Making Sense of Design Patterns
Rinke Hoekstra1,2 and Joost Breuker1
1 Leibniz Center for Law
Faculty of Law
University of Amsterdam
{hoekstra,breuker}@uva.nl
2 Department of Computer Science
Faculty of Exact Sciences
VU University Amsterdam
hoekstra@few.vu.nl
Abstract. This paper discusses the way in which design patterns may
improve the current practice of ontology engineering. It presents five re-
quirements that go beyond the current state of the art of collecting and
curating design patterns. We build on the thesis outlined in [17] that de-
sign patterns should be one of several possible outcomes of a fundamental
design decision. We emphasise their relation to structures in cognition
rather than domain dependence. This to improve our understanding of
what ontology design patterns are, and how they relate to (modelling)
expertise. We provide a definition of structural design patterns, give a
number of examples, and discuss further work.
1 Introduction
An ontology is not just any terminological knowledge base as it embodies a
specific definitional perspective. Terminological knowledge representations are
based on Minsky’s notion of a frame [23] – concepts are defined by context.
An ontology refines this notion and needs to distinguish between the inherent
and accidental properties of concepts [8,16,3]. For instance, although a typical
property such as ‘position’ is clearly relevant from a knowledge engineering per-
spective, ontologically speaking it is usually a side issue: changing the position
of an object does not make it intrinsically different.
This epistemological promiscuity of terminological knowledge representation
has led to the formulation of several design principles (cf. [17] for an overview).
Over the years several design principles have been cast in ontology engineering
methodologies to improve the quality of ontology development. Also, the wide-
spread availability of ontologies on the web has opened the door to the adoption
of pre-existing general ontologies. By providing an initial structure and a set of
basic concepts and relations, these ontologies can be a valuable jump start for
more specific ontology development.
A more recent development is the identification of design patterns that are
meant to overcome some of the limitations in the use of existing ontologies and
P. Cimiano and H.S. Pinto (Eds.): EKAW 2010, LNAI 6317, pp. 331–340, 2010.
c
© Springer-Verlag Berlin Heidelberg 2010

332 R. Hoekstra and J. Breuker
design principles. In this paper we explore a set of requirements for ontology
design patterns, which emphasise their relation to structures in cognition rather
than domain dependence.
Related Work. Ontologies are frequently quite large and heavyweight; they are
hard to mould and extend to a usable domain ontology. Design principles can be
quite abstract and difficult to translate into concrete definitions. Design patterns
offer a middle ground between the two, where the methodological principles are
made concrete in manageable building blocks [13,27, a.o.]. Furthermore, these
patterns can assist in tackling problems related to the ontology representation
language of choice [18, a.o.]. For instance, the OPPL language [20] functions as
a macro language for authoring ontologies at pattern level.
In the context of ontology development, work on Ontology Design Patterns
(ODPs) is based on the notion of knowledge pattern [10].1 Ontology design pat-
terns are typically categorised as either logical of content patterns [13,27], where
logical patterns do not introduce domain dependence. The ODP portal is a
community effort to build a catalog of high quality ontology design patterns2
Patterns undergo a peer-review based on a set of criteria before they become
‘certified’. These criteria are based on [12] and include the availability of an OWL
representation and coverage of requirements in terms of competency questions:
a pattern is a solution to one or more use cases. The content pattern should be
a small, autonomous ontology, that covers a core of cognitively relevant compo-
nents, allows some form of reasoning and reflects best practices.3
An initial evaluation of content pattern-based ontology construction [2] shows
that participants feel that patterns help, and that using design patterns results
in ‘better’ ontologies. Unfortunately this is only preliminary work, most subjects
were not very experienced knowledge engineers, and the set of available design
patterns was limited to 22. Also, the evaluation concerned the ontology engi-
neering process rather than the design patterns themselves, making it hard to
assess how the quality of these patterns contributed to the evaluation results.
The cognitive perspective of [13] has received little attention in practice: pat-
terns are mainly considered to be engineering artefacts, and essentially form a
collection of best practices. Indeed this is very useful, but in this paper we aim to
delve somewhat deeper and improve our understanding of what ontology design
patterns are, and how they relate to (modelling) expertise.
In light of this exercise we explore several requirements that support this
perspective. Admittedly these requirements differ from the criteria of the ODP
portal as they are not aimed to optimise categorisation and publication in a
catalog. We build on the thesis outlined in [17] that design patterns are language
1 We will use the terms ‘ODP’, d‘esign pattern’ and ‘pattern’ interchangeably through-
out this paper.
2 See http://www.ontologydesignpatterns.org
3 See http://ontologydesignpatterns.org/wiki/
index.php?title=Odp:EvaluationPrinciples&oldid=1714. These criteria apply
only to content patterns ([27], and 2.3).