I keep hearing on Mastodon and elsewhere how the Lemmy devs like to close github issues prematurely, are condescending to people when responding to the issues, host CCP, North Korean, and Russian propaganda on their instances and delete any criticism of those regimes ETC. Even if you don’t care about the propaganda stuff, the rest does not bode well for accessibility fixes or continued cooperation and communication between RBlind mods and lemmy staff in the future. It’s concerning, to say the least.
Yes, from kbin.social. Good luck running your own version of the kbin.social website on kbin though, the way that rblind.com is running Lemmy. Until the documentation for this process is free of sections that are just marked “todo”, and the entire thing has a replicable and well explained deploy script, it’s not even worth further discussion.
@fastfinge From what I see it’s only the firewall and mercurejs which have to-do section. Not really obscure parts
If you already have previous knowledge of hosting anything serious publicly, it’s not a problem.
But it’s part of my “less mature” in my top level comment
What bother me more is the API not working
Yes, that and the main kbin.social instance turning off federation for a while. Between that and the way keeping kbin.social up and running seems prioritized above making running instances easier, I dunno. I just don’t get the right vibes. Similarly, when people join kbin, they’re just encouraged to go to kbin.social. Unlike Mastodon or Lemmy, I haven’t really scene the developers recommending or even listing other instances. And as someone who doesn’t use PHP for much of anything anymore (and doesn’t want to), no, none of this is straightforward. Compare it to Lemmy, where deploying your own instance is just editing a couple files and running a single script away. Of course then you get to spend five days realizing the default nginx configuration isn’t useful for anything more than a single-user instance, and the postgresql config really needs tweaking, and so-on and so-on. But at least one-click Lemmy installs, helping to keep the fedevirse distributed, are within reach. It doesn’t feel that way at all for kbin. And I don’t want Kbin (or anyone else) to have control of the fediverse on one or two large instances. Also, RabbitMQ also has a todo section. I’ve been hosting things online for something like 15 years, but I have no idea what RabbitMQ even is, never mind how to set it up, or what the implications of running it might be, or how I might monitor it, or what could happen when it crashes. Rust, nginx, and postgresql all feel like well-understood, popular, and well-supported technology. If something goes wrong with them, I have pretty much endless support via Google, or if worst comes to worst, it would be easy to find a contractor to hire. The technology kbin is using feels, at least to me, much less known and understood. OK, Lemmy does use pict-rs, and I’m not aware of anything else that does, but pict-rs feels simple enough that it probably doesn’t matter, and it probably won’t fall over in ways that I would need help fixing.