| 24 May 2025 |
 K900 | But really my concern is more of data storage | 10:36:35 |
 emily | the backstop is still Google/Apple cloud services for anyone who doesn't want crazy battery drain etc. and also in practice everyone relies on one thing | 10:36:46 |
| K900 | And the legal implications of that | 10:36:48 |
| emily | of chat history and media? | 10:36:55 |
| emily | we already accept that with Discourse | 10:37:00 |
| K900 | Like, if we run our own Zulip, we'd have to deal with GDPR and such | 10:37:06 |
| K900 | In reply to @emilazy:matrix.org
we already accept that with Discourse We don't operate that | 10:37:12 |
| K900 | Flying Circus does and they accept it | 10:37:24 |
| emily | well fair re: Discourse | 10:37:30 |
| emily | but, how does that not already apply to the NixOS homeserver? | 10:37:33 |
 uep | there are no users there other than mjolnir and admin | 10:37:50 |
| emily | yeah but they're still in all the rooms | 10:37:56 |
| emily | caching everything | 10:37:59 |
| emily | I wouldn't be surprised if there's illegal material still in those caches unless anyone has been dealing with that | 10:38:10 |
| emily | like ok yes I get that there'll be ops overhead from running anything self-hosted… in which case we could also get the free FOSS cloud plan instead? | 10:39:00 |
| emily | but I figured self-hosted was a desideratum rather than a drawback | 10:39:07 |
| emily | they have data export/import so we could move from cloud to self-hosted and vice versa | 10:39:26 |
| emily | I know nothing is perfect and that any move would be a big decision but we could pick the nits of Matrix's flaws for days like this too. I'm just trying to figure out what the actual hard constraints are. if there's no end to these Matrix abuse issues in sight, it seems to me that it is really worth spinning up a test instance of something that seems to plausibly meet requirements and just feeling it out to see what the experience is like in terms of onboarding, mobile, desktop, what tools the moderation team will have, … | 10:44:10 |
| uep | an experiment with zulip was done for the discussions around setting up the SC | 10:44:59 |
| uep | i don't know what the results of that were, in terms of infra and hosting and moderation facilties, etc | 10:45:26 |
| uep | or user feedback | 10:45:30 |
| emily | right. though I think there were a whole bunch of confounders about that situation | 10:45:34 |
| emily | wrt what the user experience was like | 10:45:40 |
| uep | people tend to either love or hate zulip as a chat platform | 10:45:52 |
| emily | (since it was playing host to some of the most contentious discussions the project has had) | 10:46:05 |
| emily | (and also was invite-only etc.) | 10:46:18 |
| emily | I dunno. it just seems to me that we could easily set up "unofficial not the NixOS Zulip" and get the mod team on there and just sort of play around it with it a bit to see if we can imagine it replacing the Matrix. (I realize that this requires someone to actually do it and that I am dangerously close to volunteering myself.) | 10:47:21 |
| emily | and again I'm not married to Zulip here, it could be Mattermost or Rocket.Chat or whatever else | 10:47:33 |
| emily | it does seem like Zulip is one of the best options for various needs we have and maybe more likely to handle the scale of discussions and rooms we have | 10:48:02 |
| uep | i think it does just need someone to do that, more or less (several someones, but still) | 10:48:20 |
