Electronic Business: Concepts, Methodologies, Tools, and Applications (4-Volumes) P238 pptx

2304
Copyright © 2009, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited.
Chapter 7.21
A Basis for the Semantic Web
and E-Business:
(IۋFLHQW2UJDQL]DWLRQRI2QWRORJ\
Languages and Ontologies
Changqing Li
National University of Singapore, Singapore
Tok Wang Ling
National University of Singapore, Singapore
ABSTRACT
This chapter introduces how to effectively orga-
nize ontology languages and ontologies and how
WRHI¿FLHQWO\SURFHVVVHPDQWLFLQIRUPDWLRQEDVHG
on ontologies. In this chapter we propose the
hierarchies to organize ontology languages and
ontologies. Based on the hierarchy of ontology
languages, the ontology designers need not bear
in mind which ontology language the primitives
exactly come from, also we can automatically
DQGVHDPOHVVO\XVHWKHRQWRORJLHVGH¿QHGZLWK
different ontology languages in an integrated en-
vironment. Based on the hierarchy of ontologies,
WKHFRQÀLFWVLQGLIIHUHQWRQWRORJLHVDUHUHVROYHG
thus the semantics in different ontologies are
clear without ambiguities. Also, these semantic-
FOHDURQWRORJLHVFDQEHXVHGWRHI¿FLHQWO\SURFHVV
the semantic information in Semantic Web and
e-business.
INTRODUCTION

The Extensible Markup Language (XML) (Bray
et al., 2004) developed by the World Wide Web
Consortium (W3C) has recently emerged as a new
standard for data representation and exchange on
the Internet. However, the information exchange
based on XML is at the syntactic level (Garshol
& Moore, 2004). Nowadays, how to process and
exchange semantic information becomes very
important. Semantic Web and e-business are two
important applications which need to process the
2305
A Basis for the Semantic Web and E-Business
semantic information. Semantic Web (Lee, 1999)
means that the Web pages are annotated with the
concepts (terms and relationships) from sharing
ontologies; because Web information refers to the
sharing ontologies, computers can automatically
understand and process the semantic information.
Similarly, when different partners (agents) of e-
business refer to the sharing concepts in ontolo-
gies, they can semantically communicate with
each other. This is a semantic e-business which
is different from the traditional e-business. To
process the semantic information, the traditional
e-business is a person-to-person communication;
now with ontologies the semantic communica-
tion of e-business partners is an agent-to-agent
communication.
It can be seen that ontologies play a core role
in processing semantic information. An ontology

GH¿QHVWKHEDVLFWHUPVDQGUHODWLRQVKLSVFRPSULV-
ing the vocabulary of a topic area, as well as the
rules for combining terms and relationships to
GH¿QHH[WHQVLRQVWRWKHYRFDEXODU\*UXEHU
+RZ WR RUJDQL]H RQWRORJLHV DQG FOHDUO\ GH¿QH
the semantics in ontologies are very important.
Presently, the ontologies are built by different
organizations for their own purposes, therefore we
need to effectively organize different ontologies
together with hierarchies, then the concepts of
WKHRQWRORJLHVFDQEHHI¿FLHQWO\XVHGWRDQQRWDWH
Web pages and e-business agents, and semantic
LQIRUPDWLRQFDQ EHHI¿FLHQWO\SURFHVVHGEDVHG
on the well-organized ontologies.
7R GH¿QH RQWRORJLHV ontology languages
are required. Ontolingua (Gruber, 1992) is an
ontology interchange language which was pro-
posed to support the design of ontologies. Loom
(MacGregor, 1991), a knowledge representation
system, is used to provide deductive support. We
will further introduce the XML-based ontology
ODQJXDJHVLQWKH³%DFNJURXQG´VHFWLRQ
I n t h i s c h a p t e r, w e p r o p o s e h i e r a r c h i e s t o e f f e c -
tively organize ontology languages and ontologies
DQGGLVFXVVKRZWRHI¿FLHQWO\SURFHVVVHPDQWLF
information in Semantic Web and e-business.
The rest of this chapter is organized as follows.
,QWKH³%DFNJURXQG´VHFWLRQZHLQWURGXFHWKH
background and the motivation of this chapter. In
WKH³2QWRORJ\/DQJXDJH2UJDQL]DWLRQ´VHFWLRQ

