Update: since there are no volunteers to run it / pay for it, and the general consensus has been to shut the instance down, I’ve gone ahead and replied to the email letting the Lemmy folks that were generously hosting it for us know that we appreciate it but that they can shut down the instance. Thanks everyone for participating and being a part of this community experiment!

I think the general consensus so far is that the instance should shut down, but there have been a few commenters suggesting that they want it to stay. I tend to think people far undervalue the cost of hosting their own services, so to try and make it a bit clearer if someone did want to keep it up I think they’d need to consider the following factors:

• Multiple people to cover the cost of hosting, this needs to scale with the amount of traffic
• Multiple volunteers to handle operations, this needs to scale with the amount of traffic
• Multiple volunteers to handle moderation and accept registrations, this needs to scale with the number of users
• Someone willing to promote the instance (assuming we want it to grow, this may or may not be the case and it’s fine if we want it to just be a small space for a few XMPP projects), this workload will inversely scale with the number of users
• Some sort of governance and accountability model; this would need to scale with the number of users

Some of this we’ve gotten away without so far, and some of it can start small and scale to multiple volunteers later if the instance grows, but I really think you need a few people on all of them to prevent burnout and keep the community sustainable. I do think there’s harm in just leaving the community around to languish: this makes it a target for spammers if it’s poorly moderated, contributes to making the network look large but dead (as opposed to small and growing), and, if it becomes unstable, may create a bad experience for anyone using it when it’s down a lot and there aren’t volunteers to fix it.

If we have trusted volunteers (or maybe we can find a way to hand over the database dump without giving away anyones personal data, reset passwords and purge profiles or something; this is less important to figure out right this moment and we can figure it out only if/when we go to do the handover) for any of this, I think we can probably hand the reins over to them. If not, we can’t keep the community going whether people want to or not :)

I’m not sure; same as anything else, I’d imagine: moderation and ops is probably the most time consuming, then if course the fiscal cost of the server and image hosting

That requires us to find someone to take over hosting and maintenance, which is what the post was about.

That’s what this post is about; if someone wants to keep it, we can, but it does cost a lot to run services like this (financially, in moderation hours, in operations, in security, etc.) and a lot of people would have to take that cost on. Given that there hasn’t been much engagement at all on this server, I suspect we don’t have anyone who wants to keep it running.

My personal opinion is that the results of the experiment are that Lemmy is not a good platform for community building. I’m glad we gave it a shot, but I probably won’t be using it anymore either way for a few reasons:

• The moderation just isn’t very good (we still have Nazi imagery and accounts on the server from the early spam wave with no way to purge them, though they are blocked and can’t post anymore but the images and what not are still being served as far as I can tell)
• The UI is generally confusing and hard to use, things aren’t logically grouped, etc.
• The federation capabilities don’t seem to actually be very good (I’m sure this will improve, but there’s other software out there that’s good already so if we wanted to try this again I think we should use something else instead)
• Generally buggy/over-javascripty UI
• Generally hard to follow new content on Lemmy: it tries to imitate reddit too much I think so stuff that’s “Hot”, whatever that means, get surfaced but you have to constantly switch over to a non-algorithmic timeline if you just want to see what you’ve missed

I’d still like to find a way to build community between XMPP projects, but I’ve started to think it may be better if projects interact with other existing instances more that aren’t focused on XMPP, this spreads the message a bit better. I forget who made this argument early on, so apologies for not letting you know directly, but I think the experiment has brought me around to this way of thinking as well. I will follow up after we decide what to do with a list of possible places that it might be good for XMPP projects to join if I can remember to put it together.