23

I welcome any details that can be shared, but I understand if there are specifics that cannot be. I am primarily interested in whether the process exists, rather than what that process is.

What's been happening:

In the last few months, and especially the last week or so it seems, there is a particular user that has been blatantly abusive of the Stack Overflow site. They post questions tagged with the C# tag that seem superficially reasonable, but which are at best overly vague and poorly researched, and at worst are really just a "word salad" of click-bait programming words like "object", "boxed", "heap", "literal", "reference", and more.

Not only do the questions not contribute, the user then takes advantage of their lack of clarity, which always induces other users to post comments asking for clarification, to drag those other users into extended, fruitless discussions in the comments under the post.

If someone should happen to answer the question, this user then takes that opportunity to start the same sort of extended, fruitless discussion in the comments under that post.

In my view, this seems like straight-up trolling. I will admit, it's more sophisticated than a lot of trolling is (maybe even aided by a linguistic AI to compose these questions…that would explain all the buzz-words that when put together don't really mean much). But it's trolling, nonetheless.

Initially, it sufficed to just downvote and close the questions. But this quickly escalated to the user creating multiple user accounts, posting the same question using different accounts (and probably sock-puppet-voting on them too). When the community got wise to that tactic, they got abusive. They posted abusive comments, and even edited the questions themselves to include abusive language, including death threats.

Every time they create a new user and others post comments calling them on it, they lie and pretend that they are a brand new user, never having visited the site before. This lie doesn't last long though, because they soon after devolve into insulting the community members who are voting, flagging, and commenting on the posts, revealing themselves to be the same.

Suffice to say, such a user has no place on the Stack Exchange network at all.

What I know has been done about it:

The community has done a good job keeping up with the posts. They now get closed and deleted soon after posted, often within minutes. Moderators are apparently on top of things too; while I don't get any specific feedback on the matter, my flags are marked "Helpful" and the accounts created by the user have been suspended, and eventually deleted (of course, I don't know if the latter happens by the user themselves or by moderation staff).

But is it enough, and what else can the site do about it?

It would not surprise me at all if Stack Exchange had a mechanism for banning users by IP address. But as we all know, it's not hard to get around such bans. Use of a VPN or Tor can easily bypass a ban like that, unless Stack Exchange is willing to ban whole VPN services and Tor exit nodes.

If I recall correctly, creation of a user account requires that the user provide an email address. But of course, there are plenty of ways to get throw-away email addresses.

Because the Internet is so accommodating of anonymity, and because of the relatively open policies for Stack Exchange (i.e. nothing that requires any sort of reliable proof of identification…to be clear, I'm not proposing that this openness change), it may be impossible to verify a user's true identity in any reliable way. So in theory, a persistent user could just keep bypassing ban after ban after ban, continuing to harass the community.

From my point of view, not knowing the behind-the-scenes capabilities of Stack Exchange and its staff, it seems to me there are only two strategies we can really take here:

  • Give up. Let the user post. Let other users post answers and then suffer the endless comment discussions. Hope that the more-active users will all eventually get wise, and that the less active-users won't have to interact often enough with the user for it to be a real problem for them.
  • Keep moderating the posts and the user into oblivion. This is very time- and resource-intensive, but it keeps that particular mess off the site. For now, the user seems perfectly happy to just keep trolling, but I'm sure that as with every troll in the history of the Internet, they will eventually get bored of whatever thrill they were getting and go away.

I don't find either of these strategies particularly gratifying, but I do admit that they seem like possibly the only reliable approach.

Other than the above, is there actually anything else the Stack Exchange sites can do about users like this? Are there in fact other actions being taken behind the scenes that can more reliably put an end to the abuse of the site by users like this?

6
  • 6
    "of course, I don't know if the [account deletion] happens by the user themselves or by moderation staff" At least one of the accounts you're referring to was deleted (or destroyed) by a moderator, as it happened before the user would have been able to (there's a 24-hour timeout). – Ryan M May 25 at 3:52
  • 3
    It's happened a couple of times on Skeptics, but the mods are pretty quick to be firm there. I understand that there's some reference to IP addresses to be had, but what action to be taken may involve Staff, and the consideration that many IP's are shared. – A Rogue Ant. May 25 at 4:00
  • 15
    Just as an additional note - trolls, in general, want attention as their main or sole purpose. The best way to handle them is flag them for mods and let mods silently nuke them. This way, they don't get the attention they want, and it's far easier to destroy an account than make even a throwaway account and circumvent restrictions and evade moderators. If it's not worth it for a troll, they will eventually get bored and stop. – hyper-neutrino May 25 at 4:38
  • 2
    I can see that some of the stuff from a few hours ago was deleted from within 20 seconds to 3 minutes; they must be on a break. --- Don't give up, or take it upon yourself to do everything; many people are working on that (and the various names, predominantly one subject) and other things. Some months it's quiet, some days it's busy - that's how it goes. --- Success is better elsewhere, but some people like to try here. – Rob May 25 at 4:42
  • 3
    If you need more users with close votes for unclear / low quality posts do know SOCVR can help. – rene May 25 at 7:26
  • I would try taking it to chat and asking the user why he exhibits such a behavior? It'll be clear and it'll give him a single place to have that interaction publicly while denying him the ability to spread it out on multiple posts. Intent becomes a lot clearer if you don't let the medium hide it. I would also put down exactly what you're looking for in the interaction -- ie, no questions? more research? etc. – I Support The Boycott May 25 at 22:23
17

