The practice of structuring information models and associated data schemas such that every data element is mastered (or edited) in only one place, providing data normalization to a canonical form (for example, in database normalization or content transclusion).
