So now that decentralized forums are more than an intellectual exercise..

robobobo

kiwifarms.net
Joined
Nov 21, 2014
It seems there's a strong need for us to be able to get a decentralized forum, and last I looked around there wasn't much for that on the whole fediverse thing, at least not at a state ready for production. I would be amply happy to host a node to keep the Farms up in a form without single points of failure that the troons can organize to attack.
 
A single SQL backend with multiple HTTP server frontends, each with multiple Tor, I2P, and clearweb addresses.

And a single page Tor/I2P site operating completely independent of those, on its own URL, which keeps a list of all the URLs for the main site.
 
Last edited:
The big problem that any distributed forum has to solve is moderation. In the forum model, there has to be privileged users who have the ability to remove or edit content. How you accomplish that without a centralized database of users is not obvious to me. It might be possible, but it would take some real effort to design that system, and the end result will almost certainly be much less user-friendly than a simple Web 1.0-style forum like this.
 
the end result will almost certainly be much less user-friendly than a simple Web 1.0-style forum like this.
I'd argue that a much less user-friendly solution would greatly reduce (if not outright eliminate) the need for moderation, because most people are too lazy and/or retarded to figure out how to use them.

Usenet newsgroups are the ultimate in decentralised discussion forums. It's old tech (dates back to the late '70s), it's mostly unmoderated and and it works.

No anime avatars, no updoots, no embedded videos... just text and hyperlinks to web resources.
 
Last edited:
The big problem that any distributed forum has to solve is moderation. In the forum model, there has to be privileged users who have the ability to remove or edit content. How you accomplish that without a centralized database of users is not obvious to me.
Yeah, we had another thread gaming this out a while back.
It seems to me that there are only three alternatives: have a centralized database, do some sort of Usenet-like setup where various database instances replicate content amongst themselves (as Usenet servers did nightly), or use a blockchain of some sort. Each has its own complications.

Most existing solutions "cheat" on one or another aspect of the forum experience. FMS on Freenet is sort of similar to what a distributed forum would have to look like (in that it is a layer over a backing store where individual posts are stored) but there is no moderation or user identity in the traditional sense - everyone is their own moderator over the firehose of content.
 
Most existing solutions "cheat" on one or another aspect of the forum experience. FMS on Freenet is sort of similar to what a distributed forum would have to look like (in that it is a layer over a backing store where individual posts are stored) but there is no moderation or user identity in the traditional sense - everyone is their own moderator over the firehose of content.
This isn't really true. The way FMS works, you "subscribe" to people, who in turn subscribe to others. So for example, I might "subscribe" to Null, who in turn "subscribes" to say the 50-odd mods, who in turn subscribe to the users.

There is moderation, but each user gets to pick and choose for themselves who gets to do it. I might follow Null and, say, 5-10 banned users whom I really like, minus two or three users I don't want to see.
 
The dead sperg is correct, though the audience for Usenet in this day and age is possibly even lower than staying on TOR. The farms leaving clearnet is going to be a massive blow to its accessibility. Right now it seems that DDOS mitigation providers flaking out on us are the biggest issue, unless other internet services like DNS act to shut us down, which would require literally unprecedented acts but which I sadly couldn't completely discount given how far other services have gone out of their way to screw us. So we need to find a way to host that foils DDOS attempts without relying on a third-party mitigation service.
 
Tor works for that; the question is what to do long-term, e.g. if Dear Leader throws in the towel or dies.

If you have a 100% uncensorable and robust forum, even if it's 1993 levels of usability, that would be really useful, like a bunker that never goes down.
 
Yeah, Usenet was still vaguely usable despite the percentage of retards because the Internet was like a twentieth the size it is now when Usenet was still heavily populated. It would be completely impossible to do today; the imbeciles from TicTok alone would fill each topic with thousands of spam messages for each one that's actually useful.
 
  • Feels
Reactions: Judge Dredd
Usenet is the original example of the problem I mentioned earlier. Without moderation it turned into a cesspool of spam, retards, and angry autists. It's one of the reasons web forums even exist in the first place.
True. But the upsides to Usenet (at least as it is now) is that the spambots don't bother with it anymore and even if one group does turn to shit, it's easy enough to create a new one and start again.

Given the relatively high barrier to entry and how it's likely to scare off 99% of normies, I think now is the time to seriously consider using it again. The lack of embedded video, inability to updoot and the fact that posting isn't instant acts as a pretty big disincentive for those raised by Twitter, Reddit and 4chan to come along and shit up the place.

