A huge practical challenge when implementing bitemporality is how to deal with schema migrations efficiently whilst supporting arbitrary time-travel queries. When we were building Crux, which is referenced at the end of the post, we took the view that this is a hard problem to solve in a generic way, hence why we have opted for a schemaless document-oriented design.

I really hope to see other databases make it easier to use bitemporality in the near future, but I suspect that any DBMS which mandates a schema is fighting an up-hill battle.

Disclosure: working on Crux at JUXT :)