www.entitylogic.org - entity modelling introduced from first principles
An entity model is a system of entity types and relationships and for a reader of a model the most important relationships - those to understand first - are those that represent associations between entities and their parts. These relationships are called composition relationships and the term composition structure is used to refer to the subsystem consisting solely of entity types and compositions relationships.
Composition relationships are shown top-down, which is to say that they are drawn leaving the lower edge of the box representing type of the whole and entering the upper edge of the type representing the part, as here: (a) . This fragment signifies that there are one or more entities of type part type within the whole.
Looking at composition relationships the other way around - bottom up - then they are seen to relate entities with the contexts in which they exist and it is because of this that these are the most important relationships in an entity model - they provide context to entities.
The presence of the crows foot is representative of multiplicity - if the crows foot is present the notation asserts that there may be many parts of type part type within each entity of type whole type.
If the crows foot is absent, as here: (b) then the assertion is that there is exactly one entity of type part type within the whole.
A further distinction is made by use of a half-dashed line to represent the possibility of zero; this gives us
(c) there may be zero, one or more entities of part type within the whole:
and (d) there may be zero or one entities of type part type : .
If there are parts of different types, then the structure is shown branching as for example here: (e) or here: (f) .
Figures 1, 2 and 3 gives basic examples using the notation. More examples follow as more features of the notation are introduced. If you find yourself disagreeing with these examples - thinking that what they express is contrary to your understanding then it is reasonable to suppose that I will have achieved my aim of showing how the notation works - how it can be used to express precise models and that it is then possible to consider these models, to disagree with them or, better, to refine them.