Great. I have a workable solution to support the joined-class scenario and replicating the end revision timestamp accordingly. I'm still looking over Daniel Nuß's join table scenario, although I'm inclined to believe that is a slightly different issue all together.