the hierarchy to organize ontology languages is
proposed. We propose the hierarchy to organize
RQWRORJLHVDQGGLVFXVVKRZWRUHVROYHWKHFRQÀLFWV
in the RQWRORJ\KLHUDUFK\LQWKH³%XLOGLQJ2QWRO-
RJ\6\VWHP´VHFWLRQ+RZWRHI¿FLHQWO\SURFHVV
the semantic information in the Semantic Web
DQG HEXVLQHVV LV GLVFXVVHG LQ WKH ³6HPDQWLF
Information Processing in the Semantic Web and
(%XVLQHVV´VHFWLRQ,QWKH³&RQFOXVLRQ´VHFWLRQ
we summarize this chapter.
BACKGROUND
Some comparisons have been done to compare
different ontology languages. Although XML(S)
has no semantics, it may help bootstrap the de-
velopment of content and tools for the Semantic
Web (Gil & Ratnakar, 2002). Another comparison
(Gomez-Perez & Corcho, 2002) about ontology
languages is from three aspects, that is, (1) general
issues (partitions and documentation), (2) attri-
butes (instance attributes, class attributes, local
scope, and global scope), and (3) facets (default
value, type constraints, cardinality constraints,
and documentation). The existing works are
mainly about comparing different ontology lan-
guages, then choosing the best ontology language
to use. Different from the existing works, this
chapter is mainly about how to organize ontol-
ogy languages and ontologies with hierarchies,
therefore we mainly compare the changes of
primitives in different ontology languages. From

WKHVHFKDQJHVZHFDQ¿QGWKHFKDQJHWUHQGVRI
ontology langu ages, t hen it is motivat ed, that is, it
is very important to effectively organize different
ontology languages.
The Simple HTML Ontological Extensions
6+2(/XNH+HÀLQH[WHQGV+70/
with machine-readable knowledge annotated,
2306
A Basis for the Semantic Web and E-Business
thus the implicit semantic information can be
discovered by a computer. Although SHOE has
the XML version, it is not based on the Resource
Description Framework (RDF) (Lassila & Swick,
2004) and RDF Schema (RDFS) (Brickley &
Guha, 2004).
RDF (Lassila & Swick, 2004) is a standard
ODQJXDJHRI:&IRUGH¿QLQJRQWRORJLHV5')
GH¿QHVDVLPSOHPRGHOIRUGHVFULELQJUHODWLRQ-
ships among resources in terms of properties and
YDOXHV$UHVRXUFHUHSUHVHQWVDQ\WKLQJVSHFL¿HG
Figure 1. The hierarchy of the RDF-based ontology languages
Table 1. Some primitives of RDF and RDFS
RDF
RDFS
DAML OIL
DAML+OIL
OWL
Cat-
egory
Primitives Comment

RDF
rdf:ID Used to identify a class or property or any other resources
rdf:resource
Used to refer to a resource; a resource represents anything
VSHFL¿HGE\D85,
rdf:Property
7RGH¿QHDSURSHUW\WKH¿UVWOHWWHURIDSURSHUW\,'LVLQ
lower case
rdf:Bag An unordered collection (set) of members
rdf:Seq An ordered collection (set) of members
rdf:Alt A collection (set) of alternatives of members
RDFS
rdfs:Class 7RGH¿QHD&ODVVWKH¿UVWOHWWHURIDFODVV,'LVLQFDSLWDO
rdfs:label To provide a human-readable version of a resource name
rdfs:comment To provide a human-readable description
rdfs:domain To restrict the domain of a property
rdfs:range To restrict the range of a property
rdfs:subClas-
sOf
To indicate the specialization of a class
rdfs:subProper-
tyOf
To indicate the specialization of a property
rdfs:Container Super class of rdf:Bag, rdf:Seq and rdf:Alt
2307
A Basis for the Semantic Web and E-Business
E\ D XQLIRUP UHVRXUFH LGHQWL¿HU 85, /HH
Fielding, & Masinter, 1998). Properties are the
attributes of resources, which have either atomic
entities (strings, numbers, etc.) or other resources

