current thoughts after experimenting with some proposed discord alternatives recently:matrix: despite the honourable efforts of instance maintainers, i actually think that client apps are the achilles heel of matrix.
-
fluxer has existed before LLMs became popular, the history was squashed into one commit because the dev did not want all the messy commits they made from school (they submitted this at their uni or whatever) visible.
from the relevant blogpost:
I published the project with a squashed history because the early work happened privately, and I didn't want to make 3,000+ messy commits part of the public record. I'm proud of where things are now, and the codebase has improved a lot over the 3+ years it was developed in private.
Squashing commits during a closed source to open source transition is common practice, and it doesn't imply the project was vibe-coded.
How I built Fluxer, a Discord-like chat app
Fluxer is a free and open source instant messaging and VoIP platform built for friends, groups, and communities.
Fluxer Blog (blog.fluxer.app)
@alexia i've edited my post to correct that statement. it was definitely wrong of me to lean on that assumption without reading deeper. i was unaware of this blog post at the time and kinda rushed a surface-level evaluation of fluxer as i already wasn't a huge fan of the "discord-like" model.
that's definitely a blunder on my part, and i think the best thing i can do is give it a better chance so i can evaluate it more fairly. i'll likely make a successor to this original post when i have a greater idea of the options available, with fluxer as a particular point of interest! -
@piku if the app could be built on top of IRCv3 without stunting its future development (tech debt, limiting features, etc), this sounds good! i haven't looked into those protocols (yet) so can't make an assertion either way
@ari @piku
IRCv3 is fully backwards and forwards compatible, capabilities are negotiated upon connection -- and you can say "for capability X, you need Y and Z", so the whole mechanism is meant to allow new features while also semi-transparently handling older clients as well
so tech debt isn't an issue i'd worry about, and as someone looking into ircv3 and meaning to write my own clients, the main thing missing right now is just, people, writing software and getting involved with writing/reviewing the protocol extensions -
current thoughts after experimenting with some proposed discord alternatives recently:
matrix: despite the honourable efforts of instance maintainers, i actually think that client apps are the achilles heel of matrix. every app is missing features i would really rather have, even if the server supports them all. everything feels like a hack built on a hack, and its fragile ecosystem shows. i can't recommend it if you're moving from discord, but i welcome you to try it out and see if it works for you and your friends!
signal: if your primary use for discord is DMs and group chats, you may feel right at home with signal! i just chatted with my girlfriend for about four hours with webcam and screenshare, and it was a very pleasant experience! do note that the desktop app is rough around the edges; signal is a mobile app first.
mumble (in progress): i have a server i've kept as a voice chat fallback for discord for months- if i can convince my friends to Put Up With the huge interface downgrade, i'd love to test it on its strengths. its weaknesses are quite the blight, though. it's open-source, so i wonder could a third-party client claim superiority with a better UX and design.
fluxer: evidently vibe-coded: the commit history dates back to early january with an inhumane amount of code in its first commit (1.4m lines of code across 9k files. ok bud). assuming that didn't kill it for you already, (bold take based on speculation. read the developer's own blog post on fluxer's development before making up your mind on this) i'm concerned about the long-term maintainability of the app. i expect it to be unstable for quite some time, and self-hosting is not recommended by the developer for the time being. on the upside, they seem like a nice fellow! no ill will here. the app's design also blatantly rips from discord, which i suppose is a selling point to some, but i'd personally like to see something new.
stoat (ex. revolt): same "copying discord" comment from before applies. despite being open-source, it seems to sorely lack self-hosting support, which i consider a must-have for my future community chat app. it's failed to take hold of discord's marketshare since it launched, and i suspect there's a good reason for that.
teamspeak: even ignoring controversial military ties, i echo similar sentiments to mumble, except that teamspeak also fails by being closed-source. for my purposes, it's mumble but worse. at least it's a bit prettier, i guess...@ari@fedi.arimelody.space When I started looking into it, I thought self hosting and encryption being annoying would be the main issue with matrix, but it's 100% the clients, they're fine but dear god it's not something I could use long term
-
@Jes honestly i just don't think federation is a great fit for a community chat app. this may be a controversial take, but i'm actually perfectly happy with the independent silos provided by teamspeak, mumble, and almost all game servers. i don't really think retaining the rooms and members of dead servers is particularly useful, and i think attempting to build software that carries the burden of everyone else's activity like that in the first place was asking for trouble.
@ari the only real way to do federation in chat apps is probably more like XMPP where it's a burden of the user to store chat logs -
@alexia i've edited my post to correct that statement. it was definitely wrong of me to lean on that assumption without reading deeper. i was unaware of this blog post at the time and kinda rushed a surface-level evaluation of fluxer as i already wasn't a huge fan of the "discord-like" model.
that's definitely a blunder on my part, and i think the best thing i can do is give it a better chance so i can evaluate it more fairly. i'll likely make a successor to this original post when i have a greater idea of the options available, with fluxer as a particular point of interest!oh, ty!! not sure if the edit federated only after I replied or if it was bad timing on my part
ty nonetheless!
-
@ari the only real way to do federation in chat apps is probably more like XMPP where it's a burden of the user to store chat logs@ari but yeah connecting directly to servers for chats seems to be the most efficient way of doing it if chat history is a concern
-
@fiore fwiw, while looking into matrix hosting, i found that a webRTC connection over relays (see TURN) was the preferred method of setting up VOIP.
while it would sidestep stock mumble quite substantially, if the architecture is designed to be extensible enough, i don't see why a mumble server couldn't just negotiate a relay link between participants to support even a simple video feed!ithink murmur supporting smth like that could be cool yes
or even just . a standalone thing webrtc server , which then the client implementation could make appear as tho they were part of the same interface . idk mumble is so cool and simple , you could pretty much have a full discord replacement if you mesh mumble irc and webrtc in the right way
-
@ari@fedi.arimelody.space When I started looking into it, I thought self hosting and encryption being annoying would be the main issue with matrix, but it's 100% the clients, they're fine but dear god it's not something I could use long term
@ghost_desu i was genuinely surprised how much the onus was on clients. really ringing home that the FOSS space has more backend developers than frontend designers
-
@ari but yeah connecting directly to servers for chats seems to be the most efficient way of doing it if chat history is a concern
@Jes and perfectly sensible for 99% of what people use a community chat platform for- even on discord, where albeit outages are very infrequent, it's not the end of the world if our node goes down for a little while.
-
oh, ty!! not sure if the edit federated only after I replied or if it was bad timing on my part
ty nonetheless!
@alexia i edited around the same time, no worries! regardless, i'll be taking more time to look into fluxer properly, following this.
-
ithink murmur supporting smth like that could be cool yes
or even just . a standalone thing webrtc server , which then the client implementation could make appear as tho they were part of the same interface . idk mumble is so cool and simple , you could pretty much have a full discord replacement if you mesh mumble irc and webrtc in the right way@fiore oh yeah, to be clear, making it feel integrated would 100% be part of the plan there
-
@fiore oh yeah, to be clear, making it feel integrated would 100% be part of the plan there
tbh i like the idea of murmur being minimal as a server , thats what makes it cool and easy to host
but yea like whatever im making a nothingburger here its not like im gonna go and implement this myself , i should NOT be allowed near ui code
-
tbh i like the idea of murmur being minimal as a server , thats what makes it cool and easy to host
but yea like whatever im making a nothingburger here its not like im gonna go and implement this myself , i should NOT be allowed near ui code -
-
