🇳🇱
Mastodon Feature Suggestion
I had an idea for a way to improve and balance between the concerns of both moderation and free-speech. Its a bit radical but I think in theory it could work
Imagine mastodon, or something similar, but where a user can no add hashtags to either their posts or their profile, only the ordinary text of their choosing.
Then hashtags can be added or removed by other users via a consensus mechanism where people can vote for or promote hashtags or suggest their own (not unlike how stackoverflow does it).
Then provide all the tools for users to be able to filter out certain hashtags or promote them in their own feed.
This would allow the community to vote on what is hate speech or what is not, or what is conservative vs liberal, or any other relevant quality to a post or person. Then a user can say things like "Any post that is voted by 70% or more people to be hate speech i want filtered from my feed"..
If done well this could create a system where users are able to completely moderate a system for themselves with minimal effort.
My thoughts:
1. Requiring others to do the tagging will stifle new users. It creates a chicken/egg problem where I need my posts tagged to draw attention and gain followers, but I need people paying attention to get my posts tagged.
2. Trolls will troll. You'll get bullies who visit your profile and tag all your stuff as hate speech out of spite. The current flagging system needs a mod to actually suppress content, but it sounds like your proposal wouldn't.
@khird #1 is an issue regardless of tagging. If your a new user you have no followers, so you wont get many people who see what you post regardless. With that said, there is no reason to think you **need** tags to be seen. Many users would likely still view a feed that is inclusive of all posts and only **exclude** tags rather than use an exclusive feed and select **inclusive** tags. Ultimately, the power is with the user what they want to see. just as some users view an unfiltered federated fee now.
#2 Yes people will try. Any intelligently designed system will, however, prevent that. Stackoverflow uses the system I provided and because it is well designed prevents issue #2, not to mention prevents issue #1 with its designs as well.
Well right now #1 isn't an issue - at least not to nearly the same degree - because we can tag our own posts/profiles. People interested in calligraphy find you by searching related tags, but that wouldn't work if you had to wait for someone else to add those tags and you had no followers to do so.
StackOverflow's tags aren't used for moderation the same way as you're proposing. I can't really make it harder to find a question by maliciously tagging it, can I? (It's been a while)
@freemo but say I want to make a post about mercury language, and absolutely nobody around knows what it even is, I get no tag? Maybe other way around - give me the first vote on the tags to make the post visible, then let others reinstate/add/remove tags through voting.
Another thing is that this kind of a system will be easy to exploit, without a reputation system like in stackoverflow. Statistics don't work in small numbers, so a small group could actively silence a new/unpopular user, by exploiting filtering behavior. You could have some kind of a limit on votes cast before statistics applied, but then it kind of defeats the purpose of filtering, since it could be that "a big enough number of people who need to see the post(and vote on it) for statistics to kick in" is also "all people who will ever see the post", given that mastodon is mostly about quick continuous interactions, rather than long living publications.
Of course you can have moderators, now to moderate misbehaving taggers, but I'm not sure if that would make things easier or harder.
And finally if you're making a tag/category system pleaseee make it a tree, don't make it a list pleaseeee
@namark Makes sense you'd get a vote/tag on your own post to get the ball rolling.
I agree you'd need some sort of reputation system. I left that detail for another post not to complicate things. But I envision a trust system where instances can dictate trust level between other instances and its users. Then when a user trusts either an instance or another user then additional trust can be inferred from that. All instances with similar policies and moderation styles would have a high trust level between them and instances with dissimilar styles would have low trust levels. Each instance could have different ways insternally to rate the trust of its individual users in theory at that point,
@freemo I see, I guess the challenge there is to get people to understand and properly utilize the trust system. It sounds pretty complicated. stackoverflow's context is narrow enough for some clear objective metrics to exist, that can be diligently applied even by someone completely new, without necessarily understanding the bigger picture. You see a question you also have - vote up, answer that solves your problem - vote up, and so on.
Trust on the other hand could turn into something that people would hand out willy nilly just to be nice, treating it as likes, or something that the initial group of moderators would never give to anyone else(not necessarily maliciously, but maybe just due to lack of said objective metrics they would need to be fair).
I wonder if you boil this down enough if it'll froth up and turn into politics.
@namark
I think it could be made easy. As you downvote or uovote posts you see (not tags) that effects your trust in the individual. This could be used to imply the trust you have for the server. Blocking i dividuals would be seen as a downvote with higher weight.
Its really most of the same mechanisms we are already used to when you think about it
@freemo I often have hard time participating in voting systems that don't make it clear what the votes do. If the explanation is "vote if you trust and then some overly clever algorithm will moderate things for you" I'll likely refrain from voting (except maybe downvoting everything @design_RG posts :P).
A simple set of rules of what the trust rating does will help with that, but that goes against the idea of custom policies for specific instances. Maybe some kind of in-between - a simple set of rules but with some minor customizations available, that won't change the essence.
I'm interested what you can come up with, but expect me to runaway and hide and not help at all upon any discussion of the details.
I'm not sure it would be some "overly clever algorithm". I think the algorithm would be relatively straight forward. Everytime you vote up a users post that user gets a point. The users points determines the weight of their own votes WRT what you see. An instances votes would be the sum of the votes of its users. I dont really see a need for anythiong more complicatged than that for an algorithm.
Its no different than other highly successful systems in some ways like reddit or stackoverflow, just slightly tweaked.
But yea I do agree keep the customizations optional and provide a default.
@freemo I think this is overcomplicating things. I agree the right balance can be a difficult task sometimes but I think in general we should teach ourselves what is acceptable and what is not and only call for moderation in case it's really needed. Of course this is the ideal case and not every one is an adult or act accordingly, but is what we should seek.