as their values. For a person to understand the
semantics of a sentence, a sentence is organized
in a subject-verb-object (SVO) form. Similarly, the
fundamental design pattern of RDF is to structure
data as resource-property-resource triples. Here,
resource can represent both subject and object
in the SVO form, while property (relationship
between resources) represents the verb in the
692IRUP7KXVWKH5')¿OHVFDQEHSURFHVVHG
semantically. An RDF model can be represented
in three ways, namely, graph syntax, triple syntax,
and RDF/XML syntax. In this chapter, we focus
on the XML representation of RDF.
RDF organizes information in the SVO form,
EXWLWGRHVQRWGH¿QHWKHPDQ\VWDQGDUGprimitives
(see Table 1) required to construct ontologies.
Thus, RDFS (Brickley & Guha, 2004) is created
to provide some more basic primitives, such as
³VX E &OD V V 2I ´ D Q G³V XE3 U RS H U W \ 2 I ´ WR U H SUH V H QW 
the relationships between classes or properties).
More semantic-rich primitives are added into
the successors of RDFS, namely, U.S. Defense
Table 2. Primitive differences among DAML, OIL, DAML+OIL, and OWL
OIL DAML DAML+OIL OWL Comment
(1) Primitives included in all the four languages
Class Class Class Class
used to
GH¿QHFODVV
inverseRelation-
Of

inverseOf inverseOf inverseOf
if P1(x,y)
then P2(y,x)
FunctionalProp-
erty
UniquePro-
perty
UniqueProperty
Function-
alProperty
if P(x,y) and
P(x,z) then y=z
(2) Primitives not included in OIL, but included in the other three languages
sameCl-
assAs
sameClassAs
equiva-
lentClass
C1 = C2
sameProp-
ertyAs
samePropertyAs
equiva-
lentProperty
P1 = P2
(3) New primitives added in DAML+OIL, used by OWL
ObjectProperty
Object-
Property
relates

Resource to
Resource
DatatypeProp-
erty
Dataty-
peProperty
relates Re-
source to Literal
or data type
(4) OIL primitives not used by DAML+OIL, but used by OWL
SymmetricProp-
erty
Symmet-
ricProperty
if P(x, y),
then P(y, x)
2308
A Basis for the Semantic Web and E-Business
Advanced Research Projects (DARPA) DARPA
Agent Markup Language (DAML) (Popp, 2000),
Ontology Inference Layer (OIL) (Horrocks et al.,
2001), DAML+OIL (Connolly et al., 2001a), and
Web Ontology Language (OWL) (Harmelen et
al., 2004).
DAML (Popp, 2000), which is funded by
DARPA aims at developing a language to fa-
cilitate the semantic concepts and relationships
understood by machines. The DAML language
is based on RDF and RDFS.
OIL (Horrocks et al., 2001), from the On-To-

K n owle d ge P roj ec t , i s a n on t olog y r ep r es e nt a t ion
language that extends RDF and RDFS with ad-
ditional language primitives not yet presented in
RDF and RDFS.
Now the latest extension of DAML is
DAML+OIL (Connolly et al., 2001b), which has
some important features of OIL imported into
DAML. Presently, DAML+OIL is evolving as
OWL (Harmelen et al., 2004), and OWL is being
promoted as the Web ontology language of W3C.
OWL is almost same as DAML+OIL, but some
primitives of DAML+OIL are renamed in OWL
for more easily understanding.
,QWKH³+LHUDUFK\DQG3ULPLWLYHVRI5')DQG
RDFS-Based Ontology Languages” section, we
illustrate the hierarchies of the RDF-based ontol-
ogy languages, and we compare the primitive
differences among different ontology languages.
Note that a primitive is a basic term in ontology
ODQJXDJHVWKDWLVXVHGWRGH¿QHRQWRORJLHV,QWKH
³ 0R W LY D W LR Q´ V H FW L R QZHL Q W U R GXF H W KH P RW L YDW L RQ 
of this chapter.
Hierarchy and Primitives of RDF and
RDFS-Based Ontology Languages
RDF and RDFS are the ground of DAML, OIL,
DAML+OIL, and OWL. In Table 1, we list some
primitives of RDF and RDFS.
5')DQG5')6GH¿QHVRPHEDVLFSULPLWLYHV
and these primitives are not capable of describing
many other important concepts and relationships,