So, I do not moderate on Stack Overflow, but rather Code Golf, and I will offer a bit of insight. Maybe I'm allowed to share more, but this is just what I think I'm reasonably comfortable publicly talking about, since giving away all of our tools to the public is a great way to get exploited even more.

Due to SE's core policies about allowing anyone to participate, like you said, it does also result in more negative or harmful activity. Now, there are systems against this, such as automatic IP bans like you suggested. However, with VPNs and TOR, throwaway temporarily email address generators, and things like that, it's hard to properly filter out people we don't want on SE ever again without harming others' ability to contribute - or even at all, without caring about the latter group. I won't go into much detail about the mod tools we have available, but even with them, obviously no system is perfect.

Now, to address the question of what more can we do? Not much, to be completely honest. If it's getting serious enough to the point that the moderators are being totally overwhelmed and can't stop it, we can raise issues up to the Community Team and they have more extensive measures they can take against things like this that aren't available to us.

Until then... just keep flagging and reporting. It's annoying, yes, but the tools for us to get rid of problematic users don't actually take a whole lot of effort to use, it's just that SO, even with 24 mods, is so active that they're always busy. In fact, the effort a troll needs to go through to bypass bans, bypass all restrictions, try to evade moderators, and keep trolling is much more difficult than the few buttons I need to press to put that all to waste. So, most trolls eventually find it's not worth it.

We've dealt with things like this on the network before. In fact, there are persistent trolls that are still around. I hope this troll gets bored and goes on to do something useful with their life, but some never do. We can only keep using the tools available to us to keep the noise ratio on SE as low as possible. No system will ever be perfect, and we'll never be able to screen out all trolls before they can post or even get rid of all negative content retroactively. We just need to do our best.

So don't worry too much about it. There isn't much special that can or needs to be done. Trolls that get bad enough will eventually face the wrath of the CMs, but removing trolls is a constant process. After all, if the system was perfect at preventing bad things, we wouldn't be needed here :)

6
  • 1
    Thanks for the feedback. I found this useful, especially the confirmation that there are other processes available to moderators and Community Team, as well as that it is not adding a lot of workload to moderators' plates to handle the individual posts. – Peter Duniho May 25 at 7:24
  • since giving away all of our tools to the public is a great way to get exploited even more That's called security through obscurity, and it is bad. Releasing the tools wouldn't get you exploited more, but would allow people to improve the tools significantly, faster than trolls could take advantage of them. – forest distrusts StackExchange May 25 at 22:30
  • 1
    @forestdistrustsStackExchange I'm aware of that. Still, even though a secure system shouldn't be secure only because people don't know of its inner workings, it's still a good idea to not give those away so freely, and as we all know things take a while to get done on SE so it's better for us to find exploits when working with these tools than for a troll to find them and exploit it until we can get them fixed. Also, I'm just following what (I believe) is policy anyway. – hyper-neutrino May 25 at 22:32
  • @hyper-neutrino Nothing wrong with following policy, but infosec is my specialty, and it's shown me time and time again that releasing tools always results in those tools being improved much faster than could be done in-house. As it is, the most dedicated of trolls could find these "secrets" out anyway, but those of us who are knowledgeable in dealing with trolls and tools against them are unable to suggest or implement improvements. As it is, the trolls have an advantage. They can test boundaries and find out how to get past filters, but "we" (community) can't find out how to improve filters. – forest distrusts StackExchange May 25 at 22:34
  • @forestdistrustsStackExchange Ah. Yep - definitely agree. I for one trust open-source security far more because if someone could break it they would've and there are no secrets about how it works. Granted, the things users can do to evade moderation tools... aren't all that clever. People do them all the time. It's not too hard to figure out that a VPN would help avoid sockpuppet detection, but we still have ways of figuring it out, and there will always be trolls, it doesn't matter much if it's a new one each time or a recurring one that's found a loophole, we'll get rid of them all the same. – hyper-neutrino May 25 at 23:40
  • 1
    I appreciate all the great input; the other two answers provided very good information for me as well, but IMHO this one does the best job of directly addressing the original question, regarding what besides the measures I'd listed actually do happen. Naturally, the other advice found elsewhere on this page about minimizing interaction with trolls and letting them get bored is highly valuable. But knowing that there are other things going on to help matters is appreciated, and very much what I was asking about. – Peter Duniho May 26 at 6:05
16

There's not much you can do when the reward for the hyper-fixation is the fixation.

