you are viewing a single comment's thread
view the rest of the comments
[+] 146 points 8 months ago (62 children)
  • [–] 47 points 8 months ago (9 children)

    Yeah honestly. Running the teamspeak server executable is hardly selfhosting, and they're just another closed source proprietary service. Cool they're still around after all these years I guess but they shouldn't even be considered as a migration option.

    Stoat and Fluxer are both open source, very straightforward and familiar, and I believe self-hostable. Much easier for casual users than Matrix too.

  • source
  • parent
  • hideshow 9 child comments
  • [+] 34 points 8 months ago* (3 children)
  • [–] 21 points 8 months ago

    Yeah honestly. Like I work in IT, have my own home server, run linux on everything, etc etc etc, but even I found Matrix to be a convoluted mess, and most clients have their own issues. I can't imagine trying to get someone who's not tech-savvy to try it out.

  • source
  • parent
  • [–] 2 points 8 months ago

    I tried to get friends into Matrix before and they were confused by it. So can confirm it's not that simple for everyone.

    Though to be fair, one of them wasn't able to post without including an emoji in their message, that shit confused me too.

  • source
  • parent
  • [–] 4 points 8 months ago* (5 children)

    I'm unsure what is difficult about Matrix.

    I've had several "casual" friends register and join my space on their own.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 25 points 8 months ago
    • Signup can be tricky.
    • Managing your encryption keys is tricky for normal people (I know someone who signed up for it on incognito because they weren't sure about it yet then got a bunch of confusing popups when they signed on with their phone).
    • Room organization is missing a layer used on discord (server->room instead of list of rooms) leading to confusing moderation structures and nearly required manual organization of rooms if you're in more than 10.
    • notifications rules can be obtuse.
    • having different commands based on the clients used can lead to confusion.
    • most clients have security related popups that just confuse people (This person reset their identity!).
    • people can struggle with how to properly interact over federation, much like in the fediverse
    • screensharing tools just aren't there yet

    Things have been getting better fast for matrix, but its just not ready for the masses IMO. I still suggest it when I can when the use case makes sense.

  • source
  • parent
  • load more comments (4 replies)
  • load more comments (-1 replies)
  • [–] 28 points 8 months ago (6 children)

    Using Stoat's main server raises a privacy concern because it’s UK-based and AFAIK lacks E2EE—UK authorities could seize server data without our knowledge. That effectively means private use requires self-hosting.

    Issue with self-hosting Stoat is, it's currently more complicated than Matrix. This user created a detailed GitHub guide that documents their research and pitfalls for getting Stoat working with voice/video: https://github.com/javif89/stoat-selfhost

    The official self-hosted guide (https://github.com/stoatchat/self-hosted) looks simple at first, but if you look at the compose file, it requires FOURTEEN containers to run and doesn’t yet include voice/video support which will increase complexity.

    By contrast, TeamSpeak’s self-hosting appeal is its simplicity: only two services (or one with SQLite) and it works out of the box today.

    But I agree — moving from one closed-source silo to another isn’t ideal. I just wish Stoat were easier to run behind the scenes.

    For me, a combination of matrix for text chat and mumble for voice is the simplest and most privacy respecting way to self-host a discord alternative.

  • source
  • parent
  • hideshow 6 child comments
  • [+] 13 points 8 months ago* (4 children)
  • [–] 4 points 8 months ago* (2 children)

    There's also Movim, which doesn't even require an email, you can join instantly with just a username and password.

  • source
  • parent
  • hideshow 2 child comments
  • [–] a comment that isn't shown here: deleted, from someone you've blocked, or on another page. view the thread from it
  • [–] 2 points 8 months ago* (last edited 8 months ago)

    Fluxer is doing the same thing, no email signups right now on its homepage.

    It's no different from how lemmy/piefed function. Some instances require email, others don't. My instance, as an example, doesn't require an email to sign up, but it does require you to write a short message as to why you're interested in joining the server, and what communities are appealing to you. This weeds out 99% of bots or spammy users, and the handful that get through that are quickly banned.

