Who will moderate and what tools will they have access to? - Meta

TopAnswers Meta

Top Answer

Monica

I don't know what *will* be done, but I'd like to offer some proposals (recognizing that this question is pretty broad).

**Who will moderate?** That should be up to each community. From a platform perspective, we shouldn't care whether they choose founders, hold elections for permanent positions, draw straws every six months, or make "moderator" just another privilege that is gained through site activity. The platform's job is to provide a means of designating someone a moderator.

**What tools do we need?** This will be constantly evolving. We should start with the minimum that gets the job done. I think that means:

- A way for users to flag content for moderator attention.

- A way for somebody to hide/delete content. Initially, this probably means a manual process. Later it could be community-driven (enough flags, for example).

On day one I think that's enough. Soon after I think we will also need:

- A way to lock content -- prevent edits to a post, prevent new messages in a chat room, maybe other things. When faced with vandalism like a spew of rude comments/chat messages on a post, there should be an option that's milder than deleting the post to which the activity is attached.

- A way to address users who are behaving in ways that are disrupting the community. This could mean suspensions, but it could also mean revoking specific privileges, depending on what the privilege system ends up looking like. Or it could mean imposing rate limits.

I propose that moderators be the ones who wield these tools and make these decisions. Things baked into the system tend to be good fits for some communities and bad fits for others. Initially, give the power to the human moderators, and then see what patterns emerge.

At some point we're going to be successful enough to be facing spambots coming through Tor, and that'll need better tools. I don't know how early we need to consider that; my gut feeling is that our earliest concerns are about disruptive humans who can prevent communities from taking root by setting a bad mood, rather than insurance spammers that everybody knows to ignore.

Answer #2

Jack Douglas

We don't have a complete answer to this question yet. What we do have is our first 'moderation'-like feature, just released.

Our hope is that this will be a model of other features as they are slowly rolled out in response to need. However it is also a trial of sorts, and subject to significant changes based on feedback here, after people have tried it out for a while. The idea is to break the overall 'moderator' role down, into finer-grained areas of expertise and trust.

Before getting into the feature itself, one thing needs to be said very clearly: ***all the following actions will be public***. Your name and identicon will be attached to posts you flag, and visible to others, including the OP.

Here's how the new feature works right now:

* Every post[^1] initially carries a 'flag' button next to the 'subscribe button':
![Screenshot 2019-12-10 at 20.33.29.png](/image?hash=21a5ee117c4df5dd87e9937840564784f9bee54ddabf600e00aca96c91ba7e0a)

Each community can decide it's own 'flagging' guidelines, but the expection is that it would be for content you'd want deleted: spam, off-topic questions or junk for example.

flagging a post has two effects:
1. The post is immediately hidden from all unregistered users
2. A notification is sent to members of the post cleanup 'crew'

* Unlike regular registered users, members of the cleanup crew flag in a different way:
* As soon as a 'crew' member flags, the post is hidden from all non-crew members
* Crew members also have the option to 'counterflag' — this overrides regular flags user and makes the post visible again
* If multiple crew members flag and counterflag, the overall action is based on the majority.
* if a clear majority forms, outstanding notifications are cleared.

In normal use I expect something like the following to happen:

1. Several users flag a post drawing it to the attention of the 'crew'. Comments are made in the question chat room if necessary.
2. The first two 'crew' members to visit the post confirm the flags and the post is then effectively deleted (invisible to everyone except 'crew' and the OP).

Finally, I anticipate that each community will decide on the rules for an automatic background job. The job will flag questions that meet certain criteria (e.g. no answers or votes after 4 weeks).

[^1]: except your own — and right now it is only working on questions (answer will follow if the feature is well-received, otherwise it is back to the drawing board!)

idk I tend to think revenge like that makes the person committing it look bad so they will look for something less public. Of course there are some who just don't care, but the vast majority of revenge on SE is 'deniable' — we will find out in due course of course :(

Since the total number of posts was limited, the user could only flag so many, so we counter-flagged all of them, but with time… Maybe the ability to flag should be earned (e.g. earn 10 stars without being flagged by anyone else)?

what sort of revenge? there isn't anonymous downvoting to contend with and less passive-aggressive revenge is going to be more public and I think easier to deal with. Account deletion is always an option though of course sock puppets are an issue everywhere

vandalism is a bit strong, it's more of an irritant at this stage isn't it? Right now any action we take (and the scope is unlimited) has to be taken by devs — the mod tools simply aren't in place. We can manually mass delete flags or prevent particular users flagging. Let us know if you *want* us to take action and we'll chat about what is appropriate…

currently: the flags and counterflags simply cancel, leaving the original flags which hide the post from unregistered users. However a crew member can't both 'flag' and 'counterflag', they can switch from one to the other at any time. Longer term when crews have >3 members, initial flags will notify 3 at random, and if there is not unanimity more random notifications will go out to get more eyes on the post — that's the rough plan anyway, it might evolves as we go along!

@Jack Quick question about "If multiple crew members flag and counterflag, the overall action is based on the majority." What happens in case of a tie? For example two clean up crew members used the flag to delete, then the posts gets edited and both of them then counterflag? Does this undelete the question or does it need 3 counter flags?

@Jack nice idea on the flag experiment. I think answers are more likely to be flag-worthy than questions, so if it's not hard I suggest adding that. If there's no confirmation on flagging (I haven't tested it), could you put a little more space between the flag icon and the subscribe icon? On a desktop with a mouse I can be precise; on my tablet with my finger, maybe not so much. (If the flag requires confirmation then this becomes merely an annoyance rather than a false flag.)

I will say there doesn't seem to be any appetite for star-score-based 'privileges', and that is a good thing. Extra features and abilities should be granted on merit, but I don't think we want to go quite the same direction SO/SE did with that. Adding arbitrary levels of complexity isn't usually the right solution to a problem.

Please don't do that, it wouldn't be at all helpful ;) Yes of course we will need to handle irritations like spam and vandalism. There are discussions about that here on meta as well as elsewhere. It's not being ignored, it just hasn't been the highest priority yet. If a problem like that were to happen right now, it would be dealt with manually, much as I guess SO/SE did in the very early days. They are only really able to keep on top of problems using free labour (Charcoal, other projects, users flagging) and there's no reason to expect something radically different will arise here.

Excuse me being kind of blunt, but this is kind of a naive approach — at least at the extreme you are presenting it as. You _NEED_ some moderation tools. For example you don't want to wait around for spam to happen before you start working on a way to cut it off an the knees. If you don't believe me I'll write a bridge between my spam mailbox and the site. For good measure I'll post each one via new accounts coming from Tor. I think you'll decide you need some tools pretty quick, and that the tools need to be in quite a few hands.

I'm not a fan of rep-based privileges (including moderation) because being good at answering questions does not tell us anything about one's suitability to wield tools like account suspension and content deletion.

SE got some things right and some wrong. I don't think mods should generally be the ones enforcing scope by (e.g.) closing questions; that's for the community. Which means the community needs to actually be able to do it; 5 votes are easy to get on large sites and hard to get on small ones, for instance. So that implies some per-site customizations of thresholds, but that's beyond the scope of a moderation discussion. My point is that mods should be exception-handlers more than curators. (As community members they should of course also be curators.)

On day one we need a way to make bad stuff disappear -- spam, hate speech, porn, etc. These cases, pretty much by definition, don't require subject-matter expertise, so that *could* be a few TopAnswers-wide admins. But it shouldn't just be one person, and it's probably better for community engagement if there's at least one person within a community who has those privileges.

This question seems too broad to be answered at the moment. We simply don't know what we will *need*. There is a temptation to put things in place too early, based purely on experience gained on SO/SE. We can usefully discuss specific site moderation needs as they arise, but for the moment I agree with Tom that effort is better directed elsewhere.

Yes I do ;-) I think though one of the issues is that the best answers to this question are actually going to vary by community a little. SE got away with the cookie cutter approach because the advantage of smaller sites getting the exposure being associated with the network largely outweighed the downsides it sometimes brought, but I don't think anybody trying to foster an alternative is going to get away with a one-size-fits-all approach.

I kind of get what you're saying but I don't think I can fully agree. There are at least half a dozen solid software platforms out there for doing Q&A. The issue for me (and probably many others evaluating this) is less about the software than about how the community will come together and how bad stuff will be dealt with. For me knowing how things will be managed is more important than working out the bugs in chat and what font to use!

For now I think there are more useful things to direct the effort to and hopefully by the time we actually need janitors the site will have grown more into a clear direction, and the community will have a clearer view on what issues might arise and what would be a good way to handle it

Technologically that scales easily, but I think it will fall quite flat. Reputation is generally come by by activity and is not a measure of sensibility. This is why not just SE but almost all forums have the concept of privileged users, and rarely if never a flat scale like that for getting them. I wouldn't say try to emulate SE right off the bat, but leaving out the "humans will be humans" out of the equation will eventually be a problem. Better to think about what mechanisms you'll use to solve that problem before they blow up.

for question deletion for example, I've been pondering a rep-lock: a user hits 'delete', hiding the post from everyone with lower rep; a higher-rep user can override the lock but only a yet-higher rep user can override the override

2 Answers