for example, equivalentClass, therefore DAML,
OIL, DAML+OIL, and OWL extend RDF and
RDFS by adding some new primitives. In Table
2, we compare the primitive differences among
DAML, OIL, DAML+OIL, and OWL. We sum-
marize the differences into several cases, and for
each case, we only list a few primitives which
satisfy this case. Cases: (1) primitives included in
all the four languages; (2) primitives not included
in OIL, but included in the other three languages;
(3) new primitives added in DAML+OIL, used
by OWL; and (4) OIL primitives not used by
DAML+OIL, but used by OWL. The four dif-
ferent cases indicate the primitive relationships
among different ontology languages.
1RZZHGLVFXVV KRZWRGH¿QHDQRQWRORJ\
based on ontology languages.
Example 1. Consider a simple Person ontology
VKRZQLQ)LJXUH7KHVWDUWWDJ³UGI5')!´DW
OLQHDQGWKHHQGWDJ³UGI5')!´DWOLQH
show that this ontology complies with the RDF
syntax. Lines 1-4 specify some XML namespace
declarations (Bray, Hollander, & Layman, 1999),
WKHQZHFDQXVH³UGI´WRUHIHUWRWKHSULPLWLYHVGH-
¿QHGLQWKH85/³KWWSZZZZRUJ
rdf-syntax-ns#” (similarly for other namespaces).
7KHQDPHVSDFH³[VG´DWOLQHLVXVHGWRUHIHU
to XML Schema in which some data types are
GH¿QHG/LQHVGH¿QHDFODVV³3HUVRQ´XVLQJ
WKHSULPLWLYHRI2:/³RZO&ODVV´$OVRZHFDQ

VHHIURPOLQHVDQGWKDWSULPLWLYHV³ODEHO´
DQG³FRPPHQW´DUHIURP5')6WKHUHIRUHWKHUH
LVDQDPHVSDFH³UGIV´EHIRUH³ODEHO´DQG³FRP-
PHQW´WKDWLV³UGIVODEHO´DQG³UGIVFRPPHQW´
/LQHV  GH¿QH D GDWD W\SH SURSHUW\ ³RZO
'DWDW\SH3URSHUW\´ ³RI¿FHBSKRQH´ VLPLODUO\
ZH FDQ GH¿QH RWKHU SURSHUWLHV IRU SHUVRQ IRU
H[DPSOH³QDPH´DQGVRIRUWK/LQHVGH¿QH
D SU R SH U W \ ³F R QW D F W B QX PEH U ´ ZK L FK LVH T X LY DO H QW 
WRWKH³KRPHBSKRQH´
5HPDUN7KH³3HUVRQ´DWOLQHLVDQRULJLQDO
GH¿QLWLRQZKLOHWKH³3HUVRQ´DWOLQHKDVDKDVK
PDUN³´EHIRUHZKLFKPHDQVWKH³3HUVRQ´DWOLQH
10 is a reference. There is no namespace (URL)
2309
A Basis for the Semantic Web and E-Business
EHIRUH³3HUVRQ´DWOLQHEHFDXVH³3HUVRQ´LV
GH¿QHGLQWKHVDPH¿OHDV³RI¿FHBSKRQH´%H-
FDXVHWKH³VWULQJ´GDWDW\SHLVGH¿QHGLQ;0/
6FKHPDQRWLQWKHVDPH¿OHDV³3HUVRQ´DQGWKH
QDPHVSDFHRI;0/6FKHPDLV³[VG´WKHUHLVDQ
³[VG´EHIRUH³VWULQJ´DWOLQH
5HPDUN7KH¿UVWFKDUDFWHURIDFODVVQDPH
LVLQXSSHUFDVHIRUH[DPSOHWKH³3´LQ³3HUVRQ´
LVLQXSSHUFDVH7KH¿UVWFKDUDFWHURIDSURSHUW\
QDPHLVLQORZHUFDVHIRUH[DPSOHWKH¿UVW³R´
LQ³RI¿FHBSKRQH´LVLQORZHUFDVH
A conceptLVDWHUPGH¿QHGLQRQWRORJLHVZKLFK
includes both the resources (entities) and the prop-
HUWLHVUHODWLRQVKLSV)RUH[DPSOHWKH³3HUVRQ´

