Handbook of Software Engineering and Knowledge Engineering
f
c World Scientic Publishing Company
EXPLORING ONTOLOGIES
YANNIS KALFOGLOU
School of Articial Intelligence, University of Edinburgh, 80 South Bridge
Edinburgh EH1 1HN, Scotland
Email: yannisk@dai.ed.ac.uk
Received June 1, 2000
Revised September 27, 2000
Accepted (accepted date)
Abstract
Ontologies are studied by many scholars with diverse backgrounds and are applied in
a variety of contexts and application areas. Despite the numerous reviews published there
are many issues which still remain unclear with respect to their cost-eective deployment,
identication of tradeos, maintenance strategies and ways of integration. Furthermore,
there is no report which refers to all those issues along with the basic information in
order to be used as a road-map. This survey article aims to provide such information
in a manner which will help the interested software practitioner to comprehend the basic
principles in ontology design, understand their strengths and weaknesses, be aware of a
variety of areas where ontologies have been successfully applied, and identify tradeos and
potential solutions.
Keywords: Ontologies, knowledge sharing and reuse, knowledge management, software
design.
1. Introduction
In this survey article we are exploring ontologies with emphasis on their deploy-
ment in a wide variety of areas ranging from software design to knowledge man-
agement and information retrieval. We are not interested to provide an in-depth
analysis of the eld of ontologies in an isolated manner nor to provide a method-
ological approach for guiding the software design processes. Rather, we scrutinise
the ontologies eld as practised, mainly, in the knowledge engineering community
over the last decade, and report on their impact in software design through example
application cases, worked projects, and emerging experimental results.
Before we proceed with our survey, however, we look closer at software design.
Software design is still a young eld, and we are far from having a clear articulation
of the relevant principles. Winograd oers a parallelism of the phrase with software
engineering(hereafter, SE): \. . . is often used to characterise the discipline that is
also called software engineering - the discipline concerned with the construction of
software that is eÆcient, reliable, robust, and easy to maintain."[108]. Although
work has begun in engineering software design with the emergence of methodological
approaches[78], guidelines, and bringing in rationale[68], there is still an area that
remains unexplored: bringing into the design process explicit knowledge regarding
the domain on which the system to be developed will operate.
1

2 Handbook of Software Engineering and Knowledge Engineering
The study and modelling of that knowledge is a core theme in articial intel-
ligence(hereafter, AI) research. Having their roots in knowledge representation,
knowledge engineering methods and techniques gave AI researchers a powerful tool
for transforming contextual knowledge into machine-readable form to enable mech-
anised reasoning about a domain of interest. Ontologies are such a form of domain
knowledge. We should also note the similarity of this topic with such areas as do-
main analysis[4] and engineering(this volume, J.L.Diaz-Herrera, chapter ??), object
oriented patterns[30], etc. In this chapter, however, we are interested in ontologies
as practiced, mainly, in the AI community.
Nowadays, they are studied by many scholars who belong to dierent communi-
ties. Hence, a plethora of articles and eld reviews are available for the interested
practitioner. However, each of them is focussed on a specic area: for example, the
Uschold and Gruninger review - one of the rst comprehensive reviews published -
is concerned with design principles and methodological ways of construction with
selective references to exemplar applications[93]; the Fridman-Noy and Hafner re-
view further explores and compares design methods[29], the Chandrasekaran and
colleagues article provides a general overview of the eld[13]; and Gomez-Perez and
Benjamins devote half of their report to give a catalogue-style information on on-
tologies research[34]. Despite the bulk of resources available it is still diÆcult for
practitioners, especially those with a SE background, to locate and elicit the right
information with respect to engineering, application, and cost-eective issues. Most
of the times this information is found in dierent resources.
This survey aims to ll-in this gap. In order to do this eectively, we explore
the eld from the following angles: design, deployment, and tradeos. We explore
design issues in sections 2 to 6 where we describe what an ontology stands for, ways
of design, the role of ontological commitment, methodologies to follow when building
ontologies, and explain the various types reported in the literature. Deployment
issues are described in sections 7 to 8 with emphasis on uses of ontologies, ways
of deployment, and references to applications and in
uential projects from both
industry and academia. Lastly, we conclude our survey by discussing potential
problems, tradeos and solutions proposed and used, in section 9, followed by list
of pointers to resources for further reading, in section 10.
2. Denitions
We start our review by explaining what an ontology stands for. Although a single
denition will usually suÆce, ontologies have a peculiar characteristic: there are a
number of dierent denitions proposed and used. Even nowadays there are people
who argue about the actual meaning of the term. A reason for this is, probably,
the fact that ontologies are studied, developed and applied by people with diverse
backgrounds and interests. We do not subscribe to this pointless debate over the
meaning of the term in this article nor we will introduce yet another denition.
Rather, we brie
y review the most commonly used denitions found in the literature
in order to explain what an ontology stands for.

Exploring ontologies 3
One of the early denitions appeared in [72]. The authors dene an ontology as:
\the basic terms and relations comprising the vocabulary of a topic area as well as
the rules for combining terms and relations to dene extensions to the vocabulary".
This denition introduced the idea that ontologies can be viewed linguistically,
as extensible vocabularies regarding a topic area. In the context of knowledge
sharing, Gruber oered a short denition which became the most widely cited in the
literature: \an ontology is an explicit specication of a conceptualisation"[37]. This
denition was further enriched by Borst and his colleagues in [10], where they argued
that the specication is actually formal and the conceptualisation is shared. Studer
and colleagues analysed the terms used in the denition and provide the following
explanation: \Conceptualisation refers to an abstract model of some phenomenon in
the world by having identied the relevant concepts of that phenomenon. Explicit
means that the type of concepts used, and the constraints on their use are explicitly
dened. Formal refers to the fact that the ontology should be machine-readable.
Shared refers to the notion that an ontology captures consensual knowledge, that is,
it is not primitive to some individual, but accepted by a group"[85]. Uschold oers
a working denition which hints at the purpose of having ontologies: \An ontology
is virtually always the manifestation of a shared understanding of a domain that
is agreed between a number of agents. Such agreement facilitates accurate and
eective communication of meaning, which in turn leads to other benets such as
inter-operability, reuse and sharing"[90]. Others, consider ontologies as domain
theories[25], as vocabularies[83], as standards[58], etc. In [43] the authors oer a
clarication of terminological issues regarding the various denitions founded in
the literature. In [88], the authors relate an ontology to a knowledge base in their
denition: \An ontology provides the basic structure or armature around which a
knowledge base can be built".
Based on the denitions quoted above we summarise what an ontology stands
for: an explicit representation of a shared understanding of the important concepts
in some domain of interest. The role of an ontology is to support knowledge sharing
and reuse within and among groups of agents(people, software programs, or both).
In their computational form, ontologies are often comprised by denitions of terms
organised in an hierarchy lattice along with a set of relationships that hold among
these denitions. These constructs collectively impose a structure on the domain
being represented and constrain the possible interpretations of terms.
3. Design principles
A number of design criteria have been proposed, originally analysed in [38]. For a
thorough analysis of these criteria we point the interested reader to the aforemen-
tioned citation whereas here we brie
y recapitulate them. The criteria proposed
are: clarity, coherence, extendibility, minimal encoding bias, and minimal ontologi-
cal commitment. Clarity means that the intended meaning should be communicated
eectively. This means that ambiguity should be minimised, distinctions should be
motivated, and examples should be given to help the reader understand denitions

4 Handbook of Software Engineering and Knowledge Engineering
that lack necessary and suÆcient conditions. When a denition can be stated in log-
ical axioms, it should be. Where possible, a complete denition(a predicate dened
by necessary and suÆcient conditions) is preferred over a partial denition(dened
by only necessary or suÆcient conditions). All denitions should be documented
with natural language. Coherence means that the ontology should be internally
consistent. At the least, the dening axioms should be logically consistent. Coher-
ence should also apply to the concepts that are dened informally, such as those
described in natural language documentation and examples. Extendibility means
that one should be able to extend the existing terms in a way that does not require
the revision of existing denitions. The next two criteria help to achieve that. The
encoding bias and ontological commitment should be minimal. An encoding bias
results when representation choices are made purely for the convenience of notation
or implementation. Encoding bias should be minimised, because knowledge-sharing
agents may be implemented in dierent representation systems and styles of repre-
sentation. Minimal ontological commitment means that the ontology should make
as few claims as possible about the world being modelled, allowing the parties com-
mitted to the ontology freedom to specialise and instantiate the ontology as needed.
While making too many ontological commitments can limit extensibility, making
too few can result in the ontology being consistent with incorrect or unintended
worlds(i.e., models). For this reason, it is benecial to make ontological commit-
ments with respect to aspects intrinsic to a domain.
We should note that the above criteria are not always possible to meet by ontol-
ogy designers. A number of tradeos have been identied[38], and ways of compro-
mising between well designed ontologies and applicability have been investigated
[10]. We will not expand on this issue in this article because it is peripheral to our
topic: uses of ontologies. To support this we shift our attention to the notion of on-
tological commitment which plays an important role in using ontologies in software
systems.
4. Ontological commitment
Ontological commitment refers to agreement on the use of the shared vocabulary
by the agents commited to the ontology in question. When we say that an agent
commits to an ontology we mean that its observable actions are consistent with
the denitions in the ontology[38]. It has been said that commitment to a common
ontology is a guarantee for consistency but not for completeness, with respect to
queries and assertions using the vocabulary dened in the ontology[38].
Guarino describes the role of ontological commitment in software: \ontological
commitment should be made explicit when applying the ontology in order to facili-
tate its accessibility, maintainability, and integrity. This will lead to an increase of
transparency for the application software which based on that ontology"[40]. These
commitments are often encoded as axioms that enforce the syntactic consistency
of the denitions used. Practically, an ontological commitment is an agreement to
use a vocabulary(i.e., ask queries and make assertions) in a way that is consistent

Exploring ontologies 5
with respect to the theory that species the ontology. We build agents that commit
to ontologies and we design ontologies so we can share knowledge with and among
these agents. With a declarative specication, we can explicitly reason about dier-
ent ontological commitments. For example, we can compare two dierent proposals
for an ontology with respect to the classes of objects that they require and the
properties and relations among these objects that they postulate[93].
Guarino and colleagues argue for a greater role of ontological commitment. In
[42], the authors continue, an ontological commitment should capture and constrain
a set of conceptualisations. They propose a formalisation of ontological commit-
ments which: \oers a way to specify the intended meaning of [a logical language]
vocabulary by constraining the set of its models, giving explicit information about
the intended nature of the modelling primitives used and their a priori relation-
ships". The work of Guarino and colleagues is focussed on the design phases of
ontology. Other scholars' work aim on the deployment of ontological commitments
in applications. We will describe this work in section 8 where we summarise uses
of ontologies in software design.
5. Methodologies
The construction of an ontology is a time-consuming and complex task. Although
there are no standards to obey when building an ontology, various design guide-
lines and methodological approaches have been proposed and used. In particular,
in a comprehensive review of the eld[93] the authors report on two methodolo-
gies used in the context of the Enterprise ontology[97] and the TOVE project[20].
In the former, a skeletal methodology has been proposed[98] which identies ve
main steps: (a)identify purpose and scope, (b)build the ontology, (c)evaluation,
(d)documentation, and (e)guidelines for each phase. Step (b) is further divided
into ontology capture, coding, and integration of existing ontologies. This skele-
tal methodology was used in the construction of the Enterprise ontology but does
not explicitly deploy a formal evaluation procedure. This was the main focus of the
methodology used in the context of the TOVE project[20]. In particular, Gruninger
and Fox used a formal methodology that supported evaluation of the ontology using
the notion of competency questions[39]. The underlying philosophy is to dene a set
of queries that the ontology can answer. These queries help to assess the ontology's
competence. They evaluate the expressiveness of the ontology which is required to
represent these questions and characterise their solutions. These queries are drawn
from a number of motivating scenarios which are story problems or examples which
are not adequately addressed by existing ontologies.
Apart from the work on evaluation and construction methodologies by Uschold,
Gruninger and colleagues, others have focussed on the preliminary phases of con-
struction. In [24] the authors presented a system, called METHONTOLOGY, which
provides support for the entire life-cycle of ontology development. A distinguishing
characteristic of the METHONTOLOGY framework is that it is tailored to support
the early phases of development by employing the notion of intermediate represen-

6 Handbook of Software Engineering and Knowledge Engineering
tations. These are representations independent of the implementation language in
which the ontology will be developed. The system that support the use of these
representations is the Ontology Development Environment(ODE)[9]. An overview
of methodologies used in AI projects along with a comparison with standards from
SE literature is given in [23].
6. Types
The development methodologies reported above were used in some of the ontologies
which will be described in sections 7 and 8. Before we proceed to survey actual
implementations of ontologies we describe various types of them as reported in
the literature. Ontologies can be classied in terms of genericity. For example,
broad ontologies like CYC[57], model generic notions that forms the foundations
for knowledge representation across various domains. These are also called top-level
ontologies[13], like Sowa's ontology[84]. On the other hand, small-scale, domain-
specic ontologies are carefully tailored to the domain at question. Examples of this
type are the PhysSys ontology[10] which captures knowledge regarding physical
system processes, the AIRCRAFT ontology [100] used to represent air-campaign
planning knowledge, the PIF ontology[55] used for business process modelling, etc.
Another classication of ontologies is concerned with their purpose. There exist
task ontologies[67] that capture task-related knowledge independently of the domain
that the task is dened. Complementary to these are the method ontologies[12]
which provide denitions of the relevant concepts and relations used to specify a
reasoning process to achieve a particular task. A specic type of ontologies is the
knowledge representation ontologies. The most representative example is the Frame
ontology[37] which captures the representation primitives used in frame-based lan-
guages. It allows other ontologies to be specied using frame-based conventions, as
implemented by the Knowledge Interchange Format(KIF)[33].
Most ontologies, however, are placed under the tag domain ontology. These
are designed to support a specic domain and applications dened within that
domain. For example, the PIF ontology is concerned with the business process
modelling domain and supports the exchange of information among a variety of
business process modelling applications.
There is another type of ontology, the linguistic ontologies. The most illus-
trative examples are the Generalised Upper Model(GUM)[7], WordNet [66], and
SENSUS[54]. However, these usually have the form of a vast collection of terms
which led to another classication with regard to the level of formality. These sort
of ontologies are often called \terminological" ontologies whereas ontologies like
TOVE are called \axiomatised" ontologies.
In their overview of the eld, Uschold and Gruninger identied the following
types with respect to the degree of formality: highly informal, semi-informal, semi-
formal, rigorously formal[93]. In the informal cluster we see denitions in natural
language or at most in a structured form of natural language. In the formal cluster
we have ontologies implemented in an articial formal language(i.e., Ontolingua), or

Exploring ontologies 7
in rst order theories with formal semantics, theorems and proofs of such properties
as soundness and completeness(i.e., TOVE).
7. Engineering
Although many argue that engineering of ontologies is still in its infancy the rst
comprehensive reports covering all aspects of ontology construction and deployment
began to emerge few years ago. We selectively report here some of these eorts by
highlighting their contributions to the eld. In an experiment of ontology reuse[92],
researchers working at Boeing were investigating the potential of using an existing
ontology for the purpose of specifying and formally developing software for aircraft
design. The application problem addressed was to enhance the functionality of a
software component used to design the layout of an aircraft stiened panel. They
describe a start-to-nish process that used an existing ontology, residing on the
Ontolingua [21] server, the EngMath[36] ontology, which was then translated to the
target specication language and integrated to an engineering software component.
They then executed that component and demonstrated the benets of reusing an
existing knowledge component in the development process. The lessons learned
from that experience is that ontology reuse can be pursued on a large scale and,
under certain circumstances, it can be a cost-eective approach. We will revisit the
tradeos identied by Uschold and colleagues in their experiment in section 9 while
we continue here by reporting two studies that were focussed on the whole spectrum
of engineering ontologies: the AIRCRAFT project, and the PhysSys project.
In [100] the authors describe how they achieved reuse among ontologies them-
selves. The resulted ontology, AIRCRAFT
a
, contains knowledge about types of
US military aircraft, including data about the engines, PODs, and fuel tanks that
these aircraft can carry. The distinguishable feature of this ontology is how it has
been developed in the rst place. The process, which is described in [87], was based
on the use of a large-scale, linguistic ontology, the SENSUS[54]. A characteristic of
SENSUS is that it is actually constructed from extracting and merging information
from existing electronic resources(like the WordNet, dictionaries, GUM ontology).
The authors, used this broad coverage ontology and then devised a semi-automatic
method which made it possible to identify terms in the original ontology that were
relevant to their particular domain, and then pruned the ontology so that it in-
cluded only those terms. In addition, they enhanced the newly emerged ontology
with terms tailored to the domain of air campaign planning. These were military
terms. The resulting ontology, AIRCRAFT, is accessible through an ontology devel-
opment environment, the ontosaurus browser which supports the idea of \ontology
developed collaboratively by the system developers themselves"([87]).
In [10], a general and formal ontology, called PhysSys is presented. It covers the
domain of dynamic physical systems and it is composed of seven dierent ontologies.
This work explored a new idea in ontology engineering, that is ontology projections:
a
A demonstration version is electronically available from the URL:
http://www.isi.edu/isd/ontosaurus.html

8 Handbook of Software Engineering and Knowledge Engineering
\a
exible mechanism to link and congure ontologies into larger ones." Three
kinds of projections demonstrated in the paper, include-and-extend, include-and-
specialise, and include-and-project. The latter was used to link an ontology devel-
oped by the group of PhysSys authors to an outsourced ontology, the EngMath. The
PhysSys ontology was used as the foundation for the conceptual database schema
of a library of reusable engineering model components, the OLMECO library. The
library was evaluated by modelling and numerically simulating the existing heating
system of a general hospital in Schiedan, the Netherlands[10].
In the context of the Plinious project[101], the bottom-up method in ontology
development is discussed[102]. In contrast with the majority of approaches in ontol-
ogy construction which fall into two categories, top-down and middle-out(analysed
in [93]), the bottom-up way \proposes to lay down the meaning of complex concepts
by means of primitive meaning constituents." It has been applied to the domain of
ceramic materials and covers their properties and the processes to make them. It
was found that this approach was suitable for such a domain because, the authors
continue, it is impossible to exhaustively predict in advance which concepts will be
needed to express the knowledge found in the texts. As this domain covers chemical
substances, it was argued that listing all these substances is an open-ended task.
As such, keeping track of the regular updates in a top-down designed ontology was
impractical since it requires substantial eort and is error-prone. Consequently, the
approach used supports reasoning along two orthogonal hierarchies: \the parton-
omy formed by substances and their constituents and the taxonomy formed by
concepts and superconcepts"[102].
Other projects which provide an insight in the engineering process are the re-
engineering eort of implemented ontologies, described in [35], and the collaborative
eort in developing a common ontology for the knowledge acquisition community[8].
In particular, Gomez-Perez and Amaya describe a re-engineering process of retriev-
ing and transforming a conceptual model of an existing ontology into a new one.
The work of Benjamins and Fensel describes the Knowledge Annotation Initiative of
the Knowledge Acquisition Community ontology(in short, KA
2
), which models the
knowledge acquisition community and forms the basis to annotate its documents
on the Web
b
in order to enable intelligent access.
8. Applications and projects
A complete listing of applications of ontologies is impossible. The literature refer-
ences are huge and citing lengthy lists is not practical. However, we provide pointers
to various resources in section 10 whereas here we selectively report the most repre-
sentative ones. To do this eectively we cluster them according to their application
domain.
We start with the area of enterprise modelling. In this area we found the Enter-
prise ontology[97], which captures the organisational structure of an enterprise with
b
The ontology is accessible online from the following URL:
http://www.aifb.uni-karlsruhe.de/WBS/broker/KA2.html

Exploring ontologies 9
emphasis to activities and processes. The ontology is developed in a structured text
form and a translation in Ontolingua is also available. In the same line is the TOVE
ontologies set[20] which shares the same aims with Enterprise, but has been devel-
oped in a formal computational form and uses dierent underlying principles[28].
The dierences between these two representative ontologies for enterprise modelling
are highlighted in [93].
viewer
procedure
translator
translator
ontology
library
method
1: give me the procedure for ...
2: procedure = ?
3: procedure =
process
4: give me the process for ...
5: ? = process
6: METHOD =
process
7: give ne the METHOD for ...
8: here is the
METHOD for ...
9: here is the
process for ...
procedure for ...
10: here is the
Figure 1: Ontology as inter-lingua: Example taken directly from [93]. This illus-
trates the use of an ontology as an inter-lingua to integrate dierent software tools.
The term procedure, used by one tool is translated into the term, method used by
the other via the ontology, whose term for the same underlying concept is process.
A relevant application area is that of business process modelling. The Process
Interchange Format(PIF)[55] is among the best known in this area. The aim of
PIF is to develop an interchange format to help automatically exchange process
descriptions among a variety of business modelling and support systems such as
work
ow software,
ow charting tools, planners, process simulation systems and
process repositories. The core of PIF consists of the minimal sets of constructs
necessary to translate simple but non-trivial process descriptions. In addition, PIF
can be extended to represent local needs of individual groups with the use of Par-

10 Handbook of Software Engineering and Knowledge Engineering
tially Shared Views(PSV) described in [56]. The PIF framework has been applied
in a supply chain scenario[77] which was adopted from the Work
ow Management
Coalition(WfMC)[107]. An example of an interchange format is illustrated in gure
1.
Ontologies have also been applied to medical applications. For example, a
methodology for integrating medical terminologies was presented in [31]. This is
the aim of the ONIONS methodology[32] developed by the same group. In the same
context, the European project GALEN
c
[80] which aims at capturing information
from the clinical domain. In [1], the authors present a system, called Sophia which
acts as a knowledge server for web-based medical applications. An ontology for
bioinformatics(TAMBIS) is presented in [6]. Most of the applications in this area
are based on terminological resources like the GUM ontology [7], the CYC ontology
[57], the Unied Medical Language System(UMLS) [71], etc.
Another area to which ontologies have been applied is that of ontology-based
brokering. These are specically designed agent systems which serve as brokers
between heterogeneous systems. They use ontologies to facilitate the information
brokering task. Representative applications are: the Ontobroker[17] which was used,
among others, in the KA
2
project[8]; the onto2agent[5] used to select publicly avail-
able ontologies on the web for a given application based on a Reference Ontology
developed by the same group to classify candidate ontologies; the OBSERVER[63]
system used to provide semantically rich information to a user who subscribes to
an information management system on the web which supported by selected on-
tologies; the IMPS(Internet-based Multi-agent Problem Solving)[16] system which
uses software agents to conduct knowledge acquisition on-line using distributed re-
sources. Terminological ontologies(like WordNet) were used to underpin the whole
process.
A related area of applications is that of knowledge retrieval. A representative
application in this area is the PlanetOnto which provides an integrated set of tools
to support news publishing based on ontology-driven document enrichment[19]. To
support this project two ontology-specic tools were developed: the Tadzebao and
WebOnto both described in [18]. The former aims to support a dialectical approach
in ontology design and maintenance while the latter provides editing and browsing
facilities. The goal of Tadzebao is to provide guidance for knowledge engineers
around ongoing dialogues for designing ontologies. This can be used as a negotiation
tool for proposed changes in an ontology with the additional
exibility that Tadzebao
oers: the integration of discussion about an artefact and its representation in the
same visual metaphor. Another application in this area is the knowledge-enhanced
search approach used in the FindUR project[62]. McGuinness describes a search
tool, deployed at the AT&T research labs, which uses ontologies to improve the
search experiences from the perspectives of recall and precision as well as ease of
query formation. A similar approach which deploys content matching techniques
is described in [44] where the authors present the OntoSeek system designed to
c
The project is electronically accessible from the URL: http://www.cs.man.ac.uk/mig/galen

Exploring ontologies 11
support content-based access to the web.
In the broader context of knowledge management(hereafter, KM) ontologies are
useful to support crucial KM tasks and activities. For an overview of the eld with
emphasis on the role of AI in KM we point the interested reader to O'Leary's review
in [74]. Here, we will use O'Leary's thesis that the goal of KM is to create valuable
information by employing the so called, converting and connecting processes([75], in
order to identify the role of ontologies in KM. The processes identied were: convert
(i)individual to group knowledge, (ii)data to knowledge, (iii)text to knowledge,
and connect (iv)people to knowledge, (v)knowledge to knowledge, (vi)people to
people, and (vii)knowledge to people. We argue that ontologies could be used
in most of these processes, either by playing a major role or by supplying the
supporting infrastructure that helps an organisation to implement them. In the
following paragraph we mention indicative examples from the ontology research
literature to justify this claim.
In particular, ontologies provide part of the infrastructure for conversion pro-
cesses(i to iii as listed above) and help in the connection activities(iv to vii as listed
above). Conversion processes (i) seem to benet more from the presence of ontolo-
gies as this is the underlying principle in their construction. Methodological[93]
and collaborative approaches([87],[8]) in ontology building, convert individual to
group knowledge in the form of an ontology. Processes (ii) and (iii) use other AI
technology like data and text mining techniques with ontologies being the guide to
the `right' data or text repository[17]. Ontologies seems to be more helpful in the
connecting processes. Process (iv) is concerned with the so called, `pull' technol-
ogy, which aims at pulling knowledge residing in vast repositories to people. The
means which used to pull that knowledge are, mainly, search engines and intelligent
agents. Examples of ontology use in this area are given in [62] and [44]. Process
(v) actually highlights the main contribution of ontologies: enabling communica-
tion and interoperability between systems. The best way to cite indicative work
here is to point to reviews and collections such as [93] and [41]. Process (vi) is not
directly related to ontologies as it is more concerned with technological means such
as Intranets. However, we should mention the work on collaboration and discussion
aided by ontologies[86]. In contrast with process (iv), process (vii) is concerned
with `push' technology. Means to achieve this are designated systems that focus
on content and push knowledge to the user instead of waiting for the user to pull
out that knowledge. As in (iv), ontologies play a major role here since they are
concerned with content and semantically enriched information. Example uses are
described in [22] and [17].
Finally, after having presented the processes that help to achieve the goal of KM
we close this section on KM by describing main KM tasks and activities and how
ontologies are related to them. These are summarised in gure 2 and described in
the following paragraph.
On the right hand side of gure 2, we illustrate the main KM tasks and activities.
We identify four main KM tasks: acquiring, analysing, using, and preserving knowl-

