13
votes
Users thoughts on groups?
As of right now of course, users can not create their own groups, however this may change in the future as is stated here.
What do the users think about this? Although tildes is similar to reddit, it's obviously not got the same goals in mind, and I think it makes sense to keep groups up to the admins.
Wow. Are you storing all those links somewhere just for situations like this? I knew we'd discussed this topic a few times before, but I didn't realise we'd discussed it so many times...
I've done the same thing myself for some topics.
Yeah. I've got a couple of these rolling lists going for "Why can't we have fluff?" and "What should we call Tildes users?" And now I have one for "Why can't we create groups?" :)
I know, right? And it's always brand-new people who ask - I assume most of them are coming from Reddit, where any redditor with a handful of karma can create as many groups as they want. They don't seem to realise that they're not in Kansas any more. This site - surprise, surprise - isn't the same as other sites. That's why we're here. If I wanted a copy of Reddit, I'd just stay on Reddit.
I think it's a good idea to create groups only when there's a demonstrable demand for them, rather than letting everyone create groups on a whim. For one thing, it'll prevent duplication. For another thing, it'll aid in discovery of sub-groups. And it keeps the site cohesive, rather than fragmented.
It's about as 'official' as anything around here - meaning a general idea/concept we'll all discuss to death before implementing anything. The goal is to put humans in charge of as much as we can. The mod tools aren't meant to replace humans, just amplify the work those humans can get done with the time they donate to tildes.
I think the 'topic log' over there --> serves as a pretty good start on the idea of public mod logs. I kinda like attaching it to the topic like that instead of having it be its own messy page that people won't easily be able to find.
Edit: Just realized this thread hasn't had any changes so the log doesn't display (it's empty). If OP removes-readds a tag to this topic, it'll show up.
That is an excellent find. I've saved that as a favourite, and I'm going to trot it out every time someone asks how Tildes is going to work.
Thank you!
I don't think it'll be fully algorithm-driven. I think it'll always ultimately come down to a decision by a group of people - but that decision will be informed by statistics showing how much demand there would be for a new group, and be the result of a collective discussion, rather than just one person deciding to create their own little fiefdom.
Perhaps public votes could be held and inactive groups could be deleted after some time has passed and warnings have been issued.
One of the bigger problems with Imzy was the rapid proliferation of a huge number of groups with almost no content.
"books" wasn't enough, "books for YA about vampires" wasn't enough, there had to be groups about "this one character in this one YA series of books about vampires", and then "this other character in this other YA series about werewolves".
I have seen the future and it is bleak.
I think we should treat it like 4chan for now. If a topic becomes big enough, give it its own group.
Oh, sorry. Just joined today so I didn't know.
Thank you.
Unless the poster searches if someone else has posted on a topic (and they happen to use the same wording/keywords) wouldn't reposts still occur?
I think reposts are fine to people who have not seen a particular post but those who have seen it would probably get sick of seeing it constantly. Or someone looking to get votes will repost popular content with different titles (which I believe Tildes has measures to counter this behavior).
Looks like some of the stuff I mentioned was touched upon in this post about necroposting.
Reposts are inevitable. The good news is, the site can just pop a 'repost' tag on them automatically in most circumstances. People can then filter them out or not as is their preference. We could even link the older threads in the sidebar of the newer topic once the site becomes smart enough to detect the reposts.
Is it September already?
Imho, the amount of groups should be kept to a reasonable minimum for now. When groups start to become too full (i.e., topics are opened so numerously they "flood each other away"), subgroups would be needed.
I think the mechanics of newsgroup creation could be altered to suite Tildes: Have a group ~tildes.groups(.proposals), where proposals can be made and discussed. I am not sure, though, how to decide what proposed group should be created.
Also, once subgroups are in place (with whatever creation mechanics), I'd like to have the option to see topics of all subgroups when browsing a group. E.g., on visiting ~music.metal I would like being able to tick a checkbox labelled "show topics in subgroups". Another way of implementing this could be being able to browse multiple groups at once (e.g. by visiting tildes.net/~music.metal+~music.metal.instrumental).
Probably the way proposals would work is a user would submit a proposal, if it got enough votes and the admins liked it then it would go to a poll. Yes, you should be able to subscribe and unsubscribe from subgroups.
I think groups have, logically, two engines for creation.