³RI¿FHBSKRQH´DQG³FRQWDFWBQXPEHU´DIWHUWKH
³UGI,'´LQ)LJXUHDUHDOOFDOOHGFRQFHSWV
Motivation
Figure 1 shows the hierarchy of different ontology
languages, and from Example 1, we know that
the primitives of RDF, RDFS, and OWL are used
WRJHWKHUWRGH¿QHDQRQWRORJ\2:/LVDUHSODFH-
ment of DAML, OIL, and DAML+OIL, therefore
the primitives of DAML, OIL, and DAML+OIL
GR QRWDSSHDU LQ WKH GH¿QLWLRQ RI DQ RQWRORJ\
ZKLFK LV GH¿QHG ZLWK 2:/ +RZHYHU WKLV LV
RQO\DQH[DPSOHZKLFKLVGH¿QHGQRZZLWKWKH
OWL. In practice, before OWL appears, a lot of
RQWRORJLHVKDYHDOUHDG\EHHQGH¿QHGZLWKRQWRO-
ogy languages DAML, OIL, and DAML+OIL.
&DQ ZH VWLOO XVH WKRVHRQWRORJLHV GH¿QHG ZLWK
DAML, OIL, and DAML+OIL in the current
OWL environment? The answer is yes, in this
FKDSWHUZHFOHDUO\GH¿QHWKHKLHUDUFK\DQGWKH
relationships among different ontology languages,
and this hierarchy enables us to automatically use
WKHSUHYLRXVRQWRORJLHVGH¿QHGZLWK'$0/2,/
and DAML+OIL.
The second problem about the ontology lan-
guage is that the primitives from different ontol-
RJ\ODQJXDJHVVKRXOGEHXVHGWRJHWKHUWRGH¿QH
the ontology. For example, from Example 1 and
)LJXUHZHFDQVHHWKDW³UGI,'´³UGIVODEHO´
DQG³RZO&ODVV´DUHXVHGWRJHWKHUWRGH¿QHWKH
c o n c e p t s i n o n t o l o g i e s . H o w e v e r , i t w i l l b e a b u r d e n

Figure 2. A Person_Ontology represented using OWL language
<rdf:R DF xmlns:rdf =" />xmlns: rdfs=" />xmlns:owl=" http://www. w3. org/2002/07/owl#"
xmlns:xsd=" MLSchema#">
<owl:Class rdf:ID="Person">
<rdfs:label>Person</rdfs:label>
</owl:Class>
<owl:DatatypeProperty rdf:ID
<rdfs: domain rdf:resource="#Person" />
<rdfs: range rdf: resource="xsd#stri ng"/>
</owl: DatatypeProperty>
<owl:DatatypeProperty rdf:ID="contact_number" >
e" />
</owl: DatatypeProperty>
…
…
</rdf:R DF>
2310
A Basis for the Semantic Web and E-Business
for the ontology designer to bear in mind which
ontology language the primitives are exactly from.
Can we just use single namespace to refer to all
the primitives, and the system can automatically
translate the namespace to the proper namespaces,
IRUH[DPSOHWKHRQWRORJ\GHVLJQHUFDQXVH³RZO´
as the namespace to refer to all the primitives,
WKDWLV³RZO,'´³RZO,'´DQG³RZO&ODVV´DQG
the system can automatically translate them back
WR³UGI,'´³UGIVODEHO´DQG³RZO&ODVV´"7KH
DQVZHULV\HV,QWKLVFKDSWHUZHFOHDUO\GH¿QH
the relationships of the primitives in different

o nt olog y l an g u a ge s, a nd t h is w i ll he lp to t r a n sla t e
the single namespace to the proper namespaces.
,QWKLVZD\WKHHI¿FLHQF\RIWKHRQWRORJ\GHVLJQ
can be improved.
Moreover, the ontologies are presently de-
signed by different organizations for their own
purposes. If we can organize all of these ontologies
in an integrated environment, we can improve the
usability of ontologies, and the semantic informa-
tion in the Semantic Web and e-business can be
SURFHVVHGPRUHHI¿FLHQWO\
ONTOLOGY LANGUAGE
ORGANIZATION
In this chapter, we mainly focus on the theoretical
analysis of organizing ontology languages and
ontologies with hierarchies, so that this method
can be widely used to organize different ontology
languages and ontologies, though we implement
a prototype tool to organize ontology languages
DQGRQWRORJLHVLQWKH³$UFKLWHFWXUH´VHFWLRQ
There already exist practical tools, for example,
Daml2owl (Amin & Morbach, 2005), and so forth,
that allow the translation from one language to
another, but these tools are not general ones.
They can only be used to translate between two
V S HFL ¿ FR Q WROR J \ OD Q J X D J HV KRZHYH URX U P HW K R G 
is a general one, which can be used to organize
all the existing ontology languages as well as
the future coming ontology languages. This is
WKHPRVWLPSRUWDQWEHQH¿WRIRXUDSSURDFKRYHU

