Abstract
Deciding on the scope and level of detail of what is described by the data is an initial choice that determines the processing capabilities the DBMS will be able to offer. The relational model was established on three conceptual pillars: a declarative query language that specifies the data requirements, not how to fulfill them; the automatic translation of a declarative query into an algebraic expression that computes its results; and equivalence rules between algebraic expressions, allowing for the selection of the most efficient way to fulfill the query (E. Codd’s Turing Prize). We will highlight the advancements brought about by more recent models—such as object-oriented databases, tree-based models, and data graphs—as well as additional dimensions of modeling: time, probabilities, logical rules (ontologies), and provenance, which traces how data is produced.