Subgroups are the easy one. You have ~games, and one day you notice that things tagged with [tabletop] are coming to dominate the content of ~games. This is also annoying to some people, who probably start a thread about it (we've all seen those bitchy threads on reddit over the years). This is the point on reddit where mods usually put a topic filter in automod that auto-removes that content. That of course pisses off even more people and thus begins the content wars in that subreddit. This is no solution, it's a mess.
Here we'd simply create ~games.tabletop. It'll start with all of that content, and the rep of the users posting that content on the day it's made. Poof, overnight you make a busy group rather than a sparse empty subreddit. Now people can unsubscribe from the tabletop content if they aren't interested. Meanwhile, when epic threads happen in ~tabletop, their mods can push them up into ~games (with approval of some kind from the ~games mods). That way instead of being in ~games and seeing 100 tabletop threads a week, now you'll just see the one good tabletop thread that week that appeals to both groups. The topic is no longer spammy and fighting with other content.
Take this to an extreme and ~games has more than 1000 sub-groups. It's a gateway now, less of a community of its own. You probably don't submit much to ~games directly anymore since it's so busy - instead you submit to the right sub-groups (and we'll need a good community-finder to help with this). It's now the kind of place you host all the game-related AMAs, and push all the best threads from the sub-groups up into so everyone can see them. It's rather the opposite of a busy low-effort default now. Instead it's a curated high quality space that's stewarding the entire games community on tildes.
The second method for making groups is the top-level hierarchy stuff (basically, what we've got right now). That's probably going to come down to making posts in ~tildes and all of us talking about it. If there's enough interest we can give it a try. The only real 'bar' here is making sure it's a suitably broad topic for a hierarchy, and making sure it's got a level of interest that'll help it succeed.
There could be a third method, or maybe this is a mechanism to help with the first two. Ever make a guild in an online MMO? You need to get a certain level of buy-in from a certain number of people (signing the guild 'charter' like a petition), and it costs you in-game resources to do it (here that might be karma or rep or gold-like tokens, or nothing at all). Once you hit the threshold the guild gets made. We could probably adapt this concept into a group-starting process.
We can move groups around and re-order the hierarchy, too. It isn't set in stone. Subgroups can become root groups and vice versa. It needs to be flexible. I usually use the 'bonsai tree' analogy. Care and pruning will be required by the admins to keep it all green and healthy.
I had a crazy thought which I'm sure has issues, but I'll share it just for the conversation and thought experiment fun: What if there were no fixed groups?
What if you just posted a topic to Tildes, gave it some tags, and people used only the tags to organize the posts. This could be entirely organically created by the users who post topics.
The sidebar could contain the top 25 most used topics tags.
I imagine that this has been tried before on other platforms, why does it fail?
Actually, I think we can solve those problems. We've talked about this in depth and the solution is part of the design here.
Tildes will eventually import the tags from the content at the other end of links when people hit our 'suggest title' button. That'll save a lot of time since users only need proofread the tags it brings up, and damn near everything, everywhere is now tagged on every site. Also, as users earn trust they'll be able to edit tags and links on topics, so everyone will be able to help out if the OP misses something.
The groups are the binding element. If there's a ~bitcoin aka ~economics.cryptocurrency.bitcoin, anything submitted there is tagged with economics and cryptocurrency and bitcoin as a function of where it is submitted in the hierarchy. That's where the real uniformity comes in, based on the choice of where the content is submitted.
Any tildes group is basically ~that.groups.core.tags with a short ~nickname, that's why there are two names for every group. It's tags all the way down, a 'group' is just where we pin a community framework to a collection of tags.
That’s really interesting. Would that mean that tagging with economics and cryptocurrency and bitcoin, and it and crypto and blockchain would be like cross posting to ~economics.cryptocurrency.bitcoin and ~it.crypto.blockchain?
That would be such a powerful way of allowing natural hierarchies and relationships to develop, especially if users were gently guided towards how to add preferred tags and tag combinations to create informal subgroups.
I notice that topic urls don’t include the group, so topics could be retagged as groups and tags grow, split and fall away.
It could go that way, even to the point of having two separate comment sections on the same topic. It's also going to depend on what's easiest to code for and support, as well as the natural evolution of the forum. I don't think we're going to know the shape of the answer until after this place becomes a lot more busy and has a lot more groups and subgroups, with some bubble-up action going on. Once we get a chance to observe how that plays out we'll be in a better position to handle multi-homed topics and cross-hierarchy submissions.
