Slides available for ALA NRMIG Presentations in Midwinter 2008

Maureen P. Walsh, Metadata Librarian, Assistant Professor, The Ohio State University Libraries
Topic: Institutional Repository Metadata

(Presentation slides and handouts are available at: )

The Ohio State University’s Institutional Repository is also called “the Knowledge Bank” ( As a Metadata Librarian at OSU, Walsh addressed some of the important issues on institutional repository metadata, such as metadata schemes, crosswalking, data normalization, harvesting from the viewpoint of shareable metadata and customization of metadata display and user interfaces.

Currently, the Knowledge Bank has 40 communities and about 29, 573 records. The types of materials they deposit include journals, monographs, undergraduate thesis, conference materials, technical reports, images and