New Directions in Metadata and Linked Data

Karen Coyle keynoted the track on linked data in libraries.

Karen Coyle

We are mainly presenting text in library catalogs.  In the new world, we will need more data and less text.  We also must not mix different types of data in the same field.  The semantic web is forcing the development of standards for how you describe your data.  We must explain our data if we expect it to be used, and we also need to think about how we structure our data.

A number of registries have been developed to make semantic data available and describe library data, including the Open Metadata Registry, Dublin Core Metadata Terms, Bibliographic Ontology (BIBO), Scholarly Works Application Profile (SWAP), etc.  These allow the community to communicate.  The Library of Congress is making its data available on the web.

We must share our data.  This is beginning to occur in interesting ways.  Here are a few of the organizations involved.  The standard semantic web model is also available.  The semantic web is like the web before Mosaic.  Its goal is to make linked data usable.

Karen’s presentation as well as supporting data are available here.

Sarah Bartlett

Sarah Bartlett, senior analyst at Talis, followed Karen and looked at the place libraries have in a linked data world.  The web has made information far more open and democratic; it was formerly open only to those knowing how to search it.  We now have links between a wide variety of information types; Amazon’s site is a good example of this.

Literary theory helps us to understand how text is produced.  Writers depend on subliminal knowledge, which is impractical to rely on.  Now, technology allows them to focus on linking between varieties of text.  Here are the ways this can occur:

  • Architextuality links text to groups.
  • Intertextuality:  links between textual terms
  • Paratextuality:  elements that lie on the threshhold of the text.
  • Metatextuality:  text takes a relationship to other text.
  • Hypertextuality:  transformation of an entire text into another.

These categories can be used as an ontology.  A linked data approach can help us develop intertextual relationships.

This is the classic question.  Librarians are enthusiasts for linked data and are the main people that understand it.  Libraries remove the need for each user to reconstruct the data to find information related to their interests.  They can also make tools available to help people navigate through the data.Linked data provides a definition of relationships, allowing broad insights to surface rapidly.  We can all contribute to this effort.

Don Hawkins
Columnist, Information Today and Conference Circuit Blog Editor


  1. Tweets that mention New Directions in Metadata and Linked Data -- - December 1, 2010

    […] This post was mentioned on Twitter by SosLab, ITI Conferences. ITI Conferences said: New Directions in Metadata and Linked Data at […]