100% agree, it would end up impossible to find content you want to see and karma (or... vote?) whores would just add literally every possible tag.
It becomes chaotic with high numbers.
I more polished solution would be to have the system create (or populate a "candidate" list) groups when a tag is used often enough to create topics and those topics get activity.
That basically indicates that it's an interesting topic to lots of people so maybe it could have its own group.
I think the "group candidate" being automatically generated based on tags would be pretty interesting. It would not only incentivize people to use tags, but also for people with similar "niche" interests to tag their posts using the same tags.
This might create a culture of tagging into the site, but I wonder whether or not the "reward" of a new group creation is enough incentive. Maybe the top 5 users who used the tags become the moderators of the new group?
I strongly believe that moderation should not being given to a restricted number of people.
The main reason I got interested in tildes is for the trust system and how everything should revolve around it.
Reddit has proven that restricting access to moderation is a death sentence. Even in the few rare cases where it works out, those mods eventually burn out. Reddit doesn't handle turnover and burnout at all.
I wasn't suggesting that the 5 first were to be the only moderators ever; perhaps moderation tools could unlock as they naturally do for other groups (i.e. after a certain number of posts) and be given to everyone who has the threshold when the group candidate is accepted.
An incentive could even take the form of the kinds of "flair" that reddit has - something that doesn't really affect the experience of others but give users enough street cred that they are incentivized to tag.
It's just a matter of time before we have to open the group making process, because the idea is to continually expand the userbase and the biggest groups will end up just as bad as reddit.
Probably the biggest improvement we can implement would be to make bans democratic. Instead of mods deciding how to interpret their own rules, the subscribers should vote whether to ban a given individual.
Bingo bango bongo. 100% agree with you. Growth is only moderately important right now to keep the site populated and get it sustainable (activity-wise and ideally financially as well) but after that is achieved it's going on the back-burner until the systems designed to deal with issues that come with scale (e.g. trust/rep) are implemented. And full site democratization sounds like a horrible idea to me too, although trusted users having the ability to vote on the moderation actions in the groups where they are trusted is another story.
Obviously, the mods first decide that they would like to ban someone.
But then, instead of them making an opaque decision, it has to go to active subscribers of that sub for a 2/3 majority vote.
So, in a group with 1,000 active subscribers, we'd have to wait until 666 of them vote to ban someone. Meanwhile, that troublemaker has freedom to continue making trouble for the days or weeks it would take to get hundreds of people to vote.
That doesn't seem practical to me.
I think it's far more likely we'll employ a system of muting and strikes. If a user does something nasty they'll get silenced for a time period, and have a strike placed on their account. Enough strikes triggers a review of that user, or possibly even an automated ban. It should take more than a single incident for that to happen, everyone has a bad day once in a while. If it was a one-strike-gone policy most of us would end up banned sooner or later, myself included. :P
I kinda like the idea of placing a 'tombstone' onto a banned account too, so people can see why it happened. Not sure if we'll do that or not.
Meta is death, and this idea is sure to generate huge amounts of meta discussion, and soem of it will be toxic and awful.
Have a look at Wikipedia ANI for examples of this.
Meta is only death if everyone has to reach a universal consensus before anything gets done. Deimos is the final arbiter of what goes on around here, not a wikipedia page longer than war and peace where no one can agree on what to do.
We want those meta discussions to help shake out the best ideas. There will also come a time for any idea where it's time to stop talking and get things done. That's going to be up to Deimos and the group of developers working on Tildes on gitlab. They already have their own separate discussions there.
We'll try to inspire, they'll try to build. It's the basic convocation model for development.
agreed
This sounds like something which should be given a LOT of thought. This sort of thing could very quickly turn in to brigading and witch hunting.
And by that, we could go to literal witch hunts. Without going on the extreme, this is the kind of thoughts that gave rise to actual witch hunts and many public executions.
Mob mentality is a real problem, digital and real. We should be wary of any solution that just trusts all users regardless of knowledge about the situation and requires and amount of them to vote before taking action.
Yeah literally anything that just allows for direct democracy is going to immediately turn into a witch hunt.
Or worse, campaigning and other political dynamics. That's madness.
