1. 42

Full disclosure: I currently have the warning with 11 flags. The warning doesn’t bother me, and I’m only raising this because several other people are bothered by it.

A few months back @pushcx added a warning feature, that gives you this banner across the site if you’ve been flagged in too many comments:

Your comments have been heavily flagged across several stories in the last month. Review your standing for context on how unusual this is. Reconsider your behavior or take a break.

Currently the cutoff is 11 flags in a month. The purpose is to let honest users know they’re treading the line and give malicious users warning they’re on banwatch. To my understanding, several notorious bad actors have been banned after being warned this way.

I believe this feature is overall a net positive for the community. However, several people said the warning really bothered them. I suspect that it might be a little too harsh for people who are just above the cutoff. Any healthy forum is going to have disagreement, so it’s surprising to get a giant “you need a break” message if you disagree with people. If you go to u/<your-user-name>/standing you can see the breakdown of how everybody in the past month was flagged. We see

  1. 14 people had 11-20 flags
  2. 8 people had 21-30 flags
  3. 5 people had 31-50 flags
  4. 5 people had 51-150 flags

There’s a cluster of posters who are just above the cutoff and then a long tail of posters who are way above the cutoff. I suggest we have at least two tiers:

  1. People who are below 20 (30?) flags should get a friendlier warning. Something like “hey, several people are flagging your comments. It could just be disagreement, but you should still review the flags and see if you agree or not”
  2. People above 21 (31?) flags get the current warning.
  1. 18

    There had been a story about Emacs or something and I repeated the old joke about how it’s a great operating system but it needed a decent editor (I think…it was definitely some stupid joke about Emacs).

    The comment got flagged as off-topic, which is fair. But it got flagged by a lot of people (it was a popular thread) and it had been a slow month where I hadn’t commented much else, so the ratio was definitely tilted to the negative side.

    As a result, I got the warning.

    There were a couple of problems I had with that:

    • Because lobste.rs doesn’t show the comment’s score for ~24 hours there was no way for me to know for sure what the problem was (in my case it was obvious since I’d only had three or four comments in the previous month but for someone else it would’ve been a mystery).
    • The language of the warning is somewhat aggressive, IMHO. Making a silly joke once should not (IMHO) warrant the same warning as racism or spam or whatever. It certainly shouldn’t imply that someone who’s been on the site for years with (what I would hope is) a reasonably good reputation is in danger of being banned for telling an inoffensive joke without enough recently upvoted comments.
    • The window for analysis is perhaps too short. Someone with 18,000 karma can get threatened with being banned because they have one bad comment in an off month where they haven’t commented much else.

    In other words, I think the feature in itself is a good idea but I don’t want to risk alienating people who contribute a lot to the community, and I think for new users the warning should be perhaps more gentle unless the behavior is blatantly unacceptable.

    All this, of course, is IMHO.

    1. 4

      A special-case message when all the flags are overwhelmingly on a single comment seems appropriate, too.

      1. 4

        Quick feedback and a pointer to the issue definitely seems like a good thing. Tone is hard, and accidents and quick tongues happen. I personally want to be able to address issues promptly.

        At the bottom of your standing page, your flagged comments are listed. It’s hard to find unless you scroll to the very bottom. Since it uses the normal comment partial to render comments, flags/scores on young posts are still hidden.

        I think it would make sense to show the flag reasons on that page regardless of age, and perhaps place the list of comments (or a link to it) in a more visible place.

      2. 12

        I’ve definitely reconsidered my commenting since getting the warning. I used to reply to people who I felt were wrong instead of flagging, but since doing that just attracts flags that adds to my quota I’ve starting flagging instead.

        1. 3

          Without knowing the character of your replies (which is to say, if the replies in question are nasty and toxic, I would not feel this way) it seems at first blush that the change this inspired in you is not a constructive one.

          1. 3

            In the cases where I was flagged heavily my comments were neutral in tone.

            1. 17

              Flagging is not always done out of good faith. People may flag your otherwise neutral comment if they get personally offended. And mods explicitly wont listen to your feedback.

              IMO this whole “Reconsider your behavior or take a break” injunction is pretty stupid and condescending.

              1. 6

                Flagging as implemented on this site seems to be way too prone to abuse. Their open visibility encourages pile-ons, and the automated warning is confusing and raises umbrage. Flagging users who are trolling or spamming should be a signal to the moderators, not to the rest of the community.

