gravitated towards being able to walk into and start contributing, because it reads like such an over-componentized server stack that if someone wants to do an alt-grid, it's a clockwork of separate components that realistically on smaller profiles should be merged together. It's like the same thing of Bluesky trying to brand itself as a decentralized/federated platform, when architectually it makes it increasingly cost-prohibitive for more actors to set up their own sovereign stack in ATProto

> Resonite's architecture and dependencies are much less fitting for opem sourcing I clarified it's on the road plan that they're dropping Unity for an in-house graphics engine that will be released as an open source project, Unity is only serving as the renderer ↺

How do you know Resonite is more maintainable? Have you seen their source code?

By the fact that running a headless is far less involved than equivalent than Overte, that it doesn't break things up into a separate substreams that wouldn't need to be, that much of it's optimized that you can throw very traffic/event-heavy workloads and have a maintainable FPS while in-world?

Why divvy up so much things as separate components, especially as a server-side responsibility? https://docs.overte.org/en/latest/_images/domain-server.png

unless it's just a recycled graphic from it's High Fidelity years that was probably used to pitch to investors as 'deeply sophisticated and advanced technology' or something, rather than how it actually operates ✏

and I can assure you Resonite is not necessarily "stealing the show" wholly, as majority are clung to VRChat for dumber of reasons, a la "but everyone's on VRChat!"

Lol

As much as I hate conceding a little bit to the use of C# (Resonite): they don't have to fight with some allocation and memory management aspects as with working in C++ (Overte)

And I've heard Resonite's had far more (client) performance issues than Overte. At FOSDEM Overte had a stand where an ARM tablet was showcased running the game at reasonable performance.

If someone wants to start a Rust-based project, maybe that have a better development momentum, but then again I don't know how developed the ecosystem is yet with Rust and 3D graphics

> If someone wants to start a Rust-based project, maybe that have a better development momentum, but then again I don't know how developed the ecosystem is yet with Rust and 3D graphics https://github.com/squeaktoy/hearth boom! I got you ↺

I'm talking like throwing several +80,000 tri avatars at it, 20+ people present with full active IK

> https://github.com/squeaktoy/hearth boom! I got you Nice ↺

It's unmaintained though. The previous maintainer had a breakdown and deleted the repos

Now they don't want to do open source anymore

I am literally the only person on this planet that took it upon themselves to use a GitHub exploit to retrieve Hearth's lost source code

Imagine open source as serving in a public office: you're going to have to take constant criticism, especially from pathetic people who have never made nor contributed to anything in their life, especially people that act like they know your project more than you; and sometimes as a contributor: that you have to cooperate with people that are sometimes miserable people to work with, that you don't get the choice of working with someone else instead, etc

So there's a variety of people that can't handle dealing with constant sh't and tap out. It's unfortunate, but that's just in the nature of putting out any works of your own to the public internet, that it's a process of criticism and differences in ideas, etc.

Hopefully someone else can come along to pick up the project, as unfortunately I'm not a graphics developer; if there's stuff needed for networking, protocol stuffs, web services, identity and auth, etc, that would probably be the only domain I could contribute anything in.

Right now, ActivityPub is the lowest hanging fruit. How it takes Mastodon 5,000 to 50,000 EUR (NLnet grant) and practically a year to _start to_ implement 'quote-posts' (while most others did it in a few months, many years ago), I don't even understand..