12 Handbook of Software Engineering and Knowledge Engineering
id
en
tif
y
acquire analyse use preserve
as
se
ss
m
o
de
l
sh
ar
e
re
u
se
ap
pl
y
m
ai
nt
ai
n
ca
pi
ta
lis
e
o
rg
an
is
e
KM
information extraction/
content matching
knowledge
representation
knowledge
sharing & reuse
KBS
libraries of reusable
components
areas of
activities
supported
by ontologies
applications
tasks
experience factories application area
O N T O L O G I E S
experience repositories
(tasks/activities)
Figure 2: Ontologies in knowledge management: activities supported by ontologies
help to achieve knowledge management tasks resulting in a range of application
areas.
edge. We argue that these tasks are accomplished by activities which are supported
by ontologies. In particular, the knowledge acquisition task, is accomplished by
identifying activities which are supported by ontologies. This results in the applica-
tion area of information extraction and/or content-matching. In the same manner,
ontologies in the area of knowledge representation are used to model and assess
the environment, which are activities employed in the analysing knowledge task.
The using knowledge task, includes the apply, share, and reuse activities, which
are supported by ontologies with such application areas as knowledge sharing and
reuse, and KBSs. The last task of the KM tasks/activities diagram is preserving
knowledge. It is accomplished by activities such as organising, maintaining, and
capitalising which are partially aided by ontologies. The resulting application area
is that of libraries of reusable knowledge components and experience repositories.
The knowledge preservation task and its accompanying activities along with the
relevant ontologies are the area of overlap with experience factories
d
as denoted by
the box surrounding the task in gure 2.
The last area to report is that of systems engineering. In section 7 we already
described systems like the AIRCRAFT and the Boeing experiment with the use
of the EngMath ontology which was also used in the construction of the PhysSys
ontologies set. Other representative applications are the ATOS(Advanced Technol-
ogy Operations System)([47]) system which was designed to meet specic needs of
spacecraft operations such as the need for coordination of dierent agent applica-
tions who had to commit to a common ontology. In [26], the authors describe the
Integrated Development Support Environment(IDSE), a commercial computational
d
Kalfoglou and Robertson investigate the overlap of ontologies and experience factories in [52].

Exploring ontologies 13
environment that supports the integration of enterprise models. The integration
is underpinned by axioms representing semantic constraints and relationships be-
tween dierent tools which are interpreted and enforced semi-automatically. This
information is contained in a method ontology, the IDEF1X
e
, accessed by a truth
maintenance system that enforces rules and constraints dened in the method.
The use of ontological axioms has been inspirational and many researchers are
investigating the practicality of deploying ontologies in software design. Gruber
hints the role of these axioms in ontology deployment:
\Ontologies are often equated with taxonomic hierarchies of classes, class
denitions, and the subsumption relation, but ontologies need not be
limited to these forms. Ontologies are also not limited to conservative
denitions, that is, denitions in the traditional logic sense that only
introduce terminology and do not add any knowledge about the world.
To specify a conceptualisation one needs to state axioms that do constrain
the possible interpretations for the dened terms."[37]
A working example for the use of ontological axioms in software design is de-
scribed in [53]. The authors point out that the role anticipated by ontological axioms
is rarely delivered: to restrict the possible interpretations ontological constructs
could have. To operationalise this role and enforce it in an integrated development
environment they invented a multi-layered architecture in which ontological ax-
ioms are separated from other ontological constructs included in a system that uses
the underlying ontology. These are enforced to comply to the axiomatisation in
order to verify the consistency of the system with respect to domain knowledge
as explicitly represented in the underlying ontology[50]. Ultimately, this layered
metaphor can be extended to check the ontological axioms themselves against an-
other set of axioms, meta-axioms, which could come from another ontology. This
facilitates the conformance check of an application to ontology and can be extended
to check dependencies among ontologies themselves[51]. Moreover, it supports the
integration of ontologies in applications while preserving their identity as being a
separate layer in the multi-layer architecture. The approach is illustrated in gure
3.
One of the early contributions that used ontological commitments was that of
the Comet[61] and Cosmos [60] systems. Both systems aim at developing knowledge
bases by capturing the set of ontological commitments that dene the interdepen-
dencies among key terms in the ontology. Their role is to assess the impact of
changes in their world and provide context-specic guidance to their users on what
modules may be relevant to include in the design, and what design modications
will be required in order to include them[59]. The key idea behind this work was
to make use of the ontological commitment expressed by the underlying ontology
in the system's development process.
Similarly, in the DISCOVER project[106], the role of ontological commitment
e
Electronically accessible from the URL: http://www.idef.com/overviews/idef1.html

14 Handbook of Software Engineering and Knowledge Engineering
consistency monitor
completeness monitor
weapon
naval target
ground target
aircraft target
is bomber
ontology database
radar
detected aircraft
navy threat
ground threat
air threat
defense
system
threat controlintelligence
hostile aircraft
power supply
air supply air load
pump
coil/magnet
valve-1 reservoir valve-2
lever
bellows
air pump system
grass rabbitphotosynthesis grazing defecation
respiration respiration
p.2 p.3
p.4 p.5
p.1
conformance check of
application to ontology/ies
Mereotopology
Systems Dynamics Theory
Ecological Modelling
Energy-flow models
Air Campaign Planning
AIRCRAFT
Meta-knowledge
Applications
Figure 3: The multi-layer approach: it enforces the conformance check of an ap-
plication to ontology/ies. Applications are using meta-knowledge constructs from
various ontologies(mereotopology, system dynamics theory, etc.) in an integrated
environment which enables checks on the use of those constructs against their ax-
iomatised denitions.
was further analysed and operationalised. The authors state that ontological com-
mitment is a key issue for knowledge sharing and reuse and they applied existing
verication techniques from the KBSs literature to check the commitment of a
knowledge base to an ontology. In that project the role of the ontology was to
act as a background body of knowledge against which a knowledge base can be
validated.
There are also a number of applications related with projects undertaken by
various organisations involving academic and industrial partners. We already men-
tioned some of them in the previous sections. We complete our coverage here by
describing one of the rst projects in this area which was the Knowledge Sharing
Eort(KSE)[72] aimed to realise the benet of sharing and reusing large knowledge
bases. The distinguishable contribution of this project was the Knowledge Inter-
change Format(KIF) framework. Other projects are the High Performance Knowl-
edge Bases(HPKB) programme [15] which aims at fostering the development of
technologies that can increase the rate at which we can write knowledge bases. The
Intelligent Brokering Service for Knowledge-Component Reuse on the World-Wide
Web(IBROW
3
) project investigates means for supporting comprehensive reuse. The
idea behind this project is to provide a brokering service that plays the role of a

