What is Fact Based Modeling

Fact based modeling (FBM) is a methodology (both notation(=structure) and protocol (=structuring)) for modeling the semantics of a subject area (i.e. the universe of discourse, the domain of interest) for the purpose of, for example, developing information systems, developing or applying laws or other rule systems or sharing information.

The main purpose of fact based modeling is to capture as much of the semantics as possible, to validate intermediate and final results with the subject matter expert in his preferred language, preferably using concrete illustrations and to remain independent of the representation for a specific implementation.

Unlike Entity-Relationship (ER) modeling or object-oriented modeling, fact based modeling treats all facts as relationships (unary, binary, ternary etc.). How facts are grouped into structures (e.g. attribute-based entity types, classes, relation schemes, XML schemas) is considered a software design level, implementation issue that is irrelevant to the capturing of business semantics. Avoiding attributes in the base model enhances semantic stability and understandability. Fact based modeling facilitates natural verbalization and thus enables productive communication with all stakeholders.

Fact based modeling provides the means to capture the knowledge of the domain experts in terms of "what" (i.e. the user requirements). FBM is conceptual, hence free of any software implementation bias. Fact Based Modeling adheres to the Conceptualization and 100% principle of ISO TR9007.

FBM is based on logic and controlled natural language, whereby the resulting fact based model (the conceptual data model) captures the semantics of the domain of interest by means of fact types, together with the associated concept definitions and the integrity and derivation rules applying to populations (facts) associated with these fact types.

The fact based models are used as a basis to derive implementation specific:

•    logical data models, such as ER, UML, relational and hierarchical models,
•    physical data models, such as SQL, XSD.

The roots of fact based modeling can be traced to research into and application in business of semantic modeling for information systems during the 1970s. Subsequently, several developments have taken place in parallel, resulting in several fact based modeling “dialects”, including NIAM, ORM2, CogNIAM, DOGMA and FCO-IM.