Discussion about paper "Towards a peer-to-peer imageboard"

This. Centralization is the way to go sadly. Got no other choice.

enjoy your botnet.

Those are all federation (the worst parts of both centralization and decentralization), not decentralization. Decentralized data distribution, with an authoritative, centralized list of content is the future. The torrent site model is the correct one

There's two big factors being overlooked

You need a healthy community for any decentralized network to function

and

Self-censorship is not a solution; part of what makes imageboards great is that you're forced to wade through all the crap to find the gems.

How can you maintain a community if not everyone is seeing the same thing? How can you dissuade normalfags from wrecking your forum if you can't spam stuff that scares them aware? If the culture is to be preserved, the only way to do that is by ensuring visitors have to see everything.
Desensitize or leave

This leaves a couple of options. The first is "community policing", but since the goal is to have no accounts and no way to even link two posts together, there's no real way to prevent tampering. Voting systems have been shown to not work well to invasions. A technical solution, where threads or even boards have their ratios curved by percentage (so if every post has massive spam flags then the number of reports to be considered spam rises with the average) might alleviate some issues, but it's not perfect.
It would be feasible to allow spending a cryptocurrency to perform mod actions. Perhaps putting votes behind a curreny, so that you feel compelled to use them sparingly (since you still need multiple users voting to actually remove anything). But then you have the issue of how these points are generated and distributed. It's not impossible for a group of invaders to just sit on stacks of this "modcoin" and then coordinate elsewhere to target specific threads or opinions they don't like.
The big issue with any democratic system on an anonymous platform is that you can't verify anyone nor their intentions and you open yourself up both to actions in bad faith and the tyranny of the majority.

It's worth remembering that sites like reddit, for all their other faults, have used voting systems as a method to skew debate and cultivate opinions. It's been proven that a handful of votes at the right time can skew an entire post's performance as well as which comments get seen on that site, and once the "correct" opinions start to be everywhere it will eventually create a cycle where the moderators and everyone there demand stuff that offends them be removed.

In short, in a system like this you need to protect the right of free speech, even for those who use it to demand it be restricted, without compromising on it an inch.
I'd be interested to see what boards on any site would look like with zero moderation. Spam is a real concern, but CP seems to be posted mostly by people who want to prove the site is full of it by planting the evidence. And I don't want terrorists recruiting anywhere. But no controls seem preferable to me than letting people voluntarily avoid opinions they don't want to see. Especially if the tools allow you to pre-emptively ignore or filter comments based on keywords, etc.

Speaking of culture, how can you have event days or wordfilters? They're not necessary and are open to abuse, but they're also fun and part of the culture. A solution would ideally allow for users to collaborate to make these happen.
I don't want to rule out voting systems yet, but it's important they be structured to mediate the worst tendencies of crowds.

Attached: 1453972329433.jpg (380x385, 8.21K)

cross linking my reply to the other thread


what do you think about my flag host idea above for illegal content? then have effectively "moderators" that go through all the flags and purposefully host bad flagged threads/replies. since it's decentralized, it would work through redundancy.

what is this, SA?


using my method in the reply above, you could have wordfilters tied to the thread itself rather than the board. so when you post a thread, you can give it word filters while creating it. if enough people use word filters, you can have a word filter board-wide effectively through the normal spread of board culture.

>jew(s) attempt to look covert while shilling for (((centralization)))
Read , also considering further alternatives, and then go away.

My suggestion is nothing like SA. Moderators there are allowed to be Nazis and ban people. They come crawling back and pay for entry.
However, it would put it in line with something like Slashdot, which is really not much better. Under ideal circumstances, my intentions would be to allow users to make sparing moderation votes. But I think it needs to be designed for a worst-case, and by all evidence from sites that allow for this sort of community vote moderation it just is too temperamental and doesn't scale with the size of the site and its community. But I think as an experiment it might be worth consideration, and if the variables are all right it might be functional.

But onto the meat of the issue:

Funny, because after my previous post I was mulling over the idea of opt-out re-seeding. I reached the conclusion that any optional re-seeding would need to be opt-in, to prevent lurkers from allowing spam.
It's definitely workable but I think it has some issues.

Who has moderation power and when do they get elevated privileges?
Even if you assume that a post must be reported first before it can be reviewed (and requires multiple people who did not report it to vote on a final action) it would be potentially possible (depending on system design) to let people game this system.
In regards to moderation action, assume that a group of butthurt SJWs are hanging out in a Discord channel, coordinating to remove posts that hurt their feelings. Does a given moderation system prevent them from either deleting posts they dislike or shuffling them off the board by allowing spam and giving posts they do like a free pass?
My modcoin suggestion was deliberately suggested to make moderating costly, so a few hotpocketeers can't encamp the "review reports" job and form their own tribunal while everyone else is clueless that it's been monopolize and the board is being manipulated.

Keeping the board synchronized is important, and people looking at threads should be looking at the same thing. How do we do this without having swiss cheese threads or inline spam?

Everything still needs a unique post ID and a thread + board to attach. I'm not sure how workable making them independent and attached by context is, but it's a start. It might be functional. But it still needs to tie into those two previous systems. But re-seed disparities might make it difficult or slow.

In light of these considerations, I have a new proposal. While Federations have issues Hotwheels mentioned, I think there may be a middleground between having nodes and having peers that might mediate the downsides of both without killing their benefits entirely.

We could have nodes that act similar to trackers in a swarm. You can connect, sync with the board, and download threads.
Content itself would be P2P, alleviating bandwidth costs for the node.
Nodes can host alternative CSS and MOTD.
The admin of the node can have moderation power within their own node.


Nodes can choose to delete a post, effectively marking it as not for re-seeding.
Visitors to the node and any nodes syncing with that node would then stop receiving copies of the post.
Other nodes do not have to delete the same thread once it has been synced to their node and their peers are sharing it.
All moderation actions would be publishable to a blockchain-based moderation log.
Media gets hashed so it's not hosted.
Makes the record of moderation public.
Not optional since to delete the content you need to write to the log, which must be public so clients and other nodes can know what you're choosing to not re-share.

There's nothing stopping anyone from making a hotpocket mirror on their own node, but if other nodes choose not to re-seed posts from those boards then they'll be off in their own corner, unable to send posts towards the central node cluster.
Minimizes risk of monopolies forming and the need for an exodus.
As long as the most popular boards aren't the most moderation-heavy and they don't collaborate to form a monopoly.

I recognize this might be an issue, especially given the way normalfags beg for censorship, so there are concerns. It's important to be able to have common nodes blacklist spam or cucked nodes so they do not spread without ensuring the most heavily moderated nodes have the power to blacklist totally unmoderated ones.

Nodes could be run as Tor hidden services as well.
If the community works together it could relegate Tor users to second-class citizens but in theory should give them greater freedom to have their own network and full posting privileges, even in the worst-case scenario. Still better than the current situation.

I'd really like to hear your thoughts.

Attached: c3619005d50868facf7771650a8fb94d05a109252a72d6f5e041039dfaa144b8.jpg (500x529, 44.99K)

halfchan to my knowledge doesn't allow users to view archived posts. 4plebs and the like sites took over that arena. IBs will always be forever searchable. Anybody can run a scraper

I am interested in this because right now this website is effectively banned in Australia.

Isn't this then just voting?
Doesn't this just give moderation power to state actors or other well organised SJW groups?
I could be reading a thread and be reading messages that respond to a message I have yet to sync? In isolated nodes, or node clusters that are just starting to coalesce into their own little echo chamber, this would be a bit fragile (but maybe sync issues for these small clusters is preferable).

CP can be easily hidden in normal images