Mirror mirror on the wall
I want to talk about a common mindset that seems to rear its head in most of our processes, both public-facing and not: The idea that it is imperative to have nothing bad ever happen.
At first you might go "of course we wouldn't want bad things to happen, what's the issue?" but that's not quite what I'm referring to. The issue is the thought process that if nothing bad ever happens, it is preferable to any other system where something bad may happen. For example: Our greenlighting system is designed to ensure no bad articles hit the critique threads. Our joining process, in particular the passcode, is designed to ensure no users who join have skimmed the rule pages. Sticking on IRC was done, among other reasons, due to IRC having a higher bar of entry (seriously). All of these processes ensure something "wrong" never happens in our systems.
Unfortunately, all of this comes at a price: less good stuff happens. Just because something isn't bad it doesn't mean it is good, and designing our systems around loss prevention makes us forget to use our systems to improve the user's enjoyment of the site. We kept the sandbox under lock and key because of one troll, while sacrificing the usability of the sandbox for hundreds of users. We've stayed on IRC to not have more annoying users on the official platforms, and made it so that no one wants to use our official chat platform other than a handful of users. We've made dozens of guides "required reading," leading to new users feeling like the site is impossible to join and incredibly elitist.
We need to drop this mindset wholeheartedly, and start using systems that allow bad situations to happen, maybe even a lot of them, if they give our userbase a net gain.
This mentality of "take risks if it'll help overall" can be seen everywhere: in system programming (as linked in the thread description), in competitive games (such as fighting games or magic the gathering), and almost anywhere else that deals with risk/reward. There are of course certain areas where avoiding any bad situation is preferable, but that's only when any loss is so great it cannot be afforded. I do not believe having a coldpost reach -10 is critical to the site's functioning to such a degree we have 30 preemptive systems to avoid such an occasion.
This thread is here to discuss my post here, see what people think, and hopefully move onward with a stronger idea of what we want to see from our site.