The CaMMS Cataloging and Classification Research Interest Group held a program on Sunday, June 24 from 10:30-11:30 featuring:
OCLC FRBR Cluster Analysis
Nathan Putnam, Director, Metadata Quality, OCLC
Metadata Quality and OCLC Research have been evaluating clusters within WorldCat. The goal of this project is to characterize the coherence and completeness of OCLC FRBR clusters given WorldCat’s continued growth and enrichment of bibliographic data.
OCLC currently clusters groups of records in its discovery interfaces to focus results for similar items. Using the Functional Requirements for Bibliographic Records (FRBR) as a guide, a cluster represents a Work and individual records represent different Manifestations. For example, a cluster may contain records for both print and electronic versions of a Work, translations of the same Work, different editions of a Work, or parallel records that use different languages for cataloging the same Work. Groups of clusters are evaluated based on pre-defined categories, e.g. records in the cluster contain 1 personal name and 1 title, or many personal names and 1 title, etc. Given these categories for clusters, a rating is given to the cluster category, thus giving a sense of the accuracy of the cluster. Because the clusters are associated with a Work ID, those with high accuracy ratings are strong candidates to populate the newly defined 758 field for resource identifier. Many catalogers are interested in description at the Work level as cataloging practices shift away from record-based description to entity-based description. There is also a need for Work identifiers in BIBFRAME. A second outcome of the evaluation is to look at potential changes to the algorithm when bad clusters are found.
This presentation will give a high-level overview of the clustering algorithm used in WorldCat, the cluster review process, and next steps for populating Work IDs in WorldCat bibliographic records.
Cataloging Early English Books Online in WorldCat Discovery
Beth Guay, Continuing Resources Librarian, University of Maryland Libraries
The University of Maryland Libraries employs WorldCat Discovery as its discovery service provider and relies on KBART files for e-resources discovery and access. In late 2015, the Libraries initiated a cataloging project designed to ensure the discoverability of equivalent e-versions of its 5,062 cataloged microfilm resources in the series, Early English books, 1475-1640, and hoped to follow with its 41,306 resources in the Early English books, 1641-1700 series. The e-versions are held in ProQuest’s Early English Books Online (EEBO) collection. The project began as an attempt to adapt cataloging workflows to the new environment in which e-resources copy cataloging takes place within discovery system tools rather than MARC record set or individual record downloads into online catalogs. The presentation will cover EEBO’s relationship to scholarship and to cooperative cataloging. Based on the project findings, the presenter will: offer recommendations for maximizing existing bibliographic metadata to improve resource discovery; and open a dialogue with a goal to extend cooperative cataloging of EEBO resources beyond traditional lines.
Dangerous Liaisons: Collaborating with a Donor on Descriptive Metadata for a Digital Project at UNC Charlotte Abstract
Joseph Nicholson, Metadata Librarian, J. Murrey Atkins Library, UNC Charlotte
Involving donors in the creation of metadata for their own collections is a novel way to enrich descriptions of archival resources with voices from the communities that libraries and archives serve. But the logistical difficulties of collecting metadata from donors are considerable. How should donors be trained to create metadata? What are the best tools for gathering their metadata? What do donor metadata workflows look like? There are no clear-cut answers to these questions, and no best practices or widely shared standards to aid archives and libraries considering such a venture.
This presentation will focus on a project involving the digitization of a vast collection of photographs in Special Collections and University Archives at Atkins Library at the University of North Carolina in which the donor and photographer has been closely involved in item-level metadata creation. It will discuss the tools that project staff use to collect metadata from the donor and fold it into MODS records for UNC Charlotte’s Islandora digital repository, project workflows, donor training, and the special challenges of maintaining an amicable working relationship with a donor on a long-term, complex project. Above all, it will address the difficulties of capturing the elusive memories and knowledge of donors in metadata records that represent them faithfully and enhance user discovery.