Relations

Relations in a Top Level Data Model, and data based on it, are fundamental and have some formal characteristics. They represent what one item (thing/object/type, etc) has to do with another AND about which there is "nothing more to say". If a relationship is expressed between two items then it is logically present (TRUE). However, these binary relations are more restricted than the relations present in mathematical set theory (e.g. they can't be symmetric). They always have an order and therefore represent a set of ordered pairs, or an ordered pair that is an element of such a set.

The three foundational relationship groups in HQDM are:

• parts
• sets
• order

Topics on relations that may be added, if there is interest:

• Relations and instances of them. How to implement and query for them generally.
• Supertype-subtype relations are really important (especially considering the multiple-inheritance characteristics of HQDM) but are easy to overlook. Handling them in code may be worth documenting further.
• Specialisation of relations. The foundational relations (parts, sets and order) in HQDM have many specialisations. This is not formalised in most implementations but there is a specialisation hierarchy for most of the relations in HQDM. This, combined with the type-hierarchy, could provide a path towards a Domain Specific Language for interacting with the data (and the patterns documented on this site).
• EXPRESS as a data modelling language has the ability to redeclare relations. This is intended to assert that when a subtype inherits a relation from a supertype that relation is a required feature of the subtype (rather than just inheriting the wider-scope of that relation from its supertype). Handling this in the implementation of HQDM may benefit from further work.