    Movim currently has so few users that the main server is trying to put as few barriers as possible to adoption, other servers can and do enable the Email requirement.

    If it becomes more popular and bots or spam accounts become an issue, they could easily activate the email requirement, or even implement a system similar to what I described above. Instances that don't take appropriate measures to those threats as they become a problem can just be defederated as they are here. It's worked out pretty well so far.

  • source
  • parent
  • load more comments (1 reply)
  • load more comments (1 reply)
  • [–] 5 points 8 months ago

    The official self-hosted guide is actually quite simple and straightforward. I had it set up and going in a half hour or so, and that's even with removing Caddy and using my existing nginx reverse proxy. It's intimidating at first-glance, yeah.

    That being said, the official self-host guide is also 5 months out of date. The alternative you linked requires jumping through a bunch of hoops because it's just a small community of enthusiasts hacking together the current version of Stoat for self-hosting.

    So I acknowledge that self-hosting current version of Stoat with voice is rather complicated and frustrating right now, but hopefully it becomes as simple as the official self-hosting guide eventually.

  • source
  • parent
  • [–] 8 points 8 months ago (18 children)

    I'm not sure which horse to bet on Stoat or fluxer.app.

  • source
  • parent
  • hideshow 18 child comments
  • [+] 6 points 8 months ago (7 children)
  • [–] 15 points 8 months ago (6 children)

    One thing that worries me a little about fluxer is this:

    Finally, we can offer commercial licences to companies that want to run Fluxer internally without being bound by the AGPLv3 copyleft terms. This is enabled via a contributor-friendly CLA, but it doesn't create a separate "enterprise edition". It's still the same Fluxer software everyone else uses.

    They have a CLA on contributions. So while today Fluxer is licensed as AGPLv3, tomorrow they can pull the rug and change the license, just like everyone else has been doing.

  • source
  • parent
  • hideshow 6 child comments
  • [–] a comment that isn't shown here: deleted, from someone you've blocked, or on another page. view the thread from it
  • [–] 1 point 8 months ago* (last edited 8 months ago)

    They can’t retroactively close-source the older versions released under AGPL, but if it ever required a community fork to continue the last release of the GPL version, it would be a massive burden to maintain it, and could cause federation to break as the codebase diverges over time, which would create a rift in the community. You’d also have to hope that average users care enough about the license to jump ship to the GPL (probably now not as full-featured) version, otherwise the GPL version risks not being able to get enough funding to continue, or enough users to convince the larger communities to move over.

    As a somewhat similar real world example, the pixel-art program Aseprite once used a FLOSS license, but it switched to a proprietary license at some point. The last GPL version was forked by the community, but it never got much traction, and is now massively behind the closed source version in features and userbase.

  • source
  • parent
  • load more comments (2 replies)
  • [–] 4 points 8 months ago (9 children)

    XMPP!

    Stoat is dead in the water due to dependency on the UK and not an easy solution to deploy yet.

    Fluxer is dead in the water due to license.

  • source
  • parent
  • hideshow 9 child comments
  • [–] 3 points 8 months ago (8 children)

    I don't get why so many people are saying this. Afaik, it doesn't have channels within servers like Discord and Slack, which I feel is a defining feature in the text chat part of the apps.

  • source
  • parent
  • hideshow 8 child comments
  • [–] 2 points 8 months ago (7 children)

    channels within servers

    Oh that is like the second most common thing on XMPP! It's rooms/chats/conversations on servers/conferences/salons, etc. Like, come on, even IRC has that and that was made before I was born.

    The one thing that's complex, or at least bad in the UI I've seen for most XMPP clients, is that searchability of rooms is not very good. Like, discoverability is, but to my knowledge there's no way to actually filter for rooms based on a keyword, you either get the whole roomlist for a server or nothing.

