Information Technology Reference
In-Depth Information
which have to be discarded by the user without support or with very limited sup-
port of automated tools. Moreover, the dynamic nature of the Web requires that
the user periodically repeats this search-retrieve-filter process to localize new re-
sources of interest and to update previous ones [7].
As the Web is an ever increasing huge information space, a precise search
engine, if existent, is not enough for several user requirements, such as updated in-
formation from economic news or shoppers´ feedback. Beyond accurate and stan-
dardized metadata describing relevant pieces of information, information agents
that continually browse the Web are necessary to search for updated resources of
interest.
Most of the existent documents in the Web present a multimedia, unstructured
or semi-structured nature. Some are rendered dynamically, meaning that their con-
tent is stored inside databases, in which case the content that is located automati-
cally by information agents in the Web are just links and scripts.
The SW can be represented by a stack of systems with seven layers. Basically
the base layers are already well specified and consist of systems to process repre-
sentation schemes for character and resource identification. The following two
layers were developed afterwards the appearance of the Web and are concerned
with metadata languages for resource description, and with semantic statements
about described resources. All these four layers have standard and consolidated
specifications, which can evolve and are discussed within the W3C forum [8]. The
upper four layers are subject of research, application demonstration construction,
and standard submission. They are concerned with: the representation of informa-
tion on object categories and how objects are interrelated, named ontology; the
examination of different ontologies to find new relations among terms and data in
them, according to rules defined to be inferred; and the extent to which the infor-
mation found is both accurate and trustworthy, once machines should be able to
discover relevant and quality content more efficiently.
2.2 Information Retrieval in Wikis
In semantic wikis users edit pages with links forming a network that can be que-
ried. They, in different ways, try to offer a kind of semantic search promised by
the mentors of the SW [5]. It is interesting to highlight that a semantic wiki is a
semantic artifact that allows the creation of content semantically enriched, pref-
erably without a design process beforehand.
All of the known research efforts to create semantic wikis, aggregated in their
kernel, explore semantic representations of information objects through the use of
ontologies.
There is a semantic engine that explores: in-line annotations (metadata), in-
serted during the authoring of the wiki content; or metadata that replaces the wiki
content. These metadata allows the wiki platform to establish and organize inter
related concepts and attributes, which would represent the wiki domain of interest.
Shawn [9], SemperWiki [10], Kaukolu [11], Makna [12], Semantic Media Wiki
[13]and WikSar [14] are examples semantic wikis with in-line annotation insertion.
