LODLI: Linked Open Data Back-End System for Libraries Interoperability.

Linked Data principles offer significant advantages over current practices when publishing data. Linked Data allows library interoperability by linking to data from other organizations with authoritative data, which enriches library catalog-user search results. This paper describes LODLI, a Linked Open Data Back-End system that we designed and developed to enhance library catalog searches. We integrated our system with the Hebrew University library catalog, HUfind. While our platform can be used as is, it can also be customized by Linked Open Data providers that desire to convert their MARC records into Linked Data information library systems, making their data far more accessible. This research project faced the following challenges: finding the most efficient way to translate binary MARC into MARC records; mapping the MARC records into a variety of information models, such as Dublin Core, FRBR, RDA, OWL and FOAF, while selecting the most appropriate MARC field combinations; and providing links to resources in external datasets using a distance algorithm to identify string similarity. LODLI is a generic system to which additional ontologies can easily be added. We have demonstrated the system with two types of clients: FRBR visualization client and VIAF-extension client.