Experimenting with BIBFRAME: Reports from Early Adopters

Introduction

Experimental Mode: The National Library of Medicine and experiences with BIBFRAME

To date, the work the National Library of Medicine (NLM) has undertaken is realm of enhancing understanding of the BIBFRAME model as produced by the Library of Congress and Zepheira and, more recently, trying to change those models. At NLM, tasks are focused on developing a workable, broadly acceptable BIBFFRAME model in an experimental mode -- not a production mode -- so that community buy-in and tool development can begin in earnest. NLM is are planning to do more practical experimentation, including generating new BIBFRAME data with tools developed by others by Spring 2015; however, any data produced is not likely to be stable or useful beyond providing sample data to support NLM's idea of what the BIBFFRAME model should be/how it should work. 

Exploring BIBFRAME at a Small Academic Library

This presentation traces the development path taken by the Tutt Library at Colorado College as its explores how BIBFRAME and other Linked-data vocabularies can be used in library systems to improve operations and provide superior access and discoverability of the library's collections.

Starting with early experiments modeling bibliographic entities using Google App Engine & Solr, the library moved to using Redis as a bibliographic store for FRBR and BIBFRAME entities. Limitations with Redis lead to further experimental systems with MongoDB and Solr, leading eventually to using Fedora 4 as a Linked Data Platform supported by Elastic Search, Fuseki, and Redis that is being actively developed in the upcoming BIBFRAME Catalog for the Library of Congress. The future plans for the Tutt Library is to use the BIBFRAME Catalog a foundation for a new integrated library catalog and website.

Working with BIBFRAME for discovery and production: Linked data for Libraries/Linked Data for Production

This presentation will describe and provide updates on two collaborative and related linked open data projects Stanford is participating in—Linked Data for Libraries and Linked Data for Production. Both projects make use of BIBFRAME to explore and use linked data in the library environment.

Linked Data for Libraries (LD4L) is a collaborative project of Stanford, Cornell, and Harvard to create a semantic information store model of scholarly resources. The goal is to bring together three large pools of data—bibliographic data (transformed from MARC to BIBFRAME); person data; and various types of usage data—and link them together through an open source ontology and engineering framework, to capture the intellectual value that librarians and other domain experts and scholars add to information resources. Active since January 2014, the project is in its second year of a two year grant.

Linked Data for Production (LD4Prod) is a natural outgrowth of LD4L. The primary goal of this collaboration between 5 academic libraries (Stanford, Cornell, Columbia, Harvard, Princeton) and the Library of Congress is to actively explore metadata creation in the linked data environment using BIBFRAME, and integrating it into our technical services workflows. To that end, we will be defining specifications for tool development and infrastructure, exploring how cataloging rules mesh with BIBFRAME, and creating profiles for various subject domains.