| 21 Feb 2024 |
 @fractivore:cyberia.club | I just keep thinking, one or more humans probably need to actually go debug one or more matrix client's implementation of threads and/or implement threads in clients that don't have them. Easier said than done, I know... | 06:39:45 |
| @fractivore:cyberia.club | Between:
- Threads work perfectly on element but are very spotty everywhere else and possibly not implemented at all
and
- Threads have some support on most clients - the set of clients that accounts for 99% of users or so. These clients support threads in some capacity, but it might be a fallback like using nested replies, or it might be janky and buggy.
...
Which would be the better outcome? Would either of these conditions be sufficient for us to start using threads, or are we looking for a better state than either of those? (i.e. 3. Threads have out-of-beta support on the 99% set of clients)
| 18:09:11 |
 @joepie91:pixie.town | the issues with threads are not just implementation issues, there are also protocol-level problems and a lot of history behind them | 18:15:53 |
| @joepie91:pixie.town | one of those problems being their inflexibility on weaving in and out of threads | 18:16:21 |
| @joepie91:pixie.town | the extremely summarized version is that they ended up implementing Slack-style threads instead of some other model for what seems like business-related reasons, and so we ended up with a thread model that just fundamentally doesn't work well (or at all) for a lot of people, and that is exhausting to interact with even if the implementations worked perfectly | 18:17:49 |
| @fractivore:cyberia.club | Okay... but surely the general concept of threads is something people want in general, right? What would we need to get there? Is it just hopeless due to the state that matrix has ended up in ("governance issues" or whatever), and we'll just have to wait for the next generation federated instant messenger? | 18:30:54 |
 K900 | Honestly I don't think threads are fundamentally broken, you could build a Zulip-style UI on top of Matrix threads | 18:32:00 |
| K900 | Just no one has done it yet | 18:32:03 |
| K900 | But Slack style "spin off entire conversation" threads are definitely what the "intended UX" is | 18:32:18 |
| K900 | So most clients implement that, the stupid fallback or nothing at all | 18:32:28 |
| @joepie91:pixie.town | I would say that there is a general desire for threads that work better, yes - concretely, getting there would require spec work and implementations outside of the upstream governance process, because that is kind of stuck at the moment. I don't think that is hopeless - but it does mean that this is not really a short-term solution and that "just use threads as they are now" isn't really one either | 18:32:56 |
| @joepie91:pixie.town | In reply to @k900:0upti.me
Honestly I don't think threads are fundamentally broken, you could build a Zulip-style UI on top of Matrix threads from what I understand, some people have looked at this and came to the conclusion that it is not viable without some protocol-level changes | 18:33:16 |
| @joepie91:pixie.town | but getting anything through the spec process right now is a challenge | 18:33:53 |
 raitobezarius | In reply to @k900:0upti.me
Just no one has done it yet I'd do it if I would not foresee that doing it anywhere relevant would cause everyone else to have a utterly bad experience | 18:34:28 |
| raitobezarius | In reply to @joepie91:pixie.town
from what I understand, some people have looked at this and came to the conclusion that it is not viable without some protocol-level changes Yeah, I'd say you could do it now without protocol level changes but that would be bad for everyone else without those protocol-level changes | 18:35:02 |
| K900 | Yeah, the problem is getting clients to agree on how to handle threads | 18:35:23 |
| K900 | Which they notably already don't | 18:35:29 |
| @joepie91:pixie.town | right | 18:35:46 |
| raitobezarius | Connected issue to the discussion: https://github.com/zulip/zulip/issues/28800 | 18:36:09 |
| K900 | Honestly I really want Zulip style threads on Matrix | 18:36:50 |
| @joepie91:pixie.town | (very quick grump: the way Matrix is designed has made it heavily reliant on the core spec process to resolve disagreements like this, but their design choices for the governance means that this spec process does not actually have the capacity to handle those cases for which it has been made a critical dependency) | 18:37:08 |
| K900 | But I'll take Zulip if we can't get that | 18:37:03 |
| raitobezarius | (which absolutely doesn't address what you are talking about which is client does whatever they want and even if everything was fixed on bridging side, etc. — people using their preferred client could have an absolutely disastrous experience) | 18:37:15 |
| @joepie91:pixie.town | * (very quick grump: the way Matrix is designed has made it heavily reliant on the core spec process to resolve disagreements like this, but their design choices for the governance means that this spec process does not actually have the capacity to handle those cases for which it has been made a critical dependency. it wouldn't be such an issue if the governance design were better) | 18:37:25 |
| raitobezarius | I think a certain ship has sailed when it comes to Matrix and fighting to have Zulip seems to be hard those days | 18:41:14 |
| raitobezarius | Maybe before anything, Zulip has to be integrated into things like weechat, etc. so that it is easy to add it as a workspace, idk | 18:41:42 |
| 26 Feb 2024 |
|  aciceri joined the room. | 16:14:52 |
| aciceri | Hey, I'm co-organizing a meetup in my city (that I hope it could be a recurring event) and we were considering to create a matrix room instead of re-using the national one, would it make sense if it was hosted on nixos.org? In that case would it be possible? Who should I ask? | 16:18:13 |
 Alyssa Ross | That does make sense, and you should ask in #matrix-suggestions:nixos.org. | 16:20:25 |
| aciceri | Ok let me ask there, thanks | 16:22:07 |