prior works.
,QWKH³2SHUDWLRQVWR2UJDQL]H2QWRORJLHV´
VHFWLRQ ZH GH¿QH VRPH RSHUDWLRQV ZKLFK FDQ
be used to describe the relationships among
the primitives in different ontology languages.
%DVHGRQWKHVHRSHUDWLRQVLQWKH³(DV\8VHRI
Ontology Languages” section, we show how to
DXWRPDWLFDOO\XVHWKHH[LVWLQJRQWRORJLHVGH¿QHG
with DAML, OIL, and DAML+OIL and show
how to automatically translate the namespace to
the proper namespaces.
Operations to Organize Ontologies
We use the following operations to describe the
relationships among the ontology languages and
the primitives in ontology languages, that is,
inheritance, block, atavism, and mutation.
:HXVH³JPRH´DVWKHQDPHVSDFHEHIRUHHDFK
RSHUDWLRQ³JPRH´UHSUHVHQWV³*HQHWLF0RGHOIRU
Ontology (language) Engineering”, because these
operations are borrowed from genetics.
1. Inheritance
The inheritance relationships of ontology
languages can be seen in Figure 1. RDFS inherits
RDF; DAML and OIL inherit RDFS; DAML+OIL
inherits both DAML and OIL; and OWL inherits
DAML+OIL. The following example shows how
WRXVHWKHLQKHULWDQFHRSHUDWLRQWRGH¿QHWKHUHOD-
tionships between two ontology languages.
Example 2. A s w e k n o w, DA M L + O I L i n h e r i t s
both DAML and OIL. We use the inheritance

operation shown in Figure 3 to indicate the inheri-
tance relationship. With the inheritance operation,
we need not copy the primitives in DAML and
OIL into DAML+OIL. In DAML+OIL we only
QHHGWRGH¿QHWKHQHZSULPLWLYHVZKLFKFDQQRW
be inherited from DAML and OIL.
2311
A Basis for the Semantic Web and E-Business
2. Block
It is not enough to indicate the relationships
among ontology languages with the inheritance
operation only. Some primitives in previous on-
tology languages are not used by later ontology
languages. We need to use the block operation to
UHÀHFWWKLVUHODWLRQVKLS7KHIROORZLQJH[DPSOH
shows how to use the block operation.
Example 3. From Figure 1, we know that
DAML+OIL inherits the primitives in both DAML
D QG2,/EXWWKHSU L PLW LYH³6\ P PHWU LF3 URSHU W \´
of OIL is not used by DAML+OIL (see Table 2).
We can use the block operation to indicate that
this primitive is not used by DAML+OIL.
3. Atavism
We found that some primitives blocked by
the child ontology languages are reused by the
descendant ontology languages. To process this
kind of relationships we borrow the atavism
mechanism in genetics. Atavism means that the
characteristics of the grandparent do not appear at
the child generation, but appear at the grandchild

generation or the offspring of grandchild. We use
an example to show how to use the atavism opera-
tion to process the relationships of the primitives
in ontology languages.
Example 4.7KH³6\PPHWULF3URSHUW\´RI
OIL is blocked by its child DAML+OIL, but the
grandchild OWL again includes this primitive
(see Table 2). Without our atavism operation,
WKH GH¿QLWLRQ RI ³6\PPHWULF3URSHUW\´ RI 2,/
has to be copied into OWL, but with our atavism
operation we only need to indicate in OWL that
WKH³6\PPHWULF3URSHUW\´LQ2:/LVDQDWDYLVP
RI WKH ³6\PPHWULF3URSHUW\´ LQ 2,/ ZKLFK LV
shown in Figure 5.
)LJXUH'H¿QLWLRQRI'$0/2,/EDVHGRQLQKHULWDQFHRSHUDWLRQ
Figure 4. The use of block operation in DAML+OIL
)LJXUH'H¿QLWLRQRI³6\PPHWULF3URSHUW\´LQ2:/EDVHGRQDWDYLVP
<rdf:RDF xmlns:daml=" /> xmlns:oil="
gmoe:inheritance="daml, oil">
…
…
</rdf:R DF>
<gmoe:block rdf:resource="oil#SymmetricProperty"/>
<gmoe: atavism rdf:resource=" oil#SymmetricProperty"/>
2312
A Basis for the Semantic Web and E-Business
 0XWDWLRQRURYHUULGHRUUHGH¿QH
Furthermore, we use the mutation operation to
describe the relationship that two primitives have
the same name, but they have different semantics.

