Errors and wrinkles in HQDM as documented

[ClimbingAl]

Issues and recommended strategies for dealing with them for HQDM as originally documented

The book Developing High Quality Data Models contains, in Chapter 17, the model that the Magma Core repo implements. The original model contains 231 entities and many relationships. Type and relationship inheritance is vital for the enforcement of the model patterns that the book documents. Some errors, omissions and wrinkles have been discovered and it's worth keeping track of them and their consequences.

[ClimbingAl]

Relationship inheritance issue in entity types Aggregation, Composition and Temporal_Composition
Aggregation is a subtype of relationship (an entity type that is a reification of types of binary relations) that should inherit the member_of relationship that relates instances of relationship (and its subtypes) with class_of_relationship (or its corresponding subtype(s)). This pattern is shown in section 17.3 of the book but the listed subtypes (documented in section 17.4) don't show the member_of relationship and a corresponding subtype of class_of_relationship.

[ClimbingAl]

The entity type class_of_representation is missing a SUBTYPE OF statement
In the book, section 17.45, class_of_representation is missing a statement "SUBTYPE OF (class_of_association)". This is correct in Magma Core (possibly only in the develop branch) but, in case of using the book as a reference or using legacy code, this inheritance path is missing.

[ClimbingAl]

The entity type functional_system inherits intended_role from functional_object. However, the corresponding relation intended_role_by_class is missing from kind_of_functional_system and from kind_of_functional_system_component. From the linked diagram for kind_of_functional_system it can be seen that intended_role_by_class should be present as an equivalent commitment to natural_role_by_class for kind_of_biological_system.

The reason for this is seems to be that intended_role_by_class is introduced to kind_of_functional_object but isn't propagated down to these system related kinds as they inherit from class_of_functional_X. Addressing this formally in the model would require some careful analysis but using intended_role_by_class for kind_of_functional_system and kind_of_functional_system_component is consistent with the general X_role_by_class pattern documented by Matthew can be useful.