                1. 1

                  I think the time and point threshold to show scores is a great thing here but could perhaps be tweaked to help. Maybe make it a bit higher/lower before it shows the scores, or make comments age a bit longer before showing a score?

                  The overall moderation model here seems solid enough. But if we wanted to get real wild Slashdot’s moderation/meta-moderation system could be an interesting fit ;)

                  1. 1

                    Trouble with the time delay is that highly engaged (ie controversial) topics stay on the front page for a lot longer than most posts.

        2. 12

          If we’re having tiers then perhaps they could be based on the kind of flag? If someone is getting a lot of flags for “troll” or “unkind” or “spam” that’s more serious IMO than “off-topic”, and even more so than “me-too”, or “incorrect” (in fact, I don’t think I’d expect “incorrect” to ever trigger a warning?).

          Hopefully the warning could be customised to the kind of flag: “A number of users have flagged your comment(s) as off-topic. Please try to keep discussion on-topic on Lobst.ers”; “A number of users have flagged your comment(s) as unkind. Please be kind towards other users, and assume positive intent in all interactions” or whatever…

          1. 2

            I totally agree with this. Not all flags are created equal in my opinion.

            Also, I’m not even sure I understand why “incorrect” is a flag? I lose points for being wrong? Maybe that flag needs renamed.

            1. 1

              TBH it’s not even clear that those thing are flags, or will have any consequence.

              I think what I’d really like to see is a combination of pointed tags, (-10 incorrect, etc…) and “a mod needs to be alerted to this” flag actions, which should be separated (I didn’t even know flags got mod attention until I was told off [correctly] for misusing one).

          2. 8

            I do get some utility out of the banner–I’ve had stretches of days, even weeks, without it. When it appears, I usually do a quick check to see if I’m totally off-base.

            For people that have only gotten it a few times, I’ll paraphrase Churchill:

            You have flags? Good. That means you’ve stood up for something, sometime in your post history.

            Obviously, this bromide doesn’t address the issue of pathological trolling.

            1. 2

              I got the original ‘close your account’ banner and promptly added a uBlock rule to hide the thing so I don’t have to see it again. The way comments are flagged is starting to resemble the way they are often downvoted on the orange site which I consider to be a sad thing since it has the potential to turn this place into the same type of echo chamber as HN hade become. I greatly prefer a true diversity of opinion with a possibly more grating discourse over a virtual kumbayah. It is outside your comfort zone where you’ll find opportunities to learn.

              1. 1

                I’ve never gotten the banner; this is the first I heard of it, or the ‘standing’ page. Looks like some heavy-handed Site Boss bullshit to me.

                So, we have quotas now? And I’m under-performing? Oh dear. I conclude that I should be more outspoken.

                1. 1

                  I’ve never gotten the banner; this is the first I heard of it, or the ‘standing’ page. Looks like some heavy-handed Site Boss bullshit to me.

                  I think the tone of the current warning is much softer than the original one, reproduced below:

                      If you are outraged by this notice from a site that's full of idiots and led by mods who are power-tripping assholes,
                      you can delete your account from the bottom of your <%= link_to 'settings', settings_path %>.

                  (from a Jan 13 commit)

                  Maybe still a little aggressive, but much more constructive IMO.

                  1. 1

                    The first one is a little more in line with my own sense of humor, but you have a good point. The Mods in Their Wisdom (let’s be honest, we’re really talking about the One True Mod) are just trying to avoid the site becoming a pit of flames like some others we all know, but over-protective conformity-assurance measures can smother a community too – I’ve also seen that happen. There’s no easy answer, once we accept the basic idea of some people enforcing norms on behalf of everyone.

                    My own tastes may tend toward the laissez-faire maybe more than some, but I’m not going to run around calling people snowflakes either. I just want Lobsters to be a place where productive disagreement, with a variety of personal styles, is possible. These kinds of policies make me nervous because they seemed biased in favor of the intolerant. But if it gets too boring in here, at least I know how to go out with a bang!

                    More seriously, I think this meta post presents a valid criticism: thresholds discard nuance, and it’s not really a matter of where to set the threshold. Moreover, lumping all flag flavors together seems to defeat the purpose of having “reasoned” flags in the first place. All of this has been put in place with no prior public discussion that I’ve seen, which seems a little disrespectful. But I don’t doubt that the pavement is made of genuine Good Intentions!