There are some folks that simply never get tired of throwing stuff at the wall, and they will continue doing it relentlessly as long as they're able.

Stack Overflow has seen a dozen or so of these folks in its lifetime, some of them only stopped because they sought (or were required to seek) long-term care for the underlying causes of that behavior. Some instances resulted in folks fixating on certain users and employees and, as someone once on the receiving end of it as both, I can relate to how exhausting it can become.

My best advice in those cases is to do exactly as you stated - summarily delete the content and terminate the accounts and just hope that the person or persons responsible eventually lose the ability to so easily cycle through them. I don't want an internet tied to identity, so I understand this is just necessary work that people have to do in order to keep it that way as long as we can.

Then, there's "researchers"

... and you do have to keep a watchful eye out for them. They will conduct breaching experiments of every conceivable kind and their motives are seldom clear. Stack Overflow is huge, as are some of the network sites and very conducive to this kind of research.

Generally, researchers look at past interactions that you can fill in from the public database dumps, but some actively see how certain kinds of interactions change outcomes given the same setting.

If mods spot patterns (like university IP ranges, emails, etc) they can and do involve staff to see if those involved really know the impact of their disruptions on volunteers - and it can escalate further.

But that's more to make you feel better, not necessarily as a prescription to stop it.

Yeah, it's an ancillary cost of not wanting to create even more friction for people that want to legitimately use the sites. I completely agree with others in saying that most of this type of behavior just stops, usually within days, seldom weeks or more.

But when it gets past 'weeks or more' it can (unfortunately) mean a bit of a long haul.

1
  • 2
    There's also a different kind of academic nuance to it - those that have highly-complex research-level math constantly floating around in their head can sometimes lose touch with the natural world and their natural senses. That's actually .. pretty normal ... in some fields. So "toxic" itself can be somewhat relative. But in the end, it's about who is feeling annoyed or threatened by the behavior that matters just as much as the behavior itself. If it's affecting folks adversely, including the one doing it, it needs to be stopped unceremoniously. – Tim Post May 25 at 15:20
14

There's a pretty consistent fact about dealing with trolls...almost all trolls eventually get bored and leave. The amount of effort required to continually evade bans only for their posts to be summarily deleted with very little reach just isn't worth what little attention they get. The asymmetry between the effort required for a moderator to nuke their post and account (a couple clicks) and the effort required to evade moderation helps a lot here. Even the community just needs six red flags (see the "What effects do these flags have on a post?" section of this answer) to remove any post and alert the system of what the user's up to.

I can count on one hand the number of trolls (plus another hand if you count spammers) determined enough to keep posting after any non-trivial amount of time with all of their posts getting removed.

Thus, there generally isn't a need for anything special. If a troll is determined enough to get on that list of a handful of people, they eventually become known to community managers who make sure that all of their posts are removed from the network (obviously not going to name any such trolls, because, well, feeding trolls is bad).

4
  • " the community just needs six red flags to remove any post and alert the system" -- can you elaborate on that? While I agree with what you wrote, the most-applicable element of this answer could be more specifics about how flags are handled. In particular, I've noticed some of the posts have been censored as "spam or abusive"...does that happen because some flag threshold, maybe of a particular type of flag, has been reached? Actions the community can take without requiring a moderator's input could help things be more efficient, and would reduce workload on mods. – Peter Duniho May 25 at 6:47
  • @PeterDuniho sure thing, I added a link to the FAQ that explains the details of how red flags work. – Ryan M May 25 at 6:50
  • Thanks, that's very helpful. It sounds like, other than what we in the community are already doing, the most effective improvement (for those of us that weren't aware) is to use the "rude or abusive" flag, to "grease the wheels" of moderation, possibly even to have action taken without a moderator having to intervene. I wonder if, in that case, there's even any need for e.g. a custom flag to alert the moderation team about the user specifically. It sounds like they would get on their radar pretty quickly regardless if the post is deleted in this way? – Peter Duniho May 25 at 7:08
  • 6
    Custom flags are still helpful, because they provide context that might be missing from an isolated rude/abusive flag. Stack Overflow, for instance, has 24 moderators. It's possible that the flags go to a different moderator each time. Providing the context in a custom flag gives the moderator handling the flag the backstory. However, there's no need for multiple custom flags per post: if you know there's a custom flag on the post (for example, if a user in a chat told you they cast one), then following up with more rude/abusive flags makes sense. – Ryan M May 25 at 7:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .