Establishing and Specifying the REA model as a Domain Ontology for Business

Original specification of the REA Ontology – McCarthy, Geerts et al at University of Michigan

Transitional efforts to formalize REA Ontology using UML Diagrams – Poels, Paemeleire, Gailly et al at Ghent University

Formalization of the REA Enterprise Ontology using OWL and the UML Profile of OWL – Poels, Gailly et al at Ghent University

Background

The REA Ontology is an “event ontology” (Allen, GN and ST March, 2006) that focuses on events occurring within the realm of a company, their participating agents, affected resources, and regulating policies. REA can be used as a reference for modeling a single business cycle (e.g. sales-collection) or a chain of business cycles, connected through resource flows. Applications supported by REA-driven modeling include the design of accounting and operations management systems, auditing and internal control, and conceptual data modeling. REA has been used in a number of international standardization efforts for e-collaboration systems. For instance, REA was the basis for the business process ontology in the UMM business process and information model construction methodology, the ECIMF system interoperability enabling methodology, and the Open-EDI business transaction ontology which is part of the ISO/IEC 15944-4 standard. REA has further been proposed as a theoretical basis for the reference models that underlie ERP systems.

The Project

Frederik Gailly and Geert Poels (Ghent University) set out with two goals:

  1. To establish the REA ontology as a domain ontology – specifically a domain ontology for business, using a semiotic framework (Stamper et al, 2000; Burton-Jones et al , 2005). Within this framework, the authors sought assistance with their evaluation of the ontology’s syntactic quality specifically from Welty et al (1999) and Lassila and McGuiness (2001), and with their evaluation of the ontology’s social quality from Guarino (1998) and Uschold and Jasper (1999).
  2. To enhance the specification of the REA ontology using both a formal language representation (the Web Ontology Language or OWL) and a graphical representation (the Unified Modeling Language or UML Profile for OWL) .

We will leave it largely up to the readers to familiarize themselves with the details of how the authors achieved their goal of establishing the REA ontology as a domain ontology for business. We simply present the semiotic framework (Burton-Jones et al , 2005) in Table 1 and highlight the authors’ evaluation of the REA ontology against this framework in Table 2. Finally, we note that the authors define business as “the activity of providing goods and services involving financial, commercial and industrial aspects”.

Metric Attribute Description
Syntactic quality Lawfulness Correctness of syntax
Richness Breadth of syntax used
Semantic quality Interpretability Meaningfulness of terms
Consistency Consistency of meaning of terms
Clarity Average number of word senses
Pragmatic quality Comprehensiveness Number of classes and properties
Accuracy Accuracy of information
Relevance Relevance of information for a task
Social quality Authority Extent to which other ontologies rely on it
History Number of times ontology has been used

Table 1. Metrics  for evaluating ontologies (Burton-Jones, A et al, 2005)

Metric Attribute REA evaluation
Syntactic quality Lawfulness
  • The REA ontology has a semantically rich internal structure (say) that is not limited to the IS-A relations as found in a typical thesaurus, and that defines axioms that allow for semantic reasoning.
  • However, many details of the REA ontology internal structure and underlying logic are not explicitly specified (e.g., disjointness of concepts).
Richness
Semantic quality Interpretability
  • The authors don’t evaluate the semantic quality of the REA ontology because of their focus on the applicability of the ontology.
Consistency
Clarity
Pragmatic and
Social quality
Comprehensiveness
  • Both the pragmatic and social qualities of the REA ontology are evaluated largely in terms of the ontology’s applications and the extent to which the ontology is useful for users and their agents, irrespective of syntax and semantics.
  • The REA ontology supports a wide range of applications and users:
    • education in accounting and business
    • model-driven systems and software design
    • supply chain and e-collaboration
    • knowledge representation and retrieval
Accuracy
Relevance
Social quality Authority
History

Table 2. Evaluation of the REA ontology.

The authors’ evaluation of the syntactic quality of the REA ontology as a domain ontology for business highlighted the following deficiencies:

  • the REA ontology was a somewhat incoherent mix of textual and graphical elements that spoke to its relative immaturity
  • while some researchers (Bialecki 2001; Chou 2006; Geerts 2004) had taken steps to represent the REA ontology in a machine-readable format, none of these formalizations was widely known or generally accepted
  • like Geerts and McCarthy (1999), the authors recognized the importance of better specifying the REA ontology

Again, we will leave it largely up to the readers to familiar themselves with the details of how the authors enhance the specification of the REA ontology. The main principles of their approach included:

  • the authors apply formal procedures – the METHONTOLOGY framework (Fernandez-Lopez et al. 1997; G6mez-Perez and Rojas 1999) and the Operational Data Model (OMG, 2006) – to re-engineer the existing REA ontology
  • Three main activities are identified in this re-engineering process (Figures 2 and 3):
    • reverse engineering – the conceptualization of the REA ontology is recovered from whatever representation formats are available (e.g. text, tables, modeling diagrams)
    • re-structuring – the recovered conceptualization of the REA ontology is re-designed and expressed in a graphical representation with well-defined syntax and semantics (like the UML Profile for OWL) 1
    • forward engineering – the re-designed conceptualization is transformed into a re-engineered, formal language representation (like OWL)

 

A Business Domain Ontology Re-engineering Methodology - Process, Activities and Artifacts

Figure 2. A Business Domain Ontology Re-engineering Methodology – Process, Activities and Artifacts.

 

REA Ontology Transformation and Tools

Figure 3. The recovery, re-design, and formalization of the REA domain ontology for business.

With this overview of the methodology used by Geerts, Laurier and Poels to re-engineer the REA ontology in 2007 – 2008, we can turn to the resulting graphical representation (UML Profile of OWL) and formal language representation (OWL) of the REA domain ontology for business.

Next: Graphical representation of the REA domain ontology for business (UML Profile for OWL)

 

 

 

  1. Using UML in the re-design of the REA ontology capitalizes on the familiarity of many end-users  in the business domain with this technology.