Data Modeling: An Ontological Perspective of Pointers

Data Modeling: An Ontological Perspective of Pointers

Hock Chuan Chan (National University of Singapore, Singapore), Chuan-Hoo Tan (College of Business, City University of Hong Kong, Hong Kong) and Hock-Hai Teo (National University of Singapore, Singapore)
Copyright: © 2014 |Pages: 21
DOI: 10.4018/JDM.2014100102
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Research on ontology and conceptual data modeling builds on the premise that the world could be represented in terms of concepts (also known as “things”), their properties and their (things') relationships. A persistent effort in the representation is how to distinguish things from their properties, as well as how to denote relationships that relate two things together. Following the tradition of deriving guidelines from ontology studies, this paper proposes a need to make a clear distinction between relationships among things and relationships among things' properties. The central thrust of this paper is in proposing an ontologically guided principle: For the same relationship, working with the relationship among things will lead to better user performance than working with the relationship among things' properties. This principle, called Relationship without Pointers principle, is robustly validated by re-analyzing a set of experiment data on user data modeling performance with three database models. This principle may be applicable to other contexts that study relationships.
Article Preview

Introduction

During the system design of an Information System (IS), accurate construction of the diagrammatic representation of the real world (or studied situation/scenario) is important (Bera et al. 2010; Martin 2007; Johnson-Laird 1983). Specifically, an accurately captured representation should connote the essential semantic meanings of the focal domain, i.e., the concepts (or “things”), their properties and their interrelationships. While the diagrammatic representation, also known as conceptual data modeling, constitutes a small subset of the entire system analysis, design and development process, it plays a non-trivial role in determining the success of an IS development (Moody 2005). The widespread recognition of the importance of capturing the focal domain’s semantic meanings has led to an enduring research effort expanded towards proposing, deriving and empirically validating the related theorems (Recker et al. 2011; Recker and Rosemann 2010; Siau and Rossi 2011). These theorems span from the general research of Bunge’s ontological theory (Bunge 1977; Bunge 1979) to specific conceptual data modeling of the Bunge-Wand-Weber (BWW) model (Wand and Weber 2004, 2002, 1990).

While much of these extant theories has significantly propelled the field towards framing the overarching data modeling guidelines and principles (e.g., Tillquist et al. 2002), research has shown that there remains many unclear areas (Goldstein and Storey 1990). A core area of investigation is about things and their properties. A thing is a general term for any abstract or concrete thing that possesses properties. For instance, a person, a car, and a country are all things; a person has properties such as name and height. Recent research along this direction include Shanks and his colleagues (2008) who examine the relationship between things and the part-whole relations among them; Bodart and his colleagues (2001) question whether including optional properties would make a difference; and Gemino and Wand (2005) assess the impact of different forms of properties on conceptual data modeling complexity and clarity.

A research question along this field of exploration, yet not answered, is whether to model relationship among things or relationship among things’ properties, in the context of pointer usage. A pointer provides an implied linkage between two things, or two properties. To illustrate, we refer to two instances of one using the basic entity diagrammatic and another in relational modeling. In the entity diagrammatic, a line connecting two entities in a diagram implies a relationship between the entities. Similarly, in the relational modeling context, the same values in a primary key-foreign key pair of properties imply a relationship between the properties and consequently their relations. This inquiry is timely with the expanded interest in amalgamating object-oriented and component-based thinking (Frakes and Kang 2005) into conceptual data modeling (Evermann 2005a, 2005b). While this study focuses on one aspect of conceptual modeling practice, namely the use of pointers in data modeling, we hope through the testing of the pointer notion using three data models (to be described shortly) we are able to offer a convincing, empirically driven argument.

Our research motivation is entrenched in the dilemma entailed in the application of pointers. Specifically, the notion of pointers is deeply rooted in the object-oriented metaphor and relational system development approach. However, pointers, as will be shown subsequently, link through two things’ properties rather than to follow the philosophical wisdom of relating two things directly. Our view of this dilemma is twofold. First, people should make a distinction between relationship among things and relationship among properties of things. Second, the abstraction of a relationship through linking two properties, i.e., using the pointer notion, could violate the fundamental, semantic understanding of two related concepts. These two claims form the core thesis of this study, which we called the Relationship without Pointers (RWP) principle.

Complete Article List

Search this Journal:
Reset
Open Access Articles
Volume 28: 4 Issues (2017): 3 Released, 1 Forthcoming
Volume 27: 4 Issues (2016)
Volume 26: 4 Issues (2015)
Volume 25: 4 Issues (2014)
Volume 24: 4 Issues (2013)
Volume 23: 4 Issues (2012)
Volume 22: 4 Issues (2011)
Volume 21: 4 Issues (2010)
Volume 20: 4 Issues (2009)
Volume 19: 4 Issues (2008)
Volume 18: 4 Issues (2007)
Volume 17: 4 Issues (2006)
Volume 16: 4 Issues (2005)
Volume 15: 4 Issues (2004)
Volume 14: 4 Issues (2003)
Volume 13: 4 Issues (2002)
Volume 12: 4 Issues (2001)
Volume 11: 4 Issues (2000)
Volume 10: 4 Issues (1999)
Volume 9: 4 Issues (1998)
Volume 8: 4 Issues (1997)
Volume 7: 4 Issues (1996)
Volume 6: 4 Issues (1995)
Volume 5: 4 Issues (1994)
Volume 4: 4 Issues (1993)
Volume 3: 4 Issues (1992)
Volume 2: 4 Issues (1991)
Volume 1: 2 Issues (1990)
View Complete Journal Contents Listing