Page 173 - Building Digital Libraries

CHAPTER 7

shifted to the development and support for linked data and other
semantic principles. So, what is linked data? Very broadly, the
driving principle behind linked data is to use the Web to build
connections and relationships to other information. These rela-
tionships create linking points, and this interlinked data begins
to form what is called a knowledge graph about a concept, topic,
person, or thing.
This is vague, I know—but we see the evidence of these
knowledge graphs and “linked” data all around us. One of the
most visible examples of linked data in action is seen on Google
thousands of times a day. . . . the visual placecard Google gener-
ates around people, places, or things. For example, if one does a
search for “Alan Alda,” Google generates the following content
(see figure 7.1):
One probably sees these kinds of placecards many times each
day, but taking a closer look, one can see how this simple visual
element not only provides simple biographical information about
the topic, but that a graph starts to surface about the topic—link-
ing to people, places, and things that are closely related or have
relationships to the topic. In this case, the graph for Alan Alda not
only includes information about the movies and TV shows that
he’s been in, but it creates relationships to other actors and direc-
tors that are related to him in some way . . . that are linked to him.
Libraries have these same needs. Cultural heritage organi-
zations invest significant resources in developing metadata for
resources. This metadata often includes links to subjects, people,
and places. These access points represent potential relationships,
data that could be mined to develop the same kinds of graphs
that we see being created by Google. And in some places, this
is happening. OCLC, as the largest aggregator of library bib-
liographic and authority data, has been leading these efforts,
creating services and projects that demonstrate the potential that
creating these kinds of relationships can have for improving both
the findability and access to content. This effort is probably best
illustrated by OCLC’s VIAF (Virtual International Authority File)
project and web service.
OCLC’s VIAF project started out as an effort to understand
how the various national controlled vocabularies are related
to each other. When you think about it, it’s a basic findability
problem. National libraries around the world maintain their
own national authorized vocabularies—meaning that the same
individual is likely represented in many different forms across the
FIGURE 7.1 various discovery systems used throughout the world. VIAF was
Google Knowledge-Graph Example a project to look at how these variations could be related together,
to create relationships between different terms that represented
the same person, object, or subject. This way, a search for: 1,
ןלא דגרא ופ 1809–1849 would locate materials under the more
158