www.entitymodelling.org - entity modelling introduced from first principles - relational database design theory and practice - dependent type theory
In essence an entity model establishes, or at very least proposes, a domain of discourse. The domain can be relatively familiar consisting of entities such as 'house' , 'street' , 'person' etc. or can be technical consisting of entities such as constituent parts of grammar, for example 'noun phrase' and 'verb phrase' and so on, or constituent parts of chemical structure such as 'atom' and 'covalent bond' . The primary function of entity modelling is to define and present such domains of discourse and to do so in a way that eliminates error and ambiguity.
Sometimes the domain of discourse can be abstract such as when we use a metaphor based language term such as 'network' in the noun phrase 'social network' and by which means we seek to identify a conceptual pattern and bring to mind imagery of points and lines independently of what the points represent - whether that be social entities, railway stations or atoms. It is one of the functions of entity modelling to present and propose such deliberately abstract domains of discourse. In this one mentioned the principal entities will be 'point' and 'line' . Breakthrough abstractions in science often move into language as metaphor and this is just because abstractions have a metaphorical shape which gives them utility beyond their origins; by virtue of being abstract they are reusable and ever applicable in new situations. As diagrams of concepts, entity models give visual shape to this metaphorical shape of our abstractions. The point is made by the database pioneer Charles W. Bachman in his 1969 paper "Data Structure Diagrams" when he draws in outline a diagram representing a manufacturing system with 39 interrelated types of entity and having previously described to us these certain patterns asks of us 'Can you...find a five-level hierarchy? Two simple networks? Five compound networks?' and, of course, we can.