I've been thinking of exploring temporal modeling through introducing "one at a time" uniqueness constraints. That is, where the mapping takes care of the mapping of the history of a given N-ary fact type, by internally injecting a time role into the fact type, and enforcing the time roles...