Earlier quoted context omitted.
> [1]: https://www.slideshare.net/dberkholz/assholes-are-killing-yo ... This is the kind of thing I'm asking about. Lots of numbers are trotted out but where's the actual data? Where's the methodology? The blurb says, "This talk will teach you, using quantified data and academic research from the social sciences, about the dramatic impact assholes are having on your organization today and how you can begin to repair…
No one considers it a good option. It's usually the last option, done after a fair bit of mediation to try and improve the assholes behavior. Only when it's clear that they can't or won't do you ban.
I'm saying that's still not a reasonable measure even in that case. Why not an exponential backoff, where the first measure is that they only get one post a day. If they want to be heard they have to be more careful in how they word things and they have more time to think about how it might be received. If they transgress again, then it's upped to every three days, then once a week, then once every other week, and so on. A total ban is the limit of this more nuanced process.
No doubt this feature doesn't exist, so I'm suggesting something like this should be added because I'm not at all a fan of bans. Even this is a stopgap measure used to manage assholes because we don't yet understand what's at the root of asshole behaviour.
Edit: to clarify, I mean the backoff/retry strategy is still not ideal, but an easy first attempt at trying to reframe this as a problem we can maybe address using programming abstractions to inhibit rather than facilitate communication. Most software is focused on reducing barriers to communication, which is why banning is the only recourse, but in cases like this you obviously want to raise barriers to communication in controlled ways so you don't have use the ban hammer.