5 ms·

Vector (+Matrix) is kinda like self-hosted slack/discord with proper history sync across all your devices, file uploads, nice UI, etc. It's really good. I reco

Vector (+Matrix) is kinda like self-hosted slack/discord with proper history sync across all your devices, file uploads, nice UI, etc. It's really good. I recommended it to 7 people and every single one liked it, even got 5 of them to set up their own federated homeservers. We're thinking about moving a ~60-people skype group there as well. Only issue I had is Synapse hogging the CPU and getting laggy with a large room (#matrix:matrix.org with its 4 thousand members). I'm using scaleway's 3 EUR/month Starter VC1S server for Synapse though. Hopefully it will get even better with time.

vertex-four 10y ago
Synapse does calm down once it's sorted everything out, luckily.
amandine 10y ago
Synapse 0.18.0 has just been released, might worth giving it a try as it improves things a bit. Thanks for the support :)
kuschku 10y ago
Or you could just use IRC with quassel + quassel-webserver, and get basically all of the same, but with an actual open protocol. Which actually has support.
- mxuribe 10y ago
  Actually matrix - the protocol upon which vector and now riot depend - is an open protocol; reference here: http://matrix.org/docs/spec/ http://matrix.org/docs/spec/ And then of course, there are the matrix bridges to IRC, etc. Admittedly although i have been using vector clients AND have installed my own personal synapse server, i have no experience using/supporting the various matrix bridges, so can't speak to their quality. That being said, if you've ever been curious about re-doing some aspects of IRC for the better, you might want to take a look at matrix (the protocol), and suggest improvements; we all stand to benefit from your (and the community's) suggestions.
  - kuschku 10y ago
    > you might want to take a look at matrix (the protocol), and suggest improvements; we all stand to benefit from your (and the community's) suggestions. My suggestion: stop the project, burn the code. Seriously, by splitting up the dev effort between yet one more project, the open source community ends up with even more competing standards. You don’t even believe yourself you’ll be able to actually replace the ecosystem of IRC and so on yourself, so there will always be competing standards, and Matrix and IRC will never bridge natively. IRC, on the other hand, will reach feature parity with Matrix over time, which would be faster if you’d instead work on IRC. For IRC, currently only 3 useful clients exist where the user has the same features as if they’d run on Matrix. In the end, we’ll have to build hacky bridges the other way around, too, and end up with more developer time spent on interoperation than useful features, and people will continue using Slack. So, my suggestion: Just give up, and help improve things. Rewriting everything just creates a worse result for everyone. EDIT: I can’t answer right now (you are submitting too fast – because of the downvotes, I assume), so I’ll answer inline: > Regarding "Who is working on bouncers in IRC networks, push notifications, etc?" > Well, the IRCv3 Working Group is. How to ask for previous logs, how to handle push notifications in the future, how to integrate seemless bouncers is what they’re doing, and some networks already have seemless bouncers integrated (most notably Snoonet). > Regarding "Forking doesn’t fork the community" > That is true for file formats, but not true for chat systems. For a chat system to work, it needs critical mass. IRC still has that, Matrix doesn’t (otherwise Matrix wouldn’t even build a Matrix-to-IRC bridge). > Trying to split the community will just make things worse there.
    - Arathorn 10y ago
      Occasionally you need to re-baseline or you risk getting mired fighting backwards compatibility and the limitations of the original design. Just like Git "rewrote" Subversion. Or clang "rewrote" gcc. Or PNG "rewrote" GIF. It's incredibly shortsighted to assume that attempts to create new projects that overlap with old ones are are splitting dev effort and harm overall progress.
    - vertex-four 10y ago
      So who's working on federation and single identity across IRC networks? Who's working on a protocol which would let me ask other users for their chat logs from before I joined a channel? Who's working on WebRTC signalling over IRC? When will bouncers be built into IRC servers so I don't have to run my own? When will IRC have a standardised HTTP protocol by which I can run whatever webclient, mobile client, or otherwise that I want? When will it support proprietary push services to keep mobile battery life up? IRC has had 28 years to do any one of these things.

vertex-four 10y ago

Synapse does calm down once it's sorted everything out, luckily.

amandine 10y ago

Synapse 0.18.0 has just been released, might worth giving it a try as it improves things a bit. Thanks for the support :)

kuschku 10y ago

Or you could just use IRC with quassel + quassel-webserver, and get basically all of the same, but with an actual open protocol. Which actually has support.

mxuribe 10y ago
Actually matrix - the protocol upon which vector and now riot depend - is an open protocol; reference here: http://matrix.org/docs/spec/ http://matrix.org/docs/spec/ And then of course, there are the matrix bridges to IRC, etc. Admittedly although i have been using vector clients AND have installed my own personal synapse server, i have no experience using/supporting the various matrix bridges, so can't speak to their quality. That being said, if you've ever been curious about re-doing some aspects of IRC for the better, you might want to take a look at matrix (the protocol), and suggest improvements; we all stand to benefit from your (and the community's) suggestions.
- kuschku 10y ago
  > you might want to take a look at matrix (the protocol), and suggest improvements; we all stand to benefit from your (and the community's) suggestions. My suggestion: stop the project, burn the code. Seriously, by splitting up the dev effort between yet one more project, the open source community ends up with even more competing standards. You don’t even believe yourself you’ll be able to actually replace the ecosystem of IRC and so on yourself, so there will always be competing standards, and Matrix and IRC will never bridge natively. IRC, on the other hand, will reach feature parity with Matrix over time, which would be faster if you’d instead work on IRC. For IRC, currently only 3 useful clients exist where the user has the same features as if they’d run on Matrix. In the end, we’ll have to build hacky bridges the other way around, too, and end up with more developer time spent on interoperation than useful features, and people will continue using Slack. So, my suggestion: Just give up, and help improve things. Rewriting everything just creates a worse result for everyone. EDIT: I can’t answer right now (you are submitting too fast – because of the downvotes, I assume), so I’ll answer inline: > Regarding "Who is working on bouncers in IRC networks, push notifications, etc?" > Well, the IRCv3 Working Group is. How to ask for previous logs, how to handle push notifications in the future, how to integrate seemless bouncers is what they’re doing, and some networks already have seemless bouncers integrated (most notably Snoonet). > Regarding "Forking doesn’t fork the community" > That is true for file formats, but not true for chat systems. For a chat system to work, it needs critical mass. IRC still has that, Matrix doesn’t (otherwise Matrix wouldn’t even build a Matrix-to-IRC bridge). > Trying to split the community will just make things worse there.
  - Arathorn 10y ago
    Occasionally you need to re-baseline or you risk getting mired fighting backwards compatibility and the limitations of the original design. Just like Git "rewrote" Subversion. Or clang "rewrote" gcc. Or PNG "rewrote" GIF. It's incredibly shortsighted to assume that attempts to create new projects that overlap with old ones are are splitting dev effort and harm overall progress.
  - vertex-four 10y ago
    So who's working on federation and single identity across IRC networks? Who's working on a protocol which would let me ask other users for their chat logs from before I joined a channel? Who's working on WebRTC signalling over IRC? When will bouncers be built into IRC servers so I don't have to run my own? When will IRC have a standardised HTTP protocol by which I can run whatever webclient, mobile client, or otherwise that I want? When will it support proprietary push services to keep mobile battery life up? IRC has had 28 years to do any one of these things.