On Roles of Models in Information Systems - Intentional Perspectives on Information Systems Engineering - page 34

are associated with a unique label defined by the 1:1 mapping PID-N from all
Norwegians to the data type PNO, which is called the person number, more formally
expressed as PID-N: Norwegian ↔ PNO. We may now define the information set

NOR-WEIGHTS (PNO, WNkg2), and we may interpret the meaning of the cor-

responding data, as we have the complete semantic definition of every concept used.
This is admittedly a complicated way of defining the meaning of a data item.
On the other hand there is a certain economy involved in specification volume
because many of the concept definitions can be reused. Weight and age and person
number are reusable, as are scales and types. Essential is that domain models may
be modified without requiring a simultaneously modification of the corresponding
information models and data models.
New domain concepts can be defined relative to already existing domain con-
cepts. We don't have to modify already defined information definitions and data
models. New information set definitions can be done as additions to already existing
ones. Modifications of information systems can be done by replacing old infor-
mation sets with new ones, both new and old information being properly defined.
We can transform information sets and data structures while preserving their
semantics.
The following is an indication of what a full specification of information about
weights of Norwegians, Americans and British may look like at an early stage of
information systems engineering. In order to be able to express information associ-
ated with individuals on the level of sets of these individuals, even if we don't know
how to identify the individuals, we have to introduce a special quantitative concept
which maps 1:1 from a class concept to an unknown data type. This special map-
ping is called label and serves as a surrogate for a unique name for the members of
the class concept. Example: label(Person) stands for an unknown data type which
could serve as a unique identifier for all persons on the planet. But label is a quanti-
tative concept in its own right, and can be used to specify, e.g., PNO to be a unique
identifier for Norwegians.
Example: The objective is to specify information as a relation between data con-
cepts and UoD concepts. The data concepts are data types, and the UoD concepts
are class concepts and quantitative concepts. Scale concepts including measurement
units and precision come additionally.
First, the modeling in the domain:

=

C-concept Body, Person, American, British, Norwegian;
Q-concept weight: Body → positive real ;

Body includes Person;
Person includes American, British, Norwegian;

Then, we have to introduce appropriate data types:

datatype real Aweight, real Bweight, real Nweight, integer PNO;
Q-concept label: Norwegian ↔ PNO;

Q-concept weight: American → wA(lbs.d)

Q-concept weight: British wB(stone.ddd);

Q-concept weight: Norwegian

→

→ wN(kg.dd);