There is also anchor modeling [1] with quite a few publications [2] and even an online modelling tool [3] to play around with. It is essentially a method that y

There is also anchor modeling [1] with quite a few publications [2] and even an online modelling tool [3] to play around with. It is essentially a method that yields a highly normalized bitemporal database model so that every prior database state, schema and data, is a subset of the current database state. I was personally not able to use it yet but I really like the idea behind it. It is probably not the ideal choice in general due to its append only nature ever increasing the consumed storage space and due to its high degree of normalization which may or may not have a negative impact on performance when implemented on top of a rational database depending on the nature of your queries. But if you do not have to deal with excessive amounts of data, if you expect a lot of schema evolution, if you need traceability of all changes for auditing purposes or such, then using anchor modeling might be worth a shoot. [1] https://en.wikipedia.org/wiki/Anchor_modeling https://en.wikipedia.org/wiki/Anchor_modeling [2] http://www.anchormodeling.com/?page_id=360 http://www.anchormodeling.com/?page_id=360 [3] http://www.anchormodeling.com/modeler/latest/ http://www.anchormodeling.com/modeler/latest/