| 23 May 2025 |
 Gnuxie 💜🐝 | I'll show why | 10:02:10 |
| Gnuxie 💜🐝 | https://marewolf.me/posts/draupnir/25/02.html#priorities-a-focus-on-on-boarding-users | 10:02:31 |
| Gnuxie 💜🐝 | * | 10:03:49 |
| Gnuxie 💜🐝 | the bigger issue is just proactively getting access to events to stop flooding | 10:04:08 |
| Gnuxie 💜🐝 | from new users | 10:04:17 |
 emily | (these three messages, FWIW) | 10:05:11 |
| emily | ah, but I guess you read them :) | 10:05:19 |
| emily | "but the only have to implement the event forwarding and not the policy server itself" → that's precisely the concern wrt centralization/SPOF right? | 10:05:33 |
| emily | hm, ok. | 10:05:49 |
| emily | then the Matrix blog presentation of the feature confuses me somewhat | 10:05:59 |
| emily | since they explicitly say "In short, they're servers on the internet where you can send events to and have them be checked for spam/illegal imagery/etc." | 10:06:31 |
| Gnuxie 💜🐝 | yeah they intend to use them differently to me I think | 10:06:49 |
| Gnuxie 💜🐝 | and the point is they enable people to make tools for proactive checks | 10:08:00 |
| Gnuxie 💜🐝 | whether those tools are made is something else | 10:08:10 |
| Gnuxie 💜🐝 | since fundamentally it is impossible in matrix atm | 10:08:24 |
| emily | right. well I agree that there needs to be a blocking kind of moderation and that reactive approaches are hopeless, yeah | 10:08:26 |
| Gnuxie 💜🐝 | * since fundamentally it is impossible in matrix atm without policy servers | 10:08:37 |
| emily | in my view, it's not really a good thing for NixOS that rooms are independent of homeserver | 10:08:46 |
| emily | the split-brain issues we've had, and the reactive-only moderation, are things that wouldn't be an issue if nixos.org was authoritative for our rooms | 10:09:10 |
| emily | and the benefits we gain from federation are from users being able to be on different homeservers, not the rooms being distributed across homeservers | 10:09:28 |
| emily | of course this may not be the case for all communities | 10:09:35 |
| Gnuxie 💜🐝 | well that's another argument entirely and we can't go from saying this feature will take too long to roll out to fundamentally changing the protocol x3 | 10:11:31 |
 dgrig | The (main/only?) benefit of the current architecture is the fact that the nixos.org HS isn't a SPOF. Obviously, with this also come all the issues you mentioned (split brains, abuse, etc). | 10:11:47 |
| emily | (also wow I did not know that Matrix had spent that much on core dev for what a trash fire everything still is. how come they constantly give off the appearance of being critically endangered from being strapped for cash?) | 10:12:05 |
| emily | oh for sure. the approach seems sensible given the existing constraints | 10:12:19 |
| Gnuxie 💜🐝 | In reply to @emilazy:matrix.org
(also wow I did not know that Matrix had spent that much on core dev for what a trash fire everything still is. how come they constantly give off the appearance of being critically endangered from being strapped for cash?) bc that figure is out of the ceo's ass and was used to build a for profit buisness, not core matrix dev | 10:12:42 |
| emily | right. for us, I don't think that is any benefit at all. if the centralized NixOS infra goes down everything else is already on fire | 10:12:48 |
| emily | like, we also rely on nixos.org for builders, the cache, the Discourse forum, … | 10:13:03 |
| emily | so Matrix sticking around in the event of a total nixos.org outage is, meh | 10:13:17 |
 @magic_rb:matrix.redalder.org | What could work is if homservers could take ownership of rooms but if the homeserver is unreachable then it would delegate to another HS, kind if like a list of fallbacks | 10:13:57 |
