[Long Term] New Channel Types

I think the moderation approach is a little flawed here. There are many cases where you want to delete content to avoid people from being exposed to it. Say somebody posts some not-so-wholesome content in your wholesome community - you probably want to remove that as quickly as possible especially to avoid any younger users seeing it. In fact, in cases like these it's the most important that the user stays banned... and that does mean the evidence is gone... there is no longer concrete proof they did anything. There might be some solutions to this - but this is not addressed here! The most obvious to me would be a separate log view (like https://feedback.fluxer.com/p/1062/c/3020) that can perhaps be restricted based on roles. Bots on Discord also already support logging channels which work pretty nicely in my experience - though I'll admit they aren't very privacy friendly as the messages generally stay there forever unless you delete the channel and recreate it again (and most bots don't have a feature which automates this). This doesn't address a user appealing years after the fact either, but with your suggestion I feel that a moderator could just use the excuse that they deleted the content because it was very bad (which could also be perfectly reasonable) Something else not addressed is the problem of people sending and quickly deleting dubious messages before the moderators see them - which I have seen a lot on Discord - since you suggest that the messages should still be deletable

While I understand the privacy concern here, not logging evidence isn't really an option. I could do perfectly fine creating this functionality myself in the form a bot, but I'd prefer it to be a native feature with proper disclosure and consent beforehand (as my original post states). I fully understand that evidence / messages should be deleted after a while (especially upon user request), but your suggestion to just not log evidence is inappropriate for larger servers who value accountability on both sides (the user AND the moderator/moderation team). There could be some kind of compromise, but that compromise should not involve not logging evidence and retaining it for a short period of time.