See the following example.
Example 5.5')6GH¿QHVSULPLWLYH³&ODVV´
DQG2:/DOVRGH¿QHVSULPLWLYH³&ODVV´7KRXJK
WKHWZR³&ODVVHV´KDYHWKHVDPHQDPHWKH\KDYH
GLIIHUHQWVHPDQWLFVWKH³&ODVV´LQ2:/SHUPLWV
J UH DWH UH[SUH VVLYHQ HVV WK DQWKH ³&OD VV´L Q5 ')6 
7KXVWKHSULPLWLYH³&ODVV´LQ2:/PXWDWHVWKH
³&ODVV´LQ5')6VHH)LJXUH
Easy Use of Ontology Languages
,QWKH³2SHUDWLRQVWR2UJDQL]H2QWRORJLHV´
section, we discuss how to describe the relation-
ships among the primitives in different ontology
languages based on different operations. Based
RQWKHUHODWLRQVKLSGHVFULSWLRQLQ³2SHUDWLRQVWR
Organize Ontologies,” we can process the follow-
LQJWZRSUREOHPVHI¿FLHQWO\DXWRPDWLFDOO\
XVHWKHH[LVWLQJRQWRORJLHVGH¿QHGZLWKRQWRORJ\
languages DAML, OIL, and DAML+OIL; and
(2) automatically translate single namespace to
proper namespaces.
 8VLQJ RQWRORJLHV GH¿QHG ZLWK RQWRORJ\
languages DAML, OIL, and DAML+OIL
Before OWL appeared many ontologies have
EHHQGH¿QHG EDVHG RQ WKH RQWRORJ\ ODQJXDJHV
DAML, OIL, and DAML+OIL. Now DAML,
OIL, and DAML+OIL are being replaced by
OWL, but if we can automatically translate all
WKH RQWRORJLHV GH¿QHG ZLWK '$0/ 2,/ DQG
'$0/2,/WRWKHRQWRORJLHVGH¿QHGZLWK2:/
we can save a lot of time in building new ontolo-

gies based on OWL.
Because we have described the relationships
among different primitives in different ontology
languages, we can automatically build a mapping
between different primitives, thus the ontologies
GH¿QHGZLWK'$0/2,/DQG'$0/2,/FDQ
be automatically translated to the ontologies de-
¿QHGZLWK2:/:HXVHWKHIROORZLQJH[DPSOH
for illustration.
Example 6. Suppose that there is an ontology
LQZKLFKDFRQFHSWLVGH¿QHGXVLQJWKHSULPLWLYH
³GDPOLQYHUVH2I´LQ'$0/%HFDXVH2:/in-
herits DAML, and there are no changes for this
primitive in DAML and OWL (see Table 2), we
FDQGLUHFWO\WUDQVODWH³GDPOLQYHUVH2I´WR³RZO
LQYHUVH2I´)XUWKHUPRUHLIDFRQFHSWLVGH¿QHG
ZLWKWKHSULPLWLYH³RLO6\PPHWULF3URSHUW\´LQ
2,/ZHFDQWUDQVODWHLWWR³RZO6\PPHWULF3URS-
HUW\´EHFDXVHWKH³6\PPHWULF3URSHUW\´LQ2:/
LV DQ DWDYLVP RI WKH ³6\PPHWULF3URSHUW\´ LQ
OIL. There are no primitives in DAML, OIL, and
DAML+OIL, which are blocked or overrided in
OWL, therefore we need not consider these two
RSHUDWLRQVLQWUDQVODWLQJWKHRQWRORJLHVGH¿QHG
with DAML, OIL, and DAML+OIL to ontologies
GH¿QHGZLWK2:/
With this technique, we can automatically
XVHDOOWKHH[LVWLQJRQWRORJLHVZKLFKDUHGH¿QHG
using the ontology languages DAML, OIL, and
DAML+OIL. Therefore less effort will be paid

to build new ontologies based on OWL.
)LJXUH'H¿QLWLRQRI³&ODVV´LQ2:/EDVHGRQPXWDWLRQRSHUDWLRQ
<gmoe: mutation rdf:resource="rdfs#Class"/>
2313
A Basis for the Semantic Web and E-Business
2. Using single namespace to refer to all primi-
tives in different ontologies
From Example 1 and Figure 2, we know that
the primitives in RDF, RDFS, and OWL should
EHXVHGWRJHWKHUWRGH¿QHDQRQWRORJ\,WZLOOEHD
burden for the ontology designer to bear in mind
where each primitive exactly comes from. Based
on the organization of ontology languages in the
³2SHUDWLRQV WR 2UJDQL]H 2QWRORJLHV´ VHFWLRQ
we can use single namespace to refer to all the
SULPLWLYHVGH¿QHGLQGLIIHUHQWRQWRORJ\ODQJXDJHV
and we can automatically translate the single
namespace to the proper namespaces.
Example 7.)RUWKHSULPLWLYHV³,'´³ODEHO´
DQG³&ODVV´WKHRQWRORJ\GHVLJQHUFDQXVHWKHP
ZLWKWKHVDPHQDPHVSDFH³RZO´WKDWLV³RZO,'´
³RZOODEHO´DQG³RZO&ODVV´$VZHNQRZIURP
Figure 1, OWL inherits RDF and RDFS, thus we
FDQVHDUFKWKH³,'´SULPLWLYHERWWRPXSWKDWLV
VHDUFK2:/¿UVWO\WKHQ5')6DQGWKHQ5')
7KH³,'´SULPLWLYHLVIRXQGLQ5')WKHUHIRUHZH
FKDQJHWKH³RZO,'´WR³UGI,'´6LPLODUO\WKH
³RZOODEHO´ZLOOEHWUDQVODWHGWR³UGIVODEHO´ZKHQ
searching the ontology languages bottom up. For
WKHSULPLWLYH³&ODVV´LWLVGH¿QHGLQERWK5')6

DQG2:/DQGWKH³&ODVV´LQ2:/LVDPXWDWLRQ
RIWKH³&ODVV´LQ5')6:HZLOOXVHWKH³RZO
&ODVV´UDWKHUWKDQWKH³UGIV&ODVV´EHFDXVHWKH
³&ODVV´LQ2:/LVDQPXWDWLRQDQGLWLVWKHODWHVW
RQHDQGDFWXDOO\DOOWKHRQWRORJLHVDUHGH¿QHG
ZLWK³RZO&ODVV´UDWKHUWKDQ³UGIV&ODVV´
The number of primitives in ontology lan-
JXDJHVLVOLPLWHGDQGWKHLUUHODWLRQVKLSVDUH¿[HG
therefore based on the organization and relation-
ship descriptions of ontology languages in the
³2SHUDWLRQVWR2UJDQL]H2QWRORJLHV´VHFWLRQWKH
WUDQVODWLRQVLQ³(DV\8VHRI2QWRORJ\/DQJXDJHV´
section can be done without ambiguities.
BUILDING ONTOLOGY SYSTEM
,QWKH³2QWRORJ\/DQJXDJH2UJDQL]DWLRQ´VHFWLRQ
we describe the hierarchy of different ontology
languages based on the inheritance, block, ata-
vism, and mutation operations. These operations
can also be applied to the ontology building.
As we know, the ontologies are now built by
different organizations for their own purposes. It
is important to organize these ontologies together
in an integrated environment, then the semantic
information in one domain is more complete. Only
ZKHQWKHVHPDQWLFVLQRQHGRPDLQDUHDOOGH¿QHG
clearly, it is true that the semantic information in
Semantic Web and e-business can be processed
correctly.
,QWKLVVHFWLRQZH¿UVWO\GLVFXVVKRZWRRUJD-
nize ontologies based on the operations discussed

LQ³2QWRORJ\/DQJXDJH2UJDQL]DWLRQ´$OVRZH
summarize the guidelines for organizing ontolo-
gies, that is, different information should be put at
different hierarchies of ontologies. Furthermore,
different from primitives in ontology languages,
which will not change, the concepts in ontologies
will change. When inserting or deleting a con-
cept in ontologies, we should keep the ontologies
consistent, otherwise it will hurt the ontologies
to provide sharing information. Hence, we also
GLVFXVVKRZWRUHVROYHWKHFRQÀLFWVLQRQWRORJ\
organizations.
Architecture
To make easy the use of inheritance, block,
atavism, and mutation operations discussed in the
³2QWRORJ\/DQJXDJH2UJDQL]DWLRQ´VHFWLRQZH
build a graphical tool to implement these opera-
tions in ontology building. This tool can be applied
to ontologies as well as ontology languages.
Example 8. Figure 7 shows that, there exists
a Person_Ontology, and we need to build a Stu-
GHQWB2QWRORJ\7KHQLQWKH³3DUHQW2QWRORJLHV´