              2. 8

                Two additional thoughts beyond the core proposal:

                1. IMO the warning should make it clear that it could just be a false alarm and we should treat it as a health check. For example, it’s possible to get the warning for calling out other people’s bad behavior. A full third of my current flags, for example, are for calling someone toxic in the most recent V trainwreck.
                2. Should lower tier warnings go away earlier? Like you only see them for 7 days, not the full month.
                1. 6

                  I don’t know how much time our moderators have on their hands but IMHO this is something that’s pretty tough to handle based on numbers…

                  A very long time ago I was one of the (20 or so?) moderators on a pretty high-traffic forum. It was a computer magazine’s forum, from back when they were printed, on dead trees and whatnot, and it was one of the big ones around here. The day a new issue showed up we’d sometimes get post counts in the four-figures.

                  We briefly experimented with a “flag” thing at one point and it was pretty bad (it wouldn’t give automatic warnings, but once you got more than 20 of those, I think, it showed up in red under you avatar next to the offending post, and if you got enough, it gave you a “badge” under your avatar everywhere). People would say something not even that controversial, but in a bull-headed kind of way (“Quake III is really stupid, regardless of how cool the graphics is, the guys at id are so uncreative they couldn’t even come up with a story” – that’s how long ago it was, yeah) and they’d get flagged to death. Meanwhile, long-time lurkers would throw anti-semitic hints in low-interest topics and fly under the radar of the flag count because they just didn’t say stupid things “in front of” enough people. (Then we’d ban them and they’d make a thousand clone accounts a day to complain about how they got banned without an appeal, oh God I had so much free time back then…).

                  What did happen, though, was a flurry of “why am I getting flagged???” posts. Many of them perfectly justified – people would understandably go bananas over the fact that they said something stupid about Linux in the Linux section and got a badge of shame, whereas this other user could say this really offensive thing in an obscure topic without anyone batting an eye.

                  IIRC we disabled the badge thing, then we tried to use it in an “amortized” version after a few weeks, where you could only “upvote” (I don’t recall how we called it – it was way before Reddit & co.) posts if you thought they were cool. Then we dropped it altogether. It was useful, but we were afraid that, even if people eventually understood that it was just a possibly meaningless number, it would still encourage people to flag each other instead of discussing things.

                  Based on that experience, and on the general involution that upvoting/downvoting brough on online discourse, I don’t think flags should cause any kind of warning to pop up for anyone except moderators. Honest users who are going through a rough period and taking it off on others online are going to have an even worse time if, instead of someone (gently, if possible) telling them that ain’t cool, they get a random warning from some Ruby script. Trolls don’t care where the threshold is, and any threshold is going to be bad for chronic narcissists.

                  Basically, I think it’s just bad for people who are honest one-off offenders, and useless for anyone who isn’t and is going to get banned anyway. IMHO we should optimize things for the former kind, not the latter.

                  1. 5

                    I would be interested in seeing some statistics on comment-frequency to flag-count. Commenting and participating in conversations will naturally lead to more exposure and people who disagree with you. Reviewing my standing, 90% of the flagged comments have a 5:1-10:1 ratio of up- and down-votes, which also doesn’t seem to be properly

                    And on a related note: I can’t say for sure, sometimes I had the impression that a few of my older comments got flagged overnight after I disagreed with someone a day before (usually in meta threads…). Coincidence or just people browsing my comment history, I don’t know, but it certainly skews my statistics.

                    I have seen the warning a few times, and I usually get the feeling it contains more false-positive than comments I should reconsider. I’m not saying I’ll do it, but I can imagine that others just use uBlock to hide the notice, if it doesn’t become more accurate.

                    1. 2

                      I would be interested in seeing some statistics on comment-frequency to flag-count.


                    2. 8

                      The purpose is to let honest users know they’re treading the line and give malicious users warning they’re on banwatch

                      Does it though? I have seen a lot of comments flagged simply because the reader didn’t like what they had to say. Threatening a user with flags serves only to enforce group think. Malicious users already know what they are doing in the first place.

                      1. 3

                        After reaching out to pushcx when the feature was first implemented to make sure I wasn’t in the wrong I just added the element to ublock and forgot about it

                        1. 2

                          I personally don’t think that this fixes the issue. It seems like it just makes things more complicated to me. I think that something like aggregating (downvotes / upvotes) > 1.3 (or some threshold) over the last 30 days. If some number of posts meet that, then the person is getting downvoted more than upvoted. That would have left me without the warning and probably others, while still pointing out people who get mostly downvotes which is what I believe the intention of this feature is really wanting to point out? Number of flags means nothing, because it also a factor of how often you post.