@technology Is there a Usenet like app/website/whatever of the Fediverse?
So, I’ve sort of grown up in the Usenet, and this whole Federation thing sort of feels very reminiscent of it. To the point where I’m wondering why there isn’t a way to access it in this fashion? Any tips would be very helpful. I love the idea of this thing and would probably even begin running my own server if that becomes viable.
Upon further consideration, the smart way to do this would be to have the servers exchange community lists, with post subjects and other metadata. That’s enough information to find and evaluate a community. Post bodies and comments can then be fetched on request.
Putting all the data in the client strikes me as more bandwidth-inefficient, but whether that matters these days is questionable. Usenet clients don’t download subject lines from the server until after you subscribe to a group, but that’s a relic of the days of modems whose speed was measured in baud. And of course it’s always easier to start your own project than try to fork the existing server code. Not to mention that I haven’t bothered to look at the API yet, so my thinking may be way off-base.
Plus, I get the impression that the existing clients are half-baked, so the more the merrier. ;)
@nyan couple of hours of research in and I’m baffled. There is no apparent mechanism for discovery in ActivityPub.
It appears as if the protocol has simply ignored the literal first step of all social interaction: Observation.
Take lemmy as an example, if I visit beehaw.org directly I can click on “Communities” and get a list of them. Yet, there is no way I could identify to get this via ActivityPub. I’m completely and utterly baffled.
A quick skim of the protocol documents suggests you’re right: Activity Vocabulary doesn’t even mention the “discovery” use case, and no allowance seems to have been made for it. I can’t fathom what they were thinking.
The short-term workaround would be scraping the Web portal “Communities > Local” page of each instance periodically for a list (and cache the info centrally? Not sure.) Hopefully the raw HTML is parsable enough that you wouldn’t have to involve Selenium (which I’ve used in my day job—it’s awful) or its ilk.
The correct fix is, of course, to add the “discovery” case and messages supporting it to the protocol, either as an extension or as core for the next version. This might take years.
It might be worth asking browse.feddit.de how they get their information.
@JustusWingert
@technology @leopardboy @nyan
My guess is that this actually regarded as a feature in the wider #fediverse. At least in the Mastodon community it is. Toots not being searchable/indexed and you discover topics and people to follow by looking for hashtags and organically setup and control your own home feed. This, leading to much frustration for many who’ve moved over from Twitter.
Reasons being that it should not be easy for big tech (and others) to just slurp up and make peoples data open publicly for a wider audience through public search engines. A level of privacy and owning ones own data has been a priority over the inconvenience of not being as discoverable.
Of course the same thinking does not apply or map equally well to the #threadiverse as it does to Mastodon. I’m sure there are workarounds and a way to make this more seamless for users, but this is just getting started. It was never an issue when everything was on a small number of instances which everyone knew about.