Follow
Huh, I think I might have found a solution for the problem in Matrix where 'roles' are really difficult to implement because they break orderability assumptions in the stateres protocol
I doubt this solution will ever make it into the Matrix protocol, to be clear. I gave up on the MSC process a while ago. But I'll definitely be working on it for #Atoki.
Curiously, it looks like this problem can be solved with basically the same mechanism as the Spaces access control replication MSC that I wrote a while ago