> * The relational model: We must stress that what we are not doing is proposing some kind of “new” or “extended” relational model. Rather, we are concerned with what might be called the “classical” version of that model; we have tried to provide as careful and accurate a description of that classical model as we possibly can. It is true that we have taken the opportunity to dot a few i’s and cross a few t’s (i.e., to perform a few minor tidying activities here and there); however, the model as we describe it departs in no essential respects from Codd’s original vision as documented in references [20-22]. (Indeed, we even toyed at one point with the idea of calling the book A Relational Model of Data for Large Shared Data Banks.)
> * Type theory: Relations in the relational model have attributes, and those attributes in turn have types (also called domains). Thus, the relational model certainly assumes that types exist. However, it nowhere says just what those types must be, nor does it have much to say about what properties those types must have; in other words, the relational model and type theory are almost completely independent of each other (they are orthogonal, to use the jargon). In this book, we present a detailed theory of types that complements the relational model, and we explore some of its implications. In particular, we show that our type theory serves as a basis for a comprehensive model of type inheritance.