Tildes isn't Reddit. It will operate differently. One significant difference that's planned is that moderators of a group/sub-group will rise up from within the group, rather than being creators of the group and the creator's hand-picked helpers. As you post and comment in a group, you will gain "reputation"/"trust" within that group. As your reputation increases, you'll get increasing moderation abilities. It'll start with just editing topic tags, progress through moving posts to more appropriate groups, then gradually increase to removing comments, and eventually banning people. To move up the scale, you'll have to build up a history of good moderation at the lower levels. Higher-level mods will continually monitor lower-level mods for positive or negative contributions.
This has two consequences:
A group/sub-group will have many many moderators. I've seen one of Deimos' helpers say that a group with a million subscribers would have tens of thousands of moderators.
The only people who have the ability to ban users will be people who've worked hard for months (years?) at being a low-level moderator, and have shown they can be trusted with the power.
This isn't Reddit.
Literal LOL!
I totally agree. :)
That would get out of hand with opinion based bans I think.
A lot of what Tilde's basic philosophies rely on is unsustainable on the mid or long term (or even short) because it relies on a benevolent dictator(s) to make groups, ban people, moderate discussion, etc. This is beyond asinine and there's a reason why it's not a design that is used in politics or anything larger than small, niche communities online. It's proven to be impractical and to implode on itself given enough time.
It has its nice self-sustaining rhetoric that is explained at length in the Docs section but in the end it's nothing more than a big ego-trip that won't work as anything other than an experiment without some deep self-reevaluation and metamorphosis. We've already seen our first ban(s) due to people abusing "downvote" tags, which were even removed entirely as far as I can tell. The premises this website is built upon just don't correspond to how people interact (and no, they won't stop interacting the way that's natural regardless of how much you force a design on them) - plus it forgets tons of advances on Internet community-building and backtracks on them (namely, community self-moderation and large amounts of freedom).
I'm here to watch and participate on the experiment, with an open heart and mind but I can't see it working at all the way things are laid out right now (but I'd be happy to be proven wrong even though all I see around here just confirms it further for me). So to put it clearly: I'm definitely against having a dictator who decides what groups are allowed or not, regardless of how well-intentioned and competent.
I think the solution is to have the conversation. Later, once it's died down, go back and read it all. The answers are usually self-evident and staring us in the face once the dust settles. That's where all of Tildes' ideas are coming from, and not just on this site. This particular conversation is over eight years long now, started by Blackstar9000 in /r/theoryofreddit.
We aren't getting out of this unscathed, though. No theory ever survives contact with reality unchanged. An 'idea' is a far cry from a functional system. The only way to get to that functional system is with the rigor of experimentation and observation. Mistakes will be made, and corrected, and learned from, then made again. That's the only way anything ever gets done.
Name any other community-driven website that even tries to evolve. All of them I've seen only add features users couldn't care less about to make money, not to improve the community's ability to self-organize and govern. It may be just that easy to make progress - ditch the profit motive.
I can't imagine how you would come to this conclusion unless you've only used sites that rely on minimal moderation like Reddit.
This is more or less how every traditional forum has worked for a long ass time time now, and if it functions well or not in the long term is almost entirely reliant on choosing the right "benevolent dictators." What you're saying is so bizarre to me because this model has been tried, tested, and been the path that most high quality sites have taken for a long time now. Nevermind the issues of applying broad political ideals to all spaces regardless of goals, scale, or expected rights.
Really, the things you bring up are pretty irrelevant to what ~ is planning on bringing to the table, which has a lot more to do with the process of deciding who those "dictators" are in order to fix the biggest problem with this model: "dictators" have to treat working on their site as a full time job and manually picking out suitable "dictators" just doesn't happen fast enough at times. If the site gets overrun by normal users, which does happen sometimes, the problem is with how we decide a user understands the goals of a site and is noticed for gaining "dictatorship," which is a problem worth trying to solve. If you have a problem with this, it makes plenty of sense to go after it since experimenting and trying to fix it in different ways seems to be a pretty open thing. If your problem is with the "dictators," why are you specifically here? Why not just some vbulletin forum that's been running forever?
Although, if what you see around here already confirms it for you, maybe the issue here isn't that you think it won't work for its own goals. Maybe the site working just fine for what it wants to be just isn't desirable for you?
