At small scale the answer usually is "there are not just exact rules, you distinguish by looking at it and the standard is whatever yours is". I.e. take HN as a

At small scale the answer usually is "there are not just exact rules, you distinguish by looking at it and the standard is whatever yours is". I.e. take HN as an example: there are rules, and I assume dang has his own set of personal guidelines, but I'm fairly sure in the end someone getting banned comes down to dang checking someones history and going "ok, that's enough". Although here it's very much not a one-strike policy, which somewhat changes things, but the principle is the same. And as long as one person or a small team is doing the moderating, that's going to be consistent enough to work for a community. Also always an individual decision how much you're willing to tolerate "just annoying" - too much of that can also break a community. You also have the pattern of rule-lawyering by being obnoxious while staying just inside the formal rules if there are too strict rules and too little moderator leeway. The problem with your usual social media giant is that a) you by necessity have large teams of moderators, so you get kind of forced down the path of trying to exactly specify each rule and b), IMHO more importantly, they are a community, but contain many of them, with different contexts making different things appropriate, and people having different expectations of what's appropriate and what's not. (you see that conflict also already happening on relatively-small HN) And of course the impact on the individual is potentially larger than if they just get banned from one small forum, so your struggles of implementing it well also have outsized impact.