Enhancing the ILS with semantic data

From MARC must die!

Jump to: navigation, search

Starting an evolution is probably easier than starting a revolution. So how can we demonstrate the value of Linked Data/Semantic Web/RDF in a library context, without throwing out all the old stuff and starting from scratch? Here is one proposal...

(And it's a wiki, remember? Please feel free to add your own bright ideas! ;-)



Short term goals

  • Enhance the user's experience of the OPAC
  • Demonstrate the value semantic technologies have to offer

Long term goals

  • Demonstrate that we can live happily ever after without the old MARC data


  1. Export MARC data from the ILS, convert them to RDF (this could be really simple at first), and store it in a triplestore
  2. Create interfaces that make it easy (and fun) to enhance the data in the triplestore with...
  3. Use the data in the triplestore to enhance the old OPAC

Convert MARC to RDF

This could be relatively simple to begin with, and then expanded progressively over time. (It should be possible to run complete conversions at any time, so we can continuously refine and improve the mappings over time.)


Enhance the RDF

  • Create good interfaces for enhancing the MARC data with all the types of data we miss today
    • Identifiers
    • Relationships
  • Model properly any interesting entities that are just represented as text strings in MARC, such as
    • Subjects
    • Authors
    • Publishers


Enrich the OPAC

When a record is displayed, make a request with the ID of the record to some script that talks to the triplestore and returns useful data. This could be complete chunks of HTML or JSON/XML-data that is formatted on the ILS side, before it's incorporated into the OPAC.

Personal tools