How long it'll be until autists start coming along and shitting it up is anyone's guess, but at least the posts will be archived one at least one server somewhere for all eternity.

BTW there's a group called free.homesteading.kiwis that was started a while back. The events of the last couple of days have led to it becoming (slightly) active again.
 
The big problem that any distributed forum has to solve is moderation. In the forum model, there has to be privileged users who have the ability to remove or edit content. How you accomplish that without a centralized database of users is not obvious to me. It might be possible, but it would take some real effort to design that system, and the end result will almost certainly be much less user-friendly than a simple Web 1.0-style forum like this.
It's the big wrench in the way of true information exchange. If we got rid of moderation on a service and switched to a self-filtering style we could avoid this problem entirely. Issue is people want some information censored, for good and bad reasons (CP vs hate speech). But you can't have it both ways. A decentralized service must have no centers and no way of preventing the transmission of information.
 
If we got rid of moderation on a service and switched to a self-filtering style we could avoid this problem entirely.
There's nothing stopping us from doing that right now on Xenforo to give it a test drive. We could defrock all our mods, enable guest posting, encourage use of the "Ignore User" feature, and maybe even make up an "Ignore Post" feature to supplement it. People could make up a new version of BlockTogether to share their blocklists. Do you think it would be an improvement?
 
It's the big wrench in the way of true information exchange. If we got rid of moderation on a service and switched to a self-filtering style we could avoid this problem entirely. Issue is people want some information censored, for good and bad reasons (CP vs hate speech). But you can't have it both ways. A decentralized service must have no centers and no way of preventing the transmission of information.
You can't have a really functional community without a leader. The leader can delegate his absolute authority to sub-leaders to effect his vision. But if the leader-principle is not enforced, there is no real community, just a bunch of people shitposting. The Fediverse satisfies that requirement.
 
It's the big wrench in the way of true information exchange. If we got rid of moderation on a service and switched to a self-filtering style we could avoid this problem entirely. Issue is people want some information censored, for good and bad reasons (CP vs hate speech). But you can't have it both ways. A decentralized service must have no centers and no way of preventing the transmission of information.
A self-filtering style has the issue of being opt-in. If you generate a million spam accounts, requiring everyone to click a million times to hide them isn't sustainable.

That's why you need community moderation:
  1. I have a list of users whom I trust
  2. They have a list of users whom they, in turn, trust
  3. And so on, and so forth
  4. If I indirectly trust a user, I see his posts
Then, you can have a two-tier system for new users, something like this:
  1. New users start out with almost, but not quite, 0 trust
  2. Basically a single negrate is enough to get them hidden by default
  3. New users (e.g. users with very low trust) can post, but can't make new threads and won't bump threads
  4. If someone replies to them or posrates them, they get trust
Also, as a slight optimization, there's two separate levels of trust:
  1. Message trust (MT) = "This user isn't a spammer"
  2. Trust list trust (TLT) = "This user doesn't assign trust to spammers"
Frankly, I guess I should stop talking and >DO SOMETHING.
 
As I mentioned on poa.st, the poa.st and other twitter-style threads could be displayed instead as a Forum thread. I mean, really, it's just a top level reply and then everyone's replying to that top level reply, that data model works both for twitter threads and forum threads. As a bonus, subforums and even forums themselves could just be another "post" in the "thread."

I believe there might be challenges in data loading, but who knows?

I don't know how the Fediverse handles shit when a node dies. Do all of their posts vanish, but the replies survive? Some redundancy could be engineered, perhaps, if people are engineering a Fediverse-Forum-Threadview.
 
Just to sperg a bit: to my mind, at least half the issues revolve around the fact that the physical layer of the network (cabling and associated switchgear) are in the hands of a comparatively small number of highly influential individuals. As those people are subject to the legal and financial systems of society, a point of attack opens up for activists, politicians and other assorted despicable human scum to force corruption into the network.

Decentralizing ownership of that physical layer would remove the vulnerability. Superficially, mesh networking deployed on wide scale appears that it could offer that. I'm not acquainted with the technology, however, and remember there being known flaws in routing protocols last time I looked into it. Perhaps some of you are better able to comment.

Anyway, back to sleep. Thanks for tolerating my rambling.
 
Back