Exploring ontologies 15
mediator between customers and PSM providers to support the conguration of cus-
tomised knowledge systems that solve customers' problems. A library of reusable
components[70] has been constructed based on the work of Motta in parametric de-
sign [69]. The Knowledge Reuse and Fusion/Transformation(KRAFT) [79] aimed
to enable the sharing and reuse of information contained in heterogeneous databases
and knowledge bases. In the area of planning the SPAR[89] project draws on the
range of previous work in planning activity ontologies to create a practically useful
Shared Planning and Activity Representation.
9. Problems, tradeos, and solutions
Despite the fact that ontologies have been applied with success in a variety of
elds there are reported problems and attempts have been made to identify trade-
os and nd potential solutions. We report on the problems rst. In [73] the
author discusses impediments in the use of ontologies. He points out the diÆ-
culty in library ontologies, scale-up, interfacing and raises the issue of formality
in ontology development. O'Leary argues also for the diÆculty in establishing a
consensus: \ontologies are chosen after a political decision had been made, there-
fore it is impossible to choose an ontology that maximises the utility of all agents
in process and the group."[73]. Research in the area of studying the experts' be-
haviour provides an evidence of the apparent lack of consensus. For example, many
researchers argue that experts disagree about even well-established features of their
domain(see, [27], [65]). Others studied the behaviour of experts([82]) and found
that they often held dierent views about a supposedly standard terminology in
their eld. Furthermore, in [2] the authors state: \expert knowledge is comprised
of context-dependent, personally-constructed, highly-functional but fallible abstrac-
tions". This suggests that we should routinely expect evolution of experts' views,
especially in domains where there is disagreement on used terms. However, we have
to point out here that in situations where there is a lack of consensus among the ex-
perts regarding the domain of interest then the principle of ontology does not apply
by denition: an ontology represents consensual and commonly agreed terminology
about a domain of interest. In that respect we agree with O'Leary's thesis and as
a rule of thumb we can say that in domains where there is literally no consensus
among the domain experts then building an ontology is pointless. This, however,
should not be interpreted as a guideline to build ontologies only when experts agree:
this will rarely happen, as the studies described above suggest, therefore we have
seen the most successful ontology stories coming from domains where the `majority'
of experts agree on used terms. The issue here is to nd the right balance between
commonly agreed terminology and usability of ontology. This is actually one of
the ontology design principles: minimal ontological commitment[38]. We should
also mention that experience with task models in the KBS community indicates a
broad degree of consensus with respect to the structure of KBS tasks, like diagno-
sis, parametric design, scheduling, etc. The key factor here is the eective support
for KBS development rather than achieving community-wide consensus, a goal of

16 Handbook of Software Engineering and Knowledge Engineering
generic ontologies.
Other problematic areas have been identied: Uschold and colleagues raise the
issue of lack of translators when the representation formalisms used are not the same
in the context of their experiment for ontology reuse[92]. They argued that, \the
translation activity involved was an intensive one and lack of automatic support is
an important disadvantage". The issue of ease of reuse was also the focal point of an
empirical study performed in the context of the HPKB project. In [14], the authors
report that ease of reuse is closely related to the type of ontology: it was found
that very generic ontologies provide less support and are less useful than domain-
specic ones. The latter scored a constant 60% rate of reuse in the HPKB study
in contrast with the poor 22% rate of reuse scored by broad ontologies. However,
as the authors argue, these results should not undermine their role in structuring
ontologies: \Although the rate of reuse of terms from very general ontologies may
be signicantly lower, the real advantage of these ontologies probably comes from
helping knowledge engineers organise their knowledge bases along sound ontological
lines."[14].
Another important drawback is the lack of rigorous evaluation techniques for on-
tologies. For example, in an experiment of extending the HPKB upper ontology[3]
the author states: \. . . validation remains an important issue, i.e., the PhysSys,
EngMath and topology ontologies are capable of being validated by reference to lit-
erature in their application elds. . . but ontologies such as the HPKB upper level
and SPAR do not capture knowledge in such well understood elds, therefore this
form of validation is not possible". The issue of maintenance has also been acknowl-
edged and studied by many. Robertson neatly summarises the points made: \the
cost of producing an ontology is not just in inventing the domain-specic formal
language but in maintaining it once the system is deployed, since perfect ontologies
cannot be guaranteed. Over-commitment to perfecting an ontology causes failure
either during development(through irreconcilable arguments over what the ontol-
ogy should be) or after deployment(through inappropriate human interpretation of
inference system inputs or outputs)"[81]. In the long run this cost might hinder fur-
ther deployment of ontologies. However, it is not easily predictable and quantiable
since there are various angles of viewing this problem. For instance, if we accept
that ontology rarely stabilises then we should expect to include in our budget along
with the cost of constructing, costs for maintaining the ontology we use as well as
the system which uses it. How common is ontology instability? We don't know
since we have very little experience with the long-term use of large libraries of on-
tologies. However, this is a debatable point[64] and we nd projects where ontology
was deployed on the rationale that it was stable(i.e., in parametric design ontology:
[69]), and projects where this is not taken for granted as ontology is expected to
change over time(i.e., Aitken's HPKB experiment[3]).
We now shift our attention to potential solutions to some of the problems men-
tioned above. With respect to the problem of library ontologies, made by O'Leary,
the online libraries of ontologies(i.e., Ontolingua) are a potential solution espe-

Exploring ontologies 17
cially when the maintenance and update facilities that are envisaged[25] will be
fully integrated. To facilitate the familiarisation task, systems like OntoSaurus and
WebOnto[18] aim to help the engineer accomplishing this task. The issue of inter-
facing has attracted a lot of attention by the community. It is seen from dierent
angles: `integration', `merging', `mapping', are some of the terms used. A sum-
mary of these approaches is given in [76] whereas Visser and colleagues analyse
the nature of the problem in [104]. Some of the solutions proposed and applied
are the ontological mediation algorithms[11], the ontology clustering[105], as well
as the approaches used in projects like the creation of the AIRCRAFT ontology
and in [94]. In addition to these, the OBSERVER [63] and ONIONS[31] systems,
the Partially Shared Views(PSV) scheme[56], the encapsulation and composition
technique[46] in the context of the Scalable Knowledge Composition(SKC)
f
project
provide alternative solutions.
Even with this plethora of techniques the situation remains unsettled. There is
no comparative analysis which identies potential advantages and important draw-
backs and no common practices to be followed. This has started to change with
the proposal of frameworks that characterise ontologies, like the one originally pre-
sented by Uschold in [91] which was further analysed in [95]. These frameworks
can be used to share experiences, discuss tradeos, and disseminate knowledge re-
garding attempts to apply ontologies. A small example of this is the instantiation
of Uschold's framework, made by Kalfoglou and Robertson in the context of the
PhysSys ontologies set[52]. Another source of information is from comparative anal-
yses. For example, in [29], the authors compare and analyse the state-of-the-art in
ontology design. Ushcold and Jasper present a cost-benet analysis of three com-
monly used approaches in knowledge sharing[96]. In a larger context, Kalfoglou and
colleagues, compare various meta-knowledge types, analyse their cost-benets, and
identify pragmatic aspects in using meta-knowledge[49]. In similar fashion Menzies
and colleagues analyse issues with meta-knowledge in [64] and Kalfoglou speculates
on the role of formal ontologies in knowledge maintenance in [48].
We close this section by summarising the points made and speculating on the fu-
ture of ontology applications. We observe a shift of interest by the community from
very generic, broad ontologies to domain-related ones tailored to serve particular ap-
plications. We also saw evidence in the reported systems above that ontologies can
improve systems design in such areas as knowledge sharing and reuse and contribute
to enhance their reliability by consistency checking. This could have an impact by
reducing production costs, shortening development times and communicating con-
text among applications and across organisations. It also improves the quality of
the resulted systems with respect to verication of their correctness against domain
knowledge. However, there are serious obstacles to overcome. The most impor-
tant being, the considerably high cost of constructing an ontology from scratch,
the lengthy learning curve which has to be traversed in order to become familiar
with an ontology before integrating it in the system, the lack of rigid maintenance
f
Electronically accessible from the URL: http://www-db.stanford.edu/SKC/

18 Handbook of Software Engineering and Knowledge Engineering
strategies, and the dearth of metrics for assessing ontology.
10. Resources
As we stated earlier, an exhaustive review of the ontologies eld is impractical and
overwhelming for the reader. However, for the sake of disseminating up-to-date
information on ontologies we have selected and include here pointers to publicly
available online resources. These are:
 a comprehensive collection of ontology-related research in alphabetical order,
maintained by Peter Clark:
http://www.cs.utexas.edu/users/mfkb/related.html
 a similar collection maintained by Enrico Franconi:
http://www.cs.man.ac.uk/franconi/ontology.html
 a list maintained by Adam Farquhar:
http://ksl-web.stanford.edu/kst/ontology-sources.html
 a catalogue with classied information on ontologies prepared by Yannis Kalfoglou
for a panel debate that took place in the SEKE'99 conference:
http://www.dai.ed.ac.uk/daidb/people/homes/yannisk/seke99panelhtml.html
In addition to these periodically updated online resources there are several overviews
in the literature. These are, the Uschold and Gruninger review[93], the comparative
review of Fridman-Noy and Hafner in [29], the survey of ontology research in [13],
an overview of ontologies and PSMs in [34], and a review of planning ontologies by
Tate in [89]. There are also special issues in referred journals devoted to ontology
research: with respect to their role in IT[45], their involvement in KBSs[103], and
their uses[99]. In addition, we should mention the volume edited by Guarino in [41].
Acknowledgements
The research described in this paper is supported by a European Union Marie
Curie Fellowship (programme: Training and Mobility of Researchers).
1. N.F. Abernethy, J.F. Wu, M. Hewett, and R.B. Altman. Sophia: A Flexible, Web-Based
Knowledge Server. IEEE Intelligent Systems, 14(4):79{85, July 1999.
2. N.M. Agnew, K.M. Ford, and P.J. Hayes. Expertise in Context: Personally Constructed,
Socially Elected, and Reality-Relevant? International Journal of Expert Systems,
7(1), 1993.
3. S. Aitken. Extending the HPKB-Upper-Level Ontology: Experiences and Observations.
In A. Gomez-Perez and R. Benjamins, editors, Proceedings of Workshop on Applica-
tions of Ontologies and Problem Solving Methods, ECAI'98, Brighton, England,
August 1998.
4. G. Arango. Software Reusability/Domain Analysis Methods, pages 17{49. Ellis
Horwood, Chichester, Enland, 1994.
5. J. Aspirez, A. Gomez-Perez, A. Lozano, and S. Pinto. (onto)2agent: An ontology-based
www broker to select ontologies. In Proceedings of the Workshop on Applications
of Ontologies and Problem-Solving Methods, ECAI'98, Brighton, England, pages
16{24, August 1998.

