The person that runs the website it is posted on.

https://joinsharkey.org/ seems to be the only one that survived so far.

First you need to check if your ISP gives you a public IP (can be temporary, if you are fine with using dyndns). Otherwise you will need some tunneling service or run such yourself from a VPS.

That article is highly misleading.

A good response can be found here: https://www.moparisthebest.com/against-silos-signal/

Ejabberd is definitly for more advanced users, but you can usually get help over at joinjabber.org (which has a webclient that you can use without an xmpp account).

But yes, Snikket and Prosody is easier to set up.

I mean, XMPP is great, but if you are only interested in private chats with your family in a virtual LAN then it (and Matrix) is probably overkill and a bit of a hassle to configure without a public domain (as that is expected for federation).

Maybe a Nextcloud Talk or a Jitsi Meet instance would be simpler?

Unless you have experience with ethernet equipment and such it is probably better to start with some hosted service of an open-source app like Nextcloud or Immich or (slightly more advanced) a VPS somewhere. Doing it immediately from home with your own server has a steep learning curve.

Needs Lemmy support 🤷

The reverse, i.e. getting outage notifications and so on via xmpp is quite common.

I can’t really think of any usecase to control self-hosted apps from a messenger, but sure, technically you could write such a bot quite easily.

XMPP is truly decentralized with no single point that collects this kind of data. There have been some attempts over the years, but they always failed to capture more than a fraction of the network. A recent one is more artistic in nature, probably realizing the futility of it.

The popular Conversations app has sold 100k+ units on the PlayStore, but since you can also legally get it for free, that is probably only part of the actual users.

Please don’t recommend jabber.org, it is super outdated and lacks almost all modern features.

There is also a similar library for qt: https://www.kdab.com/embedding-servo-in-qt/

And you can use Servo in Tauri. Quite awesome developments lately.

What do you mean with bots in this context?

We have an [email protected] community here on Lemmy. Our SLRPNK instance also gives an XMPP account to every member automatically and hosts a Movim webclient for easy access.

XMPP is certainly more popular for private groups and 1:1 chatting so you will not find that many large public channels, but there is a search engine here: https://search.jabber.network/tags/

The JoinJabber project also has a curated list of recommended channels and communities: https://joinjabber.org/docs/faqs/rooms/

No, they are shutting down their publicly hosted infrastructure and say that their project is “finished” anyways, so it doesn’t matter that much as a justification. But the main point about the post is the public facing infrastructure and how they lost motivation to run it.

What? They explicitly talk about shutting down their self-hosted infrastructure which includes two git services and other targets of AI scraping. Did you even read the post?

It makes a lot of sense. Both the git repos that they hosted and things like a RSS feed-reader are things that are the prime target for AI scrapers and the same time quite database query heavy on the backend so that the scraping really has a big impact on the costs of running these services.

And yes source-code is among what is the most targeted data to ingest by AI scrapers, mainly to train coding assistants but apparently it also helps LLMs to understand logic better.

Played around a bit with hosting Flohmarkt and Manyfold. Promising fediverse projects, but still a bit early days.

Also looked a bit into running Piefed via Podman, but didn’t progress much with it. Looks easy enough though in general.

Yes and the reason they state sounds a lot like AI scraping made hosting public services such a PITA that they lost motivation to continue doing it. Lots of long running projects that used to require very little maintainance are now DDOSed by these scrapers.

1 . As you could have guessed from the community this was posted in, a software that is part of the fediverse. It has various modules for different functionality. 2. Because they made a long blogpost that explains the new features that you hopefully read before commenting?