Page 165 - Building Digital Libraries

CHAPTER 6

leverage markup syntax to enable search engines to reach into, and index,
the data in WorldCat.org. More interesting to libraries, however, is that
this data can be mined using XPath and Sax, since this marked-up data is
also embedded within the results page as XHTML. This makes WorldCat
.org a powerful linking service, since the results page can now be utilized
to retrieve URIs to a Works page, to VIAF information about the author, as
well as to a whole host of other information about the record. OCLC has
been out in front of the library community in demonstrating how semantic
data can be made manifest within library catalogs and websites today, with
minimal effort.

Moving Forward
So, what can you do today? Libraries are sitting on a precipice. On the one
hand, our community understands the work that needs to be done, and
in many cases, we have systems that support the ability to begin moving
library data into semantically aware systems. At the same time, we have
mountains of legacy data that need to be reconciled, and currently we lack
the widespread library infrastructure to enable large-production reconcili-
ation projects. It’s a tough place to be. How do we start?
Begin encoding your data now. Libraries and metadata special-
ists must stop thinking about their metadata as a collection
of strings, and must start encoding their data. For a time, this
will be slow going while library infrastructure develops, but
we cannot have a future in which libraries utilize linked data
and linked-data concepts at scale until our metadata shifts
from a series of strings to encoded objects.
Use systems that support heterogenous data profiles. While
MARC may remain the lingua franca for our traditional li-
brary catalogs, the community has an opportunity to move
beyond this and expand our bibliographic data model within
the digital repository space. Libraries should be adopting
systems that allow them to support and utilize a wide range of
data schemas. Developing models, like the Portland Common
Data Model, embrace this approach, and as library systems
continue to develop, this type of heterogeneous data modeling
should become more of the rule—as libraries are able to use
existing ontologies from a wide range of communities to not
only enable better description and discovery, but to provide
more actionable data that supports self-description.

Summary

Sadly, this chapter can only cover a small number of the metadata frame-
works that are currently available to digital library designers. As such, this
chapter focused on the three metadata frameworks that currently are most
150