Exploring ontologies 19
6. P. Baker, C. Goble, S. Bechhofer, N. Paton, R. Stevens, and A. Brass. An ontology for
bioinformatics applications. Bionformatics, 15(6):510{520, 1999.
7. J. Bateman, B. Magnini, and G. Fabris. The Generalized Upper Model(GUM)
knowledge-base: Organisation and use. In N.J.I. Mars, editor, Proceedings of the 2nd
International Conference on Knowledge Building and Knowledge Sharing(KB &
KS'95), Twente, The Netherlands, pages 60{72, Amsterdam, NL, 1995. IOS Press.
8. R. Benjamins and D. Fensel. The Ontological Engineering Initiative-KA2. In N. Guar-
ino, editor, Proceedings of the 1st International Conference on Formal Ontologies
in Information Systems, FOIS'98, Trento, Italy, pages 287{301. IOS Press, June
1998.
9. M. Blazquez, M. Fernadez, J.M. Garcia-Pinar, and A. Gomez-Perez. Building Ontolo-
gies at the Knowledge Level using the Ontology Design Environment. In Proceedings of
the 11th Knowledge Acquisition, Modelling and Management Workshop, KAW98,
Ban, Canada, April 1998.
10. P. Borst, H. Akkermans, and J. Top. Engineering Ontologies. International Journal
of Human-Computer Studies, 46:365{406, 1997.
11. A.E. Campbell and S.C. Shapiro. Algorithms for Ontological Mediation. Technical
Report 98-03, Department of Computer Science and Engineering, State University of
New York at Bualo, January 1998.
12. B. Chandrasekaran, J.R. Josephson, and R. Benjamins. The Ontology of Tasks and
Methods. In Proceedings of the 11th Knowledge Acquisition Modeling and Man-
agement Workshop, KAW'98, Ban, Canada, April 1998.
13. B. Chandrasekaran, R. Josephson, and R. Benjamins. What Are Ontologies, and Why
Do We Need Them? IEEE Intelligent Systems, 14(1):20{26, January 1999.
14. P. Cohen, V. Chaudhri, A. Pease, and R. Schrag. Does prior knowledge facilitate the
development of knowledge-based systems? In Proceedings of the Sixteenth National
Conference on Articial Intelligence, AAAI'99, Orlando, FL, USA, pages 221{
226, July 1999.
15. P. Cohen, R. Schrag, E. Jones, A. Pease, A. Lin, B. Starr, D. Gunning, and M. Burke.
The DARPA High Performance Knowledge Bases project. AI Magazine, 19(4):25{49,
1998.
16. L. Crow and N. Shadbolt. Acquiring and Structuring Web Content with Knowledge
Level Models. In R. Studer and D. Fensel, editors, Proceedings of the 11th European
Workshop on Knowledge Acquisition, Modelling and Management(EKAW'99),
Dagstuhl, Germany, pages 83{101. Springer Verlag, May 1999.
17. S. Decker, M. Erdmann, D. Fensel, and R. Studer. Ontobroker: Ontology Based Access
to Distributed and Semi-Structured Information. In R & et.al. Meersman, editor,
Proceedings of DS-8, Semantic Issues in Multimedia Systems, Boston, MA, USA,
pages 351{369, 1999.
18. J. Domingue. Tadzebao and WebOnto: Discussing, Browsing, and Editing Ontologies
on the Web. In Proceedings of the 11th Knowledge Acquisition, Modelling and
Management Workshop, KAW'98, Ban, Canada, April 1998.
19. J. Domingue and E. Motta. Planet-Onto: From News Publishing to Integrated Knowl-
edge Management Support. IEEE Intelligent Systems(In Press), 2000.
20. Enterprise Integration Laboratory. EIL. TOVE Project, University of Toronto, Canada.
available from http://www.ie.utoronto.ca/EIL/tove/ontoTOC.html, July 1995.
21. A. Farquhar, R. Fikes, and J. Rice. The ontolingua server: a tool for collaborative ontol-
ogy construction. International Journal of Human-Computer Studies, 46(6):707{
728, June 1997.
22. D. Fensel, V.R. Benjamins, E. Motta, and B. Wielinga. UPML: A Framework for
knowledge system reuse. In Proceedings of the 16th International Joint Conference

20 Handbook of Software Engineering and Knowledge Engineering
on Articial Intelligence, IJCAI'99, Stockholm, Sweden, pages 16{21, August 1999.
23. M. Fernandez. Overview for Methodologies for Building Ontologies. In Proceedings
of the IJCAI-99 Workshop on Ontologies and Problem-Solving Methods(KRR5),
Stockholm, Sweden, August 1999. Available from: http://sunsite.informatik.rwth-
aachen.de/Publications/CEUR-WS/Vol-18/.
24. M. Fernandez, A. Gomez-Perez, and N. Juristo. METHONTOLOGY: From Ontolog-
ical Arts Towards Ontological Engineering. In Proceedings of the AAAI-97 Spring
Symposium Series on Ontological Engineering, Stanford, CA, USA, pages 33{40,
March 1997.
25. R. Fikes and A. Farquhar. Distributed Repositories of Highly Expressive Reusable
Ontologies. IEEE Intelligent Systems, 14(2):73{79, March 1999.
26. F. Fillion and C. Menzel. Using ontologies to enable enterprise model integration. in Ap-
pendices of the paper: "Ontologies: principles, methods and applications", Uschold,M.
and Gruninger,M., The Knowledge Engineering Review, 11(2), p.130-136, 1996.
27. A. Finkelstein, D. Gabbay, A. Hunter, J. Kramer, and B. Nuseibeh. Inconsistency
handling in multi-perspective specications. IEEE Transactions on Software Engi-
neering, 20(8):569{578, 1994. also as a Research Report DoC 93/2.
28. M.S. Fox and M. Gruninger. On Ontologies and Enterprise Modelling. In Proceedings
of International Conference on Enterprise Integration Modelling Technology 97.
Springer-Verlag, 1997.
29. N. Fridman-Noy and C.D. Hafner. The State of the Art in Ontology Design: A Survey
and Comparative Review. AI Magazine, 18(3):53{74, 1997.
30. E. Gamma, R. Helm, R. Johnson, and J. Vlissides. Design Patterns: Elements of
Reusable Object-Oriented Software. Addison-Wesley, 1995.
31. A. Gangemi, D. Pisanelli, and G. Steve. Ontology Integration: Experiences with Medical
Terminologies. In N. Guarino, editor, Proceedings of the 1st International Confer-
ence on Formal Ontology in Information Systems, FOIS'98, Trento, Italy, pages
163{178, June 1998.
32. A. Gangemi, G. Steve, and F. Giacomelli. ONIONS: An Ontological Methodology for
Taxonomic Knowledge Integration. In P. van der Vet, editor, Proceedings of the Work-
shop on Ontological Engineering, ECAI'96, Budapest, Hungary, August 1996.
33. R. Genesereth and R. Fikes. Knowledge Interchange Format. Computer Science
Dept., Stanford University, 3.0 edition, 1992. Technical Report, Logic-92-1.
34. A. Gomez-Perez and R. Benjamins. Overview of Knowledge Sharing and Reuse Com-
ponents: Ontologies and Problem-Solving Methods. In Proceedings of the IJCAI-
99 Workshop on Ontologies and Problem-Solving Methods(KRR5), Stockholm,
Sweden, August 1999. http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-
WS/Vol-18/.
35. A. Gomez-Perez and MD. Royas-Amaya. Ontological Reengineering for Reuse. In
Proceedings of the 11th European Workshop on Knowledge Acquisition, Modeling
and Management,(EKAW99), Dagstuhl, Germany, pages 139{157, May 1999.
36. T. Gruber and G. Olsen. An ontology for engineering mathematics. In J. Doyle,
P. Torasso, and E. Sandewall, editors, Proceedings of the Fourth International Con-
ference on Principles of Knowledge Representation and Reasoning, San Mateo,
CA, USA., pages 258{269, 1994.
37. T.R. Gruber. A Translation Approach to Portable Ontologies. Knowledge Acquisition,
5(2):199{220, 1993.
38. T.R. Gruber. Towards principles for the design of ontologies used for knowledge sharing.
International Journal of Human-Computer Studies, 43:907{928, 1995.
39. M. Gruninger and M.S. Fox. Methodology for the Design and Evaluation of Ontolo-
gies. In Proceedings of the IJCAI'95 Workshop on Basic Ontological Issues in

