02.08.11

Modellierung: Datentypen, Klassen und Terminologien

Für die Modellierung ist es sinnvoll, diese drei Bereiche zu unterscheiden: Ein Artikel von Steven Waldren beschreibt das so:

"Primitives are items that abstract away all the underlying complexity that does not have to be addressed by the data modeling exercise. They are the building blocks of the model. Primitives are also items that are not specific to the domain you are modeling and would be reused across domains. Data types (e.g. String, Character, Integer, Date, Time, etc.) are typically primitives. From these primitives more complex types can be built. I posit for health care there are two categories of primitives that are needed: Data types and Controlled Vocabularies."

"(The) top layer items, which I will use the term classes, have characteristics that need to be modeled. These characteristics maybe called attributes, properties, facets, or slots. These are probably the closest to the generic term “data element.” For example, the Actor class could have attributes of Name, Address, Occupation, etc. Each of these attributes would be defined with a name and a data type. The data type could be a primitive data type or another class in the domain."

Und: "There is a need to create a value set or subset of the controlled vocabulary for a specific instance of an attribute in a class. "

Keine Kommentare:

Kommentar veröffentlichen