Page 51 - Building Digital Libraries

CHAPTER 3

the library need only create a formal agreement with the entity that main-
tains the data, and this agreement establishes that the data are important and
must be maintained indefinitely. In certain cases, relying on organizations
with an inherent interest in the resource to be responsible for preserving that
access is appropriate as well as cost-effective. Just as it is perfectly reasonable
to trust the Patent Office to provide perpetual access to patents and that the
appropriate departments will maintain court proceedings, entities vested in
the ongoing provision of an information service, such as specialized scien-
tific communities, are often in a better position than the library to provide
access to that service. Having said this, other organizations have different
priorities than libraries, and providing access to such materials from other
access mechanisms can be difficult unless their systems are designed to be
used in this way.

Organizing Content and
Assigning Metadata

An object’s value is ultimately defined by use, and people must find objects
before they can use them. For this reason, metadata must be assigned to
or extracted from acquired objects so they may be searched and relation-
ships between them understood. The first step in developing a workflow
for adding or extracting appropriate access points is to consider what types
of materials will be added, who will use them and how, how they will be
searched, and what the expected size of the repository will be. If the reposi-
tory consists of Portable Document Format (PDF) documents or simple
image files that are uploaded to a server, the processing required will be
substantially different from processing web pages or an eclectic collection of
multimedia resources. Likewise, completeness and consistency of metadata
are far more important for repositories that contain millions of items than
for collections containing only a few thousand items.
A repository’s usefulness depends heavily on how well it associates
similar items and allows users to find what they need. There are many ways
to organize electronic resources, but one of the time-tested ways to provide
access is to create or extract metadata that is embedded in the resource or
stored in a separate record. All repository software relies on metadata to help
users find things—in fact, the traditional library catalog card is metadata
printed on paper.
The quality and completeness of metadata are critical to its useful-
ness. Although it’s administratively convenient to store information in its
original form and rely on automated means to identify access points, this
approach rarely leads to adequate long-term access. Resources—especially
nontextual ones such as those consisting of images, video, or sound—lend
themselves poorly to keyword searching. File formats become obsolete or
require users to install software which may be expensive or difficult to find,

36