Exploring ontologies 21
Knowledge Sharing, Montreal, Quebec,Canada, August 1995.
40. N. Guarino. Formal Ontology and Information Systems. In N. Guarino, editor, Pro-
ceedings of the 1st International Conference on Formal Ontologies in Information
Systems, FOIS'98, Trento, Italy, pages 3{15. IOS Press, June 1998.
41. N. Guarino, editor. Formal Ontology In Information Systems, Frontiers in Articial
Intelligence and Applications. IOS Press, June 1998. ISBN: 90-5199-399-4.
42. N. Guarino, M. Carrara, and P. Giaretta. Formalizing Ontological Commitments. In In
Proceedings of the 12th National Conference on Articial Intelligence(AAAI'94),
Seattle, Washington, USA, 1994.
43. N. Guarino and P. Giaretta. Ontologies and Knowledge Bases: Towards a Terminologi-
cal Clarication. In Proceedings of the 2nd International Conference on Knowledge
Building and Knowledge Sharing(KB&KS'95), Twente, The Netherlands, April
1995.
44. N. Guarino, C. Masolo, and G. Vetere. OntoSeek: Content-Based Access to the Web.
IEEE Intelligent Systems, 14(3):70{80, May 1999.
45. N. Guarino and R.(eds.) Poli. The Role of Ontology in the Information Technology.
International Journal of Human-Computer Studies, 43(5/6):623{965, 1995.
46. J. Jannink, S. Pichai, D. Verheijen, and G. Wiederhold. Encapsulation and Compo-
sition of Ontologies. In Proceedings of the AAAI'98 Workshop on Information
Integration, Madison, WI, USA, July 1998.
47. M. Jones, J. Wheadon, D. Whitgift, M. Niezatte, M. Timmermans, R. Rodriquez, and
R. Romero. An agent-based approach to spacecraft mission operations. In N.J.I. Mars,
editor, Proceedings of 2nd International Conference on Knowledge Building and
Knowledge Sharing(KB&KS'95), Twente, The Netherlands, pages 259{269. IOS
Press, April 1995.
48. Y. Kalfoglou. The Role of Formal Ontologies. In Proceedings of the 11th In-
ternational Conference on Software Engineering and Knowledge Engineering,
SEKE'99, Kaiserslauten, Germany, pages 401{405, June 1999. Position paper pre-
sented at the Panel: Knowledge Maintenance. Also as: Research Paper No.952, Dept.
of AI, University of Edinburgh.
49. Y. Kalfoglou, T. Menzies, K-D. Altho, and E. Motta. Meta-knowledge in sys-
tems design: panacea...or undelivered promise? The Knowledge Engineering Re-
view(submitted), 2000.
50. Y. Kalfoglou and D. Robertson. A Case Study in Applying Ontologies to Augment
and Reason about the Correctness of Specications. In Proceedings of the 11th In-
ternational Conference on Software Engineering and Knowledge Engineering,
SEKE'99, Kaiserslauten, Germany, pages 64{71, June 1999. Also as: Research Pa-
per No.927, Dept. of AI, University of Edinburgh.
51. Y. Kalfoglou and D. Robertson. Managing Ontological Constraints. In Proceedings of
the IJCAI-99 Workshop on Ontologies and Problem-Solving Methods(KRR5),
Stockholm, Sweden, http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-
WS/Vol-18/, August 1999. Also as: Research Paper No.948, Dept. of AI, University of
Edinburgh.
52. Y. Kalfoglou and D. Robertson. Applying Experienceware to support ontology deploy-
ment. In Proceedings of the 12th International Conference on Software Engineer-
ing and Knowledge Engineering, SEKE2000, Chicago, IL, USA, pages 266{275,
July 2000.
53. Y. Kalfoglou, D. Robertson, and A. Tate. Using Meta-Knowledge at the Application
Level. Journal of Articial Intelligence Research, submitted, 2000. Also as: Research
Paper No.956, Dept. of AI, University of Edinburgh.
54. K. Knight and S. Luk. Building a Large Knowledge Base for Machine Translation.

22 Handbook of Software Engineering and Knowledge Engineering
In Proceedings of the American Association of Articial Intelligence Conference-
AAAI 94, Seattle, USA, July 1994.
55. J. Lee, M. Gruninger, Y. Jin, T. Malone, A. Tate, G Yost, and other members of the
PIF working group. The PIF Process Interchange Format and framework. Knowledge
Engineering Review, 13(1):91{120, February 1998.
56. J. Lee and T. Malone. Partially Shared Views: A Scheme for Communicating between
Groups Using Dierent Type Hierarchies. ACM Transactions on Information Sys-
tems, 8(1):1{26, 1990.
57. D.B. Lenat and R.V. Guha. Building large knowledge-based systems. Represen-
tation and inference in the Cyc project. Addison-Wesley, Reading, Massachusetts,
1990.
58. W. Mark. Ontologies as Representation and Re-Representation of Agreement. In Pro-
ceedings of the 5th International Conference on Principles of Knowledge Rep-
resentation and Reasoning, KR'96, Massachusetts, USA, 1996. Position paper
presented on the panel: Ontologies: What are they and where's the research.
59. W. Mark, J. Dukes-Schlossberg, and R. Kerber. Ontological Commitment and Domain-
Specic Architectures: Experience with Comet and Cosmos. In N.J.I.Mars, editor, Pro-
ceedings of the 2nd International Conference on Knowledge Building and Knowl-
edge Sharing(KB & KS'95), Twente, The Netherlands, pages 33{45, April 1995.
60. W. Mark, J. Schlossberg, S. Tyler, and J. McGuire. Cosmos: A System for Support-
ing Engineering Negotiation. Concurrent Engineering: Research and Applications,
2:173{182, 1994.
61. W. Mark, S. Tyler, J. McGuire, and J. Schossberg. Commitment-Based Software De-
velopment. IEEE Transactions on Software Engineering, 18(10):870{884, October
1992.
62. L.D. McGuinness. Ontological Issues for Knowledge-Enhanced Search. In N. Guarino,
editor, Proceedings of the 1st International Conference on Formal Ontology in
Information Systems(FOIS'98), Trento, Italy, pages 302{316. IOS Press, June 1998.
63. E. Mena, V. Kashyap, A. Illarramendi, and A. Sheth. Domain Specic Ontologies for
Semantic Information Brokering on the Global Information Infrastructure. In N. Guar-
ino, editor, Proceedings of the 1st International Conference on Formal Ontology
in Information Systems(FOIS'98), Trento, Italy, pages 269{283. IOS Press, June
1998.
64. T. Menzies, K-D. Altho, Y. Kalfoglou, and E. Motta. Issues with Meta-Knowledge.
International Journal of Software Engineering and Knowledge Engineering(to
appear), 10(4), August 2000.
65. T. Menzies and B. Clancey. Special Issue on Situated Cognition. International Journal
of Human-Computer Studies, 49, 1998. Editorial.
66. G.A. Miller. WORDNET: an online lexical database. International Journal of Lexi-
cography, 3(4):235{312, 1990.
67. R. Mizoguchi, J. van Welkenhuysen, and M. Ikeda. Task ontology for reuse of problem-
solving knowledge. In N.J.I. Mars, editor, Proceedings of the 2nd International Con-
ference on Knowledge Building and Knowledge Sharing(KB & KS'95), Twente,
The Netherlands, pages 46{57, Amsterdam, NL., 1995. IOS Press.
68. T.P. Moran and J.M. Carroll. Design Rationale: Concepts, Techniques, and Use.
Lawerence Erlbaum Associates, 1996. ISBN: 0-8058-1567-8.
69. E. Motta. Reusable Components for Knowledge Models: Case Studies in Para-
metric Design Problem Solving, volume 53 of Frontiers in Articial Intelligence
and Applications. IOS Press, 1999. ISBN: 1-58603-003-5.
70. E. Motta, D. Fensel, M. Gaspari, and R. Benjamins. Specications of Knowledge Com-
ponents for Reuse. In Proceedings of the 11th International Conference on Soft-