  • source
  • parent
  • hideshow 7 child comments
  • [–] a comment that isn't shown here: deleted, from someone you've blocked, or on another page. view the thread from it
  • [–] 5 points 8 months ago* (5 children)

    I think Iambalicious may be confusing terms. AFAIK no XMPP client has discord-like rooms within channels. The Movim client is actively working on implementing that feature (it can also do group video calls and screen sharing), but it's the only one doing that unless I'm mistaken.

  • source
  • parent
  • hideshow 5 child comments
  • [–] a comment that isn't shown here: deleted, from someone you've blocked, or on another page. view the thread from it
  • [–] 1 point 8 months ago* (2 children)

    There's a bit more to it than just their visual organization. In Discord, a user only needs to join a single community to access all of that community's rooms (they don't have to manually join each one to have it in their feed).

    The admins of that community can then seamlessly create or delete rooms within that community (the users don't need to do anything for those changes to be seen and applied on their end), and can independently adjust what the base requirements are to view, enter, or interact with each room, and then give an individual granular permissions of what rooms are visible within that community.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 8 months ago (1 child)

    Oh yeah that already exists in XMPP, it's called "Server spaces" and grouping can be done by admin or by room owners, but it's experimental so there's no direct way to know what servers or clients are using it or not.

  • source
  • parent
  • hideshow 1 child comment
  • load more comments (1 reply)
  • load more comments (1 reply)
  • load more comments (-1 replies)
  • [–] 12 points 8 months ago

    It’s all about friction. As long as the user has to pick an instance they will always hesitate to pick any federated service. The average user will always choose the path of least resistance.

    Proprietary services spend a lot of time trying to reduce friction, and it works.

    The only solution I can think of would be a three part one:

    1. The main app of a federated service automatically rotates between a pool or reliable, reputable, non-extremist instances where the user can log in with an email and password.
    2. The federated service makes it trivial to migrate accounts amongst instances.
    3. the user can log into their instance threw any other instance perhaps threw oauth.

    This would of course require some federated account login system. Hard but not impossible. It could be some sort of Casandra style ring based account service where nodes are part of the ring.

    This eliminates the new user friction.

    1. Download app
    2. Sign up
    3. Login

    It works anywhere any time with corpo style low friction. You don’t need to think about instances at all till you are ready to.

  • source
  • parent
  • [–] 5 points 8 months ago (7 children)

    Stoat, both its app and website refuse to open on my mobile data. I doubt it's only happening to me. Teamspeak at least lets people host and have control of their own servers.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 3 points 8 months ago (2 children)

    Im seeing a huge increase in people using stoat. And its been fun.

    If anyone is interested in Retro Games: https://stt.gg/GJh5JHy2

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 8 months ago (8 children)

    Yeah... I hate watching people make the same mistake over and over. I guess we just have to take the lead and build the communities that we need over on Stoat and Matrix.

  • source
  • parent
  • hideshow 8 child comments
  • [+] 3 points 8 months ago (3 children)
  • [–] a comment that isn't shown here: deleted, from someone you've blocked, or on another page. view the thread from it
  • [–] 2 points 8 months ago* (1 child)

    having a tool intended for a purpose doesn't necessarily mean it fits that description without issues. the xmpp world is filled with a heterogenous plethora of clients of various quality and encryption duct-taped on top of about a third (made up/guesstimated number based on nothing but my gut feeling having gone down that rabbit hole partly). even if i convinced all of my friends to switch, there wpuld be no client for all platforms and learning miltiple programs to use a single protocol is bad UX imho

  • source
  • parent
  • hideshow 1 child comment
  • load more comments (2 replies)
  • [–] 2 points 8 months ago

    I just spent a week trying to set up my own server and good lord it was such a battle I gave up. Matrix? Up in like half an hour. Shame because my friends are so much more interested in Revolt lmao. Just gonna give them some time to sort out their business before trying again.

  • source
  • parent
  • load more comments (2 replies)