I totally absolutely 100% agree: expecting the current model of Tildes to scale up as the site grows is utter idiocy. It can't operate under that scenario, with hand-picked benevolent dictators.
That's why I'm glad that Deimos and his team have something different in mind than that.
This system is so ripe for circle-jerking people into full moderator power and overall abuse that it's not even funny - especially when you consider each invite just means a free +1 potential circle-jerking account. I shouldn't be surprised, considering how much of Tildes' aesthetics are absolutely naive.
The full, public moderation log is the one saving grace here but that would actually work a billion times better if it had an actual sane system to work in tandem with instead of a backwards joke.
Tildes is going to have to rethink everything moving forward. I doubt it will keep the impetus, especially if the dev(s) go through all the work of implementing their systems only to then find out they are 100% obviously exploitable in practice. I'm pretty sure he stated in the blog he can't work on Tildes for very long without proper funding so I'm guessing the website is pretty much dead if he follows through with that senseless plan.
You're coming out pretty harsh on a website that is still in alpha, with major features such as trust, subgroups, and tagging not being completely implemented yet. I mean, yes, many of the things Tildes hopes to be are somewhat of a long shot. But I don't think fair to dismiss the whole thing as a naive experiment considering we haven't really started to experiment yet.
You're absolutely right. It's in alpha. Not beta, not a finished product, alpha. There's obviously going to be problems, and we're going to have to address those as they become apparent.
What would a better system look like? The folks that brainstormed this put a lot of thought into why this is a better approach than what forums and social sites have traditionally done, so I'm curious how you'd go about solving those same problems.
Not at all: I'm discussing their plans for the future as laid out in the Docs section and linked comments, I'm not talking at about the current state of the website.
Are you aware that the current invite-only model for Tildes is temporary? The long-term plan is for Tildes to be open for anyone to sign up.
I would also point out that a similar moderation system is already in practice over at StackOverflow and StackExchange.
However, if you don't like what you see on Tildes, that's perfectly fine. It's not for everyone. There are plenty of other forums out there for people who want something different.
It may be a good idea in moderation. One thing I've been thinking is that perhaps religions should be in a separate group from humanities, just because religion could have so many sub-groups in addition to the other things covered under humanities (such as philosophy) that it could quickly become overly cumbersome if large categories like religion and philosophy are grouped together. This doesn't necessarily require user-made groups though, so I'm not sure. Anyway, the Buddhist subs on Reddit are the ones I'm most heavily involved in, so when I joined here it was the first thing I noticed.
That's where subgroups will come in.
It will be. There will be a ~humanities.theology and a ~humanities.philosophy and a ~humanities.history and a ~humanities.linguistics - and eventually there'll be a ~humanities.theology.religion.buddhism. Sub-groups don't exist yet on Tildes, but they are absolutely part of the medium-term plan here.
I've even seen people explaining that a cumbersome sub-group name like ~humanities.theology.religion.buddhism will also be accessible via a simpler alias like ~buddhism.
Stay tuned!
Well there is also the problem that Buddhism is a non-theistic religion. So theology may be the wrong subgroup name. Maybe just straight up make it religion?
Theology is also the study of religions, not only the study of gods.
Anyway, it's all academic now (pardon the pun). It'll be months before we start splitting ~humanities up into sub-groups, and splitting off a religion/faith group or sub-group. We can make the decision then, based on how people are posting in ~humanities, and what the various religious-related topics look like, and so on.
The main point I wanted to convey is that you will eventually get a general group about religion, and a specific group about Buddhism (regardless of where they sit in the hierarchy).
Faith doesn't quite work because Buddhism isn't a faith based religion either. I think the only one that works for everyone is spirituality
I'm starting to think I like the idea of tags better than groups. Especially when there are so few groups. Need a way to subscribe to multiple tags. And worry about people gaming tags later.
I've read about the dream of an quasi-organic growth of subgroup hierarchies. The germ of it maybe already exists to a limited extent, with more user freedom currently, in the form of tags. Sort of. Tags tend to be more specific, and often rely on tagging conventions.
Think of twitter (but don't think of it too hard) and how the terrible (and brutally ugly) hack of so-called hashtags freed users to create their own databases of categorization, grouping, and (in some cases) lateral communication.
Vive la tags! (Also known, in the good old days, as keywords.)