This happened to me on another online community recently. I sent a PM but did so based on mistaken identity. A harmless mistake. The content, however, is not so harmless and might leave bad feelings for the unintended recipient.
I’ve been to a few communities that have such a feature (such as MUSHes). Where clicking a link (or for a MUSH, typing a command) before the PM is read by the recipient. The action retracts it and deletes it so the recipient doesn’t view the contents and never knows it was sent in the first place.
I understand if the current system doesn’t allow this and cannot be hardcoded in. My questions ask ways to deal with this if a user messages an admin that has permission to view all PMs and can query/edit the database.
My first thought is having the user edit out the content but if they don’t make the ninja edit window then the original message shows up in the edit history and negates the entire point of editing it. The other idea is having the admin go in and hard-delete it from view and that opens the whole “breaching privacy” can of worms that’s been debated ad infinitum. Don’t want to go there, please. My last thought is if the sender of the initial PM can hit the trash can button and that equals a retraction. The button would be available if the recipient didn’t read the message yet.
So what should be standard practice for this if using a Discourse-run community?
(Also, my concern has nothing to do with replies within a PM already in progress. This has to do with the very first message of a PM.)