Exploring ontologies 23
ware Engineering and Knowledge Engineering, SEKE'99, Kaiserslauten, Ger-
many, pages 36{43, June 1999.
71. National Library of Medicine, Bethesda, Maryland, USA. UMLS Knowledge Sources,
1997.
72. R. Neches, R.E. Fikes, T. Finin, T.R. Gruber, T. Senator, and W.R. Swartout. Enabling
Technology for Knowledge Sharing. AI Magazine, 12(3):36{56, 1991.
73. D. O'Leary. Impediments in the use of explicit ontologies for KBS development. In-
ternational Journal of Human-Computer Studies, 46(2):327{337, 1997.
74. D. O'Leary. Knowledge Management Systems: Converting and Connecting. IEEE
Intelligent Systems, 13(3):30{33, June 1998.
75. D. O'Leary. Using AI in Knowledge Management: Knowledge Bases and Ontologies.
IEEE Intelligent Systems, 13(3):34{39, June 1998.
76. S. Pinto, A. Gomez-Perez, and J. Martins. Some Issues on Ontology Integration.
In Proceedings of the IJCAI-99 Workshop on Ontologies and Problem-Solving
Methods(KRR5), Stockholm, Sweden, August 1999. http://sunsite.informatik.rwth-
aachen.de/Publications/CEUR-WS/Vol-18/.
77. S.T. Polyak. A Supply Chain Process Interoperability Demonstration using the Pro-
cess Interchange Format(PIF). Research Paper No.917, Department of Articial Intel-
ligence, University of Edinburgh, February 1998.
78. C. Potts. Supporting Software Design: Integrating Design Methods and Design Ratio-
nale. In P.T. Moran and M.J. Carroll, editors, Design Rationale: Concepts, Tech-
niques, and Use, chapter 10, pages 295{321. Lawrence Erlbaum Associates, 1996.
79. A. Preece, K. Hui, P. Gray, P. Marti, T. Bench-Capon, D. Jones, and Z. Cui. The
KRAFT Architecture for Knowledge Fusion and Transformation. In Proceedings of
the 19th SGES International Conference on Knowledge-based Systems and Ap-
plied Articial Intelligence(ES'99), Cambridge, England. Springer Verlag, Decem-
ber 1999. Best Technical Paper Award.
80. A.L. Rector, W.A. Nowlan, and the GALEN Consortium. The GALEN Project. Com-
puter Methods and Programs in Biomedicine, 45:75{78, 1995.
81. D. Robertson. Pitfalls of Formality in Early System Design. In Proceedings of the 1998
ARO/NSF Monterey Workshop on Increasing the Practical Impact of Formal
Methods for Computer-Aided Software Development, Carmal, California, 1998.
82. M. Shaw. Validation in a Knowledge Acquisition System with Multiple Experts. In pro-
ceedings of the International Conference on 5th Generation Computer Systems,
pages 1259{1266, 1988.
83. D. Skuce. Viewing Ontologies as Vocabulary: Merging and Documenting the Logical
and Linguistic Views. In Proceedings of the IJCAI'95 Workshop on Basic Onto-
logical Issues in Knowledge Sharing, Montreal, Quebec, Canada, August 1995.
84. J. Sowa. Knowledge Representation: Logical, Philosophical, and Computational
Foundations. Brooks Cole Publishing Co., Pacic Grove,CA,USA, 2000. ISBN: 0-534-
94965-7.
85. R. Studer, V.R. Benjamins, and D. Fensel. Knowledge engineering, principles and
methods. Data and Knowledge Engineering, 25(1-2):161{197, 1998.
86. T. Summer and S. Buckingham-Shum. From Documents to Discourse: Shifting Con-
ceptions of Scholarly Publishing. In proceedings of the CHI'98: Human Factors in
Computing Systems, Los Angeles, CA, USA, pages 95{102. ACM Press, 1998.
87. B. Swartout, R. Patil, K. Knight, and T. Russ. Toward Distributed Use of Large-
Scale Ontologies. In Proceedings of the 10th Knowledge Acquisition, Modeling and
Management Workshop(KAW'96),Ban,Canada, November 1996.
88. W. Swartout and A. Tate. Ontologies - Guest editors' introduction. IEEE Intelligent
Systems, 14(1):18{19, January 1999.

24 Handbook of Software Engineering and Knowledge Engineering
89. A. Tate. Roots of SPAR - Shared Planning and Activity Representation. The Knowl-
edge Engineering Review, 13(1):121{128, 1998.
90. M. Uschold. Knowledge level modelling: concepts and terminology. The Knowledge
Engineering Review, 13(1):5{29, February 1998.
91. M. Uschold. Where are the Killer Apps? In Gomez-Perez,A. and Benjamins,R., edi-
tor, Proceedings of Workshop on Applications of Ontologies and Problem Solving
Methods, ECAI'98, Brighton, England, August 1998.
92. M. Uschold, P. Clark, M. Healy, K. Williamson, and S. Woods. An Experiment in
Ontology Reuse. In Proceedings of the 11th Knowledge Acquisition Workshop,
KAW98, Ban, Canada, April 1998.
93. M. Uschold and M. Gruninger. Ontologies: principles, methods and applications. The
Knowledge Engineering Review, 11(2):93{136, November 1996.
94. M. Uschold, M. Healy, K. Williamson, P. Clark, and S. Woods. Ontology Reuse and
Application. In N. Guarino, editor, Proceedings of the 1st International Conference
on Formal Ontology in Information Systems(FOIS'98), Trento, Italy, pages 179{
192. IOS Press, June 1998.
95. M. Uschold and R. Jasper. A Framework for Understanding and Classifying On-
tology Applications. In Proceedings of the IJCAI-99 Workshop on Ontolo-
gies and Problem-Solving Methods(KRR5), Stockholm, Sweden, August 1999.
http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-18/.
96. M. Uschold, R. Jasper, and P. Clark. Three Approaches for Knowledge Sharing: A
Comparative analysis. In Proceedings of the 12th Knowledge Acquisition, Modelling
and Management Workshop, KAW'99, Ban, Canada, October 1999.
97. M. Uschold, M. King, S. Moralee, and Y. Zorgios. The enterprise ontology. The
Knowledge Engineering Review, 13(1), February 1998. Also available as AIAI-TR-
195 from AIAI, University of Edinburgh.
98. M. Uschold and King.M. Towards a methodology for building ontologies. In Pro-
ceedings of the IJCAI-95 Workshop on Basic Ontological Issues in Knowledge
Sharing, Montreal, Canada, 1995.
99. M. Uschold and A.(eds.) Tate. Putting Ontologies to Use. The Knowledge Engineer-
ing Review, 13(1):1{128, 1998.
100. A. Valente, T. Russ, R. MacGrecor, and W. Swartout. Building and (Re)Using an
Ontology for Air Campaign Planning. IEEE Intelligent Systems, 14(1):27{36, January
1999.
101. P. van der Vet and N. Mars. Structured system of concepts for storing, retrieving, and
manipulating chemical information. Journal of Chemical Information and Com-
puter Science, 33:564{568, 1993.
102. P. van der Vet and N. Mars. Bottom-Up Construction of Ontologies. IEEE Transac-
tions on Knowledge and Data Engineering, 10(4):513{526, 1998.
103. G. van Heijst, A. Schreider, and B.(eds.) Wielinga. Using Explicit Ontologies in KBS
Development. International Journal of Human-Computer Studies, 46(2/3):183{292,
1997.
104. P.R.S. Visser, D.M. Jones, T.J.M. Bench-Capon, and M.J.R. Shave. Assessing Het-
erogeneity by Classifying Ontology Mismatches. In N. Guarino, editor, Proceedings
of 1st International Conference on Formal Ontologies in Information Systems,
FOIS'98, Trento, Italy, pages 148{162. IOS Press, June 1998.
105. P.R.S. Visser and V.A.M Tamma. An Experiment with Ontology-based Agent Cluster-
ing. In Proceedings of the IJCAI-99 Workshop on Ontologies and Problem-Solving
Methods(KRR5), Stockholm,Sweden, August 1999. http://sunsite.informatik.rwth-
aachen.de/Publications/CEUR-WS/Vol-18/.
106. A. Waterson and A. Preece. Verifying Ontological Commitment in Knowledge-based

Exploring ontologies 25
Systems. Knowledge-Based Systems, 12:45{54, April 1999.
107. WfMC. Work
ow Management Coalition: Abstract Specication. WFMC-TC 1012,
WfMC, October 1996. Interoperability demonstration presented at the 1996 Business
Process and Work
ow Conference in Amsterdam.
108. T. Winograd. Bringing Design to Software. Addison-Wesley, 1996. ISBN: 0-201-
85491-0.