Re: The Theoretical Foundations of the Relational Model

Alfredo Novoa wrote:
> > ORM offers the least inflexible method for capturing conceptual designs
> And what about using text?

I'm not talking about the representation of the design, and besides, ORM has a textual form. I'm talking about postponing conceptually unimportant decisions (mainly the decision about which things should be entities and which attributes) until the logical or physical design stage (when this decision must be made). The decision about fine- or coarse-grained inheritance hierarchies, or whether inheritance should be used at all, is also omitted from an ORM model. In that respect, ORM is neutral on the relational/OO question, and offers itself as common ground. It captures the conceptual map without saying or implying anything about what physical structures might be needed to efficiently represent and manipulate the information.

> I don't understand the need for doing all graphically.

Agree. Parsers for graphical languages are not at all well-understood, whereas those for textual ones are, so any graphical language should always be mapped to a textual language so that it can be formally verified.

--
Clifford Heath