Database Reference
In-Depth Information
system. Its dominant paradigm for studying organizations is the
PSI-paradigm
(Performance
in Social Interaction). The matching model type is the
white-box-model
.
Organization science and organization engineering are complementary fi elds. The
former is particularly useful for
managing
organizations (strategic, tactic and operational
managing
management), while the latter is especially useful for
changing
organizations (redesign/re-
changing
engineering of business processes, forming networks of organizations, etc.).
The PSI-paradigm states that an organization consists of people who, while communicat-
ing, enter into and comply with commitments (social interaction) about the things they bring
about in reality (performance). This reality therefore is to a large extent an inter-subjective
reality. Put differently, in their social interaction people engage in obligations about actions
to take, and reach agreement about the results of those actions. The PSI-paradigm is made
more specifi c and operational in DEMO as described later. DEMO belongs to a group of
modeling approaches that are all based on the Language/Action Perspective (e.g., Goldkuhl,
1996; Medina-Mora, Winograd, Flores & Flores, 1992). Van Reijswoud and Dietz (1999)
provide a detailed description of DEMO.
Object-Role Modeling
(ORM) is a fact-oriented approach for modeling information at a
changing
organizations (redesign/re-
Object-Role Modeling
conceptual level. An overview of ORM is given in Halpin (1998a), and a detailed treatment
in Halpin (2001a). ORM includes a family of closely related variants, including Natural
Information Analysis Method (NIAM) (Wintraecken, 1990), Natural Object Relationship
Method (NORM) (De Troyer & Meersman, 1995), Predicator Set Model (PSM) (ter Hofstede,
Proper & van der Weide, 1993), and Fully Communication Oriented Information Model-
ing (FCO-IM) (Bakema, Zwart & van der Lek, 1994). Unlike Entity-Relationship (ER)
modeling (Chen, 1976) and the class diagram technique of the Unifi ed Modeling Language
(UML) (OMG UML RTF, 2003), ORM makes no use of attributes as a base construct, in-
stead expressing all fact types as relationships. This attribute-free approach leads to greater
semantic stability in conceptual models and conceptual queries (Bloesch & Halpin, 1997;
Halpin, 1998b) and enables ORM fact structures to be directly verbalized and populated
using natural language sentences.
ORM supports mixfi x predicates of any arity (unary, binary, ternary, etc.), so its con-
straints and derivation rules can also be directly verbalized in sentential form. For details
on business fact and rule verbalization in ORM, see the series of articles initiated by Halpin
(2003). Moreover, ORM's graphic constraint notation is far more expressive than that of
UML class diagrams or industrial ER versions. ORM is now supported by a number of
modeling tools, which can automatically transform ORM schemas into physical database
schemas (e.g., see Halpin, Evans, Hallock & MacLean, 2003). For such reasons, ORM is
being increasingly used for conceptual analysis of information, as well as ontology specifi -
cation (Spyns, Meersman & Jarrar, 2002), and is currently being considered as a candidate
for a standard business rule modeling language within the Object Management Group.
Both DEMO and ORM treat fact types as fundamental, and require that their models
be expressible in natural language sentences. This suggests that the approaches may be
synthesized in a natural way, resulting in a more powerful method for business modeling.
This chapter discusses the fi rst attempts to explore the feasibility of this synthesis, and
identifi es some lessons learned, using a running example of a library application to illustrate
the main ideas.
The following section summarizes the essential concepts and model types underlying
the DEMO approach, and discusses how the library application is modeled using DEMO.
Next, the chapter explains the main concepts and notations of ORM, and shows how the
