Page 219 - Building Digital Libraries

CHAPTER 9

allowing the service to create specialized indexing, facets, and ancillary
services around the content being mined. Outside the library community,
search engines are excellent examples of a managed search. Search engines
utilize crawlers to capture and index websites. One of the direct benefits of
this approach is that managed systems provide total control over issues like
relevancy, indexing, and query languages. At the same time, the primary
disadvantage is that unless the content is indexed, it cannot be found. Within
the library community, this is very much how third-party comprehensive
discovery tools work. Tools like ProQuest’s Summon or EBSCOhosts’s
EDS index large sets of library content. These tools provide a high level of
relevance and speed to the user, but the search is limited to the content that
has been indexed.

Federated Search and

Digital Libraries

How federated search fits into the larger picture of digital libraries may not
be readily apparent at first glance. For most organizations, how a federated
search product will interact with their digital repository infrastructure is
likely the farthest thing from their minds—but by neglecting this consid-
eration, they put their digital collections at a distinct disadvantage and
ultimately shortchange their users’ overall experience. Federated search tools
are often thought of only in relation to electronic serial content. Organiza-
FIGURE 9.3
Managed Search Diagram tions purchase access to thousands of journals online, and federated search
tools have been springing up since 2000 that promise to provide a single
search interface for all of these resources. However, as a digital repository
program develops, collections and projects will often become siloed. The
development of these information silos is often unintended, and often
occurs because of funding sources or software platform choices. For exam-
ple, an organization may use DSpace for their electronic theses collection,
use Sufia to house their archival image collection, and use ArchivesSpace
to house their EAD content. Separated by software platform, each of these
projects represents three separate interfaces that a user would need to query
in order to locate content from these digital projects. These separate inter-
faces marginalize these collections by placing them outside of the organiza-
tion’s mainstream query interface. What’s more, this would represent three
new query interfaces existing on top of a library’s traditional search tools
like the integrated library system and electronic journal pages. In all, a user
may have to search 8 to 10 different locations just to cast a net that is broad
enough to query most of the important organizational resources for a topic.
While federated search tools certainly have the ability to provide a more
unified search interface for vended serial content, they are probably best
suited for the unification of discovery for locally developed digital collec-
tions. This is due in large part to the fact that an organization has the ability

204