More Advanced Applications - User-Centered Data Management - page 57

architectures, the lower-level modules respond to a user request by producing an XML document
with the key data required as a response. A higher level subsequently transforms this into a web page
using XSLT, a notation that allows generic descriptions of transformations of XML to other formats,
notably HTML. This multi-layer approach is similar to the presentation-functionality separation
found in traditional pre-web user interface architectures such as Seeheim ( Pfaff and ten Hagen ,
1985 ) and MVC ( Krasner and Pope , 1988 ).
RDF (Resource Description Framework) was originally developed as a way to add semantics to
web pages such as the page author ( Powers, S. , 2003 ; W3C , 2010 ). This is possible to do inside a web
page using the < meta > and < link > tags in the page header, but this means that you have to have the
whole web page to hand and, furthermore, that the information has been entered fully and correctly
by the page author. RDF allows information to be described separate from the page itself, maybe
added by a third party. However, RDF has been adopted more widely as a knowledge representation
framework by assigning URIs to non-web resources such as people, books and countries. (Note,
URIs, not necessarily URLs, uniquely identify the resources but may not point to an actual location
on the web.)
Figure 4.9 shows a short RDF description of this topic. The data is represented a as series
of triples: < subject, predicate, object > , where each triple tells you some property of the subject.
For example, the triple with predicate 'dc:title' has subject (what it is about) 'mc:ucdm' and object
“User-Centered Data Management.” While this is a minimalist representation compared to tables in
a relational database or the hierarchical structure of XML, in fact, anything that can be represented
in either tables or XML can also be described (albeit verbosely) in triples.
@prefix mcb: <http://www.morganclaypool.com/books/> .
@prefix dc: <http://purl.org/dc/elements/1.1/> .
mcb:ucdm dc:title "User-Centered Data Management" .
mcb:ucdm dc:author "Giuseppe Santucci" .
mcb:ucdm dc:author "Stephen Kimani" .
mcb:ucdm dc:author "Alan Dix" .
mcb:ucdm dc:author "Tiziana Catarci" .
Figure 4.9: Describing this topic in RDF (N3 notation).
While XML requires that all the data are lumped together, RDF allows knowledge to be
represented in fragments and pieced together as needed. For example, Figure 4.10 has some additional
information about the topic, and this could be located in a completely different repository. This is
particularly valuable in the open environment of the web where different sources may bring different
data.
Note that Figure 4.10 also shows the use of 'dc:title' (shorthand for ' http://purl.org/dc/
elements /1.1/ #title ' ) from the Dublin Core ontology. By using a commonly known ontology,
this makes it possible for computer tools to understand the meaning of the properties. Note also that
some of the triples defined simple properties of an entity, such as the name of the topic, while others
denote a relationship between two entities, for example, that the topic belongs to a particular series.