Atomicity and Semantic Normalization

This paper contrasts two different approaches to designing relational databases that are free of redundancy. The Object-Role Modeling (ORM) approach captures semantics in terms of atomic (elementary or existential) fact types, before grouping the fact types into relation schemes. Normalization by decomposition instead focuses on “non0loss decomposition” to various, and progressively more refined, “normal forms”. Traditionally, non0loss decomposition of a relation requires decomposition into smaller relations that, upon natural join, yield the exact original population. Non-loss decomposition of a table scheme (or relation variable) requires that the decomposition of all possible populations of the relation scheme is reversible in this way. This paper shows that the dependency requirement for “all possible populations” is too restrictive for definitions of multi-valued and join dependencies over relation schemes. By exploiting ORM modeling heuristics, the authors offer new definitions of these data dependencies and non-loss decomposition, to enable these concepts to be addressed at a truly semantic level.