Page 174 - Building Digital Libraries

Sharing Data—Harvesting, Linking, and Distribution

common heading in the United States of: Poe, Edgar Allan, 1809–1849.
VIAF, for the first time, provides a way for researchers to visualize the rela-
tionships between these national authority files and create linkages between
these link terms. However, given OCLC’s vast amount of authority and
bibliographic data, the cooperative was able to expand the project beyond
authority data and begin exploring secondary relationships to a person,
place, or subject—examining relationships like coauthors, related works,
and publishers. Like Google’s knowledge graph, VIAF utilizes data from a
wide range of sources to build new knowledge through the relationships
between data.
On its own, this kind of work would be exciting. But it’s the open nature
of linked data and the focus on machine-actionable content that has impli-
cations for the future. While Google
doesn’t share the underlying data that
it uses to build its knowledge graphs,
OCLC does. Looking at the example of
Edgar Allan Poe, VIAF produces a per-
manent URL for the information. When
requested through a browser, the system
generates an HTML representation of
the graph for human consumption. But if
the request is made by a machine, OCLC
makes the VIAF data available in various
data serializations. Here’s a snippet of the
data when the request is for XML data
(see figure 7.2).
VIAF produces a machine-readable
object about Edgar Allan Poe, and in
this object, one sees not only the links
between the different forms of the
names, but bibliographic information,
links to other numeric identifiers in
other systems, and also the secondary
relationships. This is a typical example
of how linked data can be developed,
and the transformative power that it can
provide to libraries. Utilizing this data,
libraries could conceivably build systems
that provide better discovery, or auto-
mated authority control, or that even
provide support for greater international
accessibility—but we don’t do this. As
interesting and exciting as the work of
OCLC’s VIAF is, it is also exceedingly
rare in the cultural heritage community.
As of this writing, a large number of FIGURE 7.2
organizations are experimenting with VIAF XML Representation
159