The last couple of years has seen the emergence of the term Linked Data, which covers the pragmatic implementation of techniques and technologies espoused by the Semantic Web. The term Semantic Web itself, following Sir Tim Berners-Lee’s first using it in 2001, has had somewhat of a bad press, being associated with artificial intelligence and the need for web publishers to add lots of metadata to their pages.
Our guest on the Gang this month, Karen Coyle, has recently published a Library Technology Report - Understanding the Semantic Web: Bibliographic Data and Metadata. She was therefore in an great position to provide an introductory overview of the Semantic Web and how it could apply to the data rich world of libraries. Describing herself as a semantic web opportunists, not a semantic web purist, she was ideally placed to get this month’s conversation going. Gang members Carl Grant and Marshall Breeding soon joined in the conversation, making it clear we were talking about the practical implementation of semantic web technologies, not the big ‘vision’.
Applying the principles of linked data to the bibliographic world leads you away from current practice of libraries creating and/or copying full bibliographic records on their own systems towards linking out to authoritative sources for things such as subjects. The Library of Congress publishing their subject headings as linked data, is an obvious early example of this approach.
Concerns about holding local copies of data, for performance reasons, and how ‘local practices’ could impact the adoption of this very different way of working for libraries, were discussed as some of the aspects of this topic which has the potential to disrupt future library metadata practices.