Information Technology Reference
In-Depth Information
tWo semAntiC Web
ArChiteCtures And
tWo dAtA models
Due to the main activities of collaboration, reus-
ing and sharing digital resources over the World
Wide Web, effective information retrieval must
be an imperative part of collaboration.
As an effort to improve information retrieval,
researchers have endeavored to find more efficient
information organization methods. Their efforts
can be summarized into three major categories of
methods: term lists, classification/categorization,
and relationship groups (Zeng, 2005). With the
realization that these three groups of informa-
tion organization methods did not significantly
improve information retrieval (Smeaton & Berrut,
1996; Voorhees, 1994), a few newer information
organization methods that focus more on rela-
tionships among the information units have been
recently studied.
One noticeable method is the Semantic Web.
The Semantic Web emerged as a dynamic web
for sharing data on the current static web in 1998
(Berners-Lee, 1998). The Semantic Web Archi-
tecture was released in 2000; however, there is
little penetration into current web and informa-
tion systems. There are criticisms (Ian Horrocks,
Bijan Parsia, Peter Patel-Schneider, & Hendler,
2005; Patel-Schneider, 2005) of the Semantic
Web, and those problems will be examined in
this chapter.
This chapter introduces an alternative data
model to cope with some criticisms of the cur-
rent Semantic Web Architecture and presents a
study that explored the alternative data model to
measure user performance.
This chapter is organized as follows: In Section
2.1, we address criticisms of two Semantic Web
Architectures and describe the two data models.
In Section 2.2, we discuss the differences between
the two data models. In Section 2.3 we provide
guidelines for choosing appropriate data models.
Section 2.4 presents user performance using a
Topic Maps-based ontology system; Section 2.5
concludes with future directions in the context
of collaboration.
Problems of the First semantic
Web Architecture
According to Patel-Schneider (2005), the current
architecture (See the left-hand side of Figure 1)
for the Semantic Web has problems when expres-
sive Semantic Web languages such as RDF are
integrated. The Resource Description Framework
(RDF) is a language for representing information
about resources in the World Wide Web. World
Wide Web Consortium supports RDF; however,
RDF is not suitable for the Semantic Web (Patel-
Schneider, 2005). Patel-Schneider criticizes that
RDF is not sufficient to encode complex syntactic
information in triple form.
Due to the limitations of RDF and first-order
logic, ontology engineers and domain experts are
concerned about the difference between users and
domain experts. There is a similar issue between
users and indexers. When an indexer chooses a
term that a user does not utilize, the user has a
hard time finding relevant resources.
The burden of knowing what could link to
what information should be removed from both
domain experts and users (Holm, 2001). The
system must present and process the informa-
tion without requiring it know the complicate
and non-agreed relationships between the data.
Instead, the system can show built-in relation-
ships to users and let them navigate and decide
the relevance of information. Topic Maps (TM)
explicitly show the structure and relationships
among digital resources.
In order to resolve the issue of RDF triple for
the current Semantic Web Architecture, different
data models such as Topic Maps can be used.
Topic Maps do not have triple issues or rules is-
sues. Users can see explicit relationships among
resources. Recent change of the first Semantic
Web Architecture (the right-hand side of Figure
Search WWH ::
Custom Search