| 24 May 2025 |
 uep | but.. ugh, there really should be a better way | 09:45:27 |
| uep | and it's not even clear to me that this will work reasonably for users unless more clients support spaces | 09:45:50 |
 emily | does Draupnir support per-room plicy? | 09:46:07 |
| emily | * does Draupnir support per-room policy? | 09:46:08 |
| emily | oh, I guess Cat implied no | 09:46:15 |
 Gnuxie 💜🐝 | In reply to @emilazy:matrix.org
does Draupnir support per-room plicy? no | 09:46:22 |
| uep | nope. | 09:46:42 |
| Gnuxie 💜🐝 | and it likely never will | 09:46:52 |
| emily | I think we can just point people to the entrance room on the website etc., and link to the space from that | 09:46:55 |
| emily | the space just acts like a room if your client doesn't support spaces, AFAIK | 09:47:03 |
| uep | part of it is simply that they store that config as custom events in a special admin room, but it's still just laziness in the implementation of that | 09:47:42 |
| emily | but also this sucks – people can join the unofficial Discords in two clicks. do we really want to cede the ground of a place people can actually quickly ask a Nix question to them? | 09:47:44 |
| emily | (I know we might have no better option for Matrix. I'm just saying, I feel like the asking-a-question-on-Zulip-from-scratch experience starts to look better.) | 09:48:04 |
| emily | ("Matrix federates so people can bring an existing account" – yeah but entering an email and password is less work than joining a lobby and asking to be let in!) | 09:48:23 |
| uep | all true | 09:48:36 |
| emily | (because of code architecture issues, or design decision?) | 09:48:46 |
| uep | as is the fact that any of this faff is letting the spammers win | 09:48:53 |
| Gnuxie 💜🐝 | there's no "letting the spammers win" please stop. Yes Matrix is unsafe, yes the people making decisions were irresponsible and stupid. None of that is new and spammers didn't change it. | 09:50:43 |
| Gnuxie 💜🐝 | In reply to @emilazy:matrix.org
(because of code architecture issues, or design decision?) both sort of | 09:51:02 |
| emily | well, letting them post CSAM is also letting the spammers win. we do have to respond in some form | 09:51:57 |
| emily | but I agree with what I think Gnuxie 💜🐝 means, which is that ultimately this is Matrix project failures | 09:52:09 |
| emily | but it is also true that I am sure that if we had someone intent on spamming a Zulip with CSAM they'd succeed and we probably wouldn't have great tools to deal with it there either. | 09:53:11 |
| emily | so there is certainly a cat-and-mouse, pragmatism element: so far nobody seems to actually want to target NixOS that badly, only Matrix | 09:53:28 |
| uep | the spammers are (to the extent we can guess any motives) trying to drive users away and make matrix a worse experience. To the extent that we make the new user joining experience harder in response, is some sort of success for them. That's all I mean. | 09:54:19 |
| emily | OTOH, insofar as Matrix's failures are exposed by spam waves that cause projects to make a pragmatic decision to use something else to dodge them, that seems like basically the correct thing to happen to Matrix because of those failures. | 09:54:20 |
| emily | I mean, of course ideally the pipeline from bad project management to people not using your thing doesn't involve abuse spreading illegal imagery to hundreds of servers… the spammers can go to hell regardless :/ | 09:55:16 |
| emily | but I don't think we should have an attitude of "we need to stay on Matrix to not let them win". (I don't think anyone has this attitude though… so maybe what I'm saying is pointless) | 09:55:46 |
| uep | right | 09:56:12 |
| emily | we moved off freenode awful fast when it turned out a certain sysop had made the dubious decision of selling the network to the Crown Prince of Korea :p | 09:57:53 |
| Gnuxie 💜🐝 | We are developing the gradual access control protection for draupnir. But it is slow and we aren't actively working on it because we also hold the burdens for homeserver admins at the moment (to stop invitations for one thing). And there are lots of other things that need immediate attention. So I can't actually promise anyone when it will be ready. | 10:04:22 |
