Page 170 - Building Digital Libraries

Sharing Data—Harvesting, Linking, and Distribution

For the library community, this means learning a new language and
new rules. How much should libraries engage in social networks or work
with these toolkits? What is the price in information currency that they are
willing to pay to aggregate statistics or enable integration with common
social media authentication platforms? What’s more, how much information
should the library share? Should one’s digital library be largely transparent,
encouraging tools like Google and Bing to index content, or should the
community support a more walled-off model that focuses on supporting
access through larger aggregators like the HathiTrust or the Digital Public
Library of America? These are key questions for the community, but ones
that we believe have significant historical precedent, if one looks closely
at the development of OCLC and other similar initiatives. Libraries have
always been strongest and most effective when they reduce the number of
barriers and actively work to promote the sharing and integration of their
content. To this end, digital library developers and managers need to come
to an understanding that library services need to be transparent. This means
that services need to be not only open, but documented. Unlike commercial
information providers, the library is most successful when it gets users to
information in the most unobtrusive way possible. Digital library develop-
ers and managers should work to do a better job of not tripping over their
own technology, and work to find ways to provide more direct paths to
information. In some cases, this may mean looking outside the library com-
munity for usable or compatible technologies, or engaging with standards
bodies outside the library community to develop pathways for libraries to
more easily interact in the new information environment. And we see this
happening with Schema.org and the larger Dublin Core community . . .
communities that don’t primarily serve the library community, but which
have encouraged and made space for members of the library community
to serve and contribute.

Metadata Doesn’t Want to Be Free . . .
If It Did, It Would Be Easy

Concepts can develop into mythologies that take on a life of their own.
“Information wants to be free” has morphed into “metadata wants to be
free,” and within the library community, I believe that it does. Libraries have
invested significant capital and energy to develop protocols and standards
that enable the sharing of bibliographic information. This is why MARC was
developed: so the Library of Congress could share its metadata. And this
2
is what spurred on the development of Z39.50—a need to provide a remote
protocol to share bibliographic data. Since libraries first started creating
electronic bibliographic records in the early 1970s, they have worked to
develop technologies that would simplify the process of sharing metadata.

155