cross-posted from: https://lemmy.world/post/4058689
Please be creative! Will there be more communities? Will we refer to each other based on the instances we belong to? Will there be beefs between instances? Will there be doomed romances of two peoples meeting from different places of the fediverse?
In such a widespread usage, there would probably not be “the fediverse” anymore, but a galaxy of clusters of interconnected instances. Spam would be a serious problem, so instances would switch to whitelisting instances they want to federate with instead of just occasionally “defederating” from them. It would not only happen because of spam, by the way, but also because of political/cultural/ideological divergences. Maybe even because of laws.
There would be a boom of innovations, made possible because of the data openly accessible and the fact that we would finally have a standard on which to build upon to create third party applications (which, from a developer perspective, was the promise of the web-2.0 and its APIs, but never truly materialized). You would see alternative frontends for everything, and applications that allow to get new insights or use your data in new and smart ways.
The big businesses would still be around, by the way. They would open their own instances, publish lot of ads and add cool features found nowhere else so that most people join their instances, which would quickly become the go to instances for everyone, dwarving all other instances. We would spend a lot of time evangelizing so that people join smaller instances instead, but our folks would answer that it’s less convenient, they would have less easy to use features and their account is already at BigCo anyway. Plus, to fight spam, terrorism, child pornography, nazis or whatever is the scarecrow then, they would severely limit the possibility for small instances to interop with them, adding arbitrary technical barriers that most implementers won’t succeed in hoping. But we won’t care that much, because we will have our own alternative networks with more content on them than ever.
I loved that, really got me thinking … especially the first paragraph I never thought in this direction before
That was called Usenet. Or in the SF novel A Fire Upon the Deep, it was called “the net of a million lies”.
Usenet has some interesting differences from what we’re using now:
Clients and servers:
- There is no default UI. Everyone is using client software of their choice.
- Your client software — not the server — is responsible for keeping track of which newsgroups you’re subscribed to, and which messages you’ve read in those groups.
- Posts are copied from server to server based on whether the receiving server is configured to accept posts in a particular newsgroup (or collection of newsgroups).
- Posts are not retained forever, except by archive services like Deja News (later Google Groups). Servers automatically expire posts to save disk space. Expiration can be tuned; a server might keep
sci.mathposts forever, but expirealt.binaries.*posts in a day because binaries (i.e. images and other non-text media) are large. - Server peering relationships are arranged by the server admins; there’s no default open federation.
- If a server goes offline for a while, it can reconnect to its peers and pick up posts that it missed. But because of expiration, it might have just missed some posts forever.
Newsgroups:
- Newsgroups look like containers-of-posts in the UI, but are actually implemented as topic tags in post headers.
- A newsgroup identity is global, unlike a Lemmy community. A newsgroup doesn’t have a home server.
- A single post can be in multiple newsgroups without duplication. This is the original meaning of “crossposting”. (Reposting the same message to different groups is “multiposting”.)
Newsgroup management:
- Newsgroups are arranged in hierarchies. Some of these are global (e.g.
sci.*oralt.*) while others are regional (likene.*for New England) and were originally unlikely to be carried outside of a geographical region. - New newsgroups in most global hierarchies are created through a formal discussion and voting process, intended to ensure that a new group is well-placed and has an audience.
- In the original moderation system, you post to moderated groups by emailing your post to the moderator, who posts it for you.
- This was largely replaced by “retromoderation” which is what we think of moderation today: anyone can post, but the moderator is allowed to cancel bad posts. The first modbot was invented for Usenet.
- Federated deletion is handled through “cancel messages”. Supposedly, only the author, moderator, or a server admin were supposed to be able to cancel a message. However, forged cancels were a common problem.
Threading:
- There is no post/comment distinction; they’re all just messages, commonly called posts. A post can be marked as a reply to another post; a new thread is started by a post that is not a reply to any other post.
- Clients typically present posts in a threaded tree structure, but the server doesn’t need to know about threads, just posts. Threading is a client-side feature; the server just gives your client the posts, and your client constructs the threaded view by following references in post headers.
- Because newsgroups and threading are independent, a single thread can actually span multiple newsgroups! A conversation can start in
comp.lang.lisp, then someone mentions their cat, then a reply crossposts torec.pets.cats, a later reply dropscomp.lang.lispand now it’s a cat thread. This is either pleasingly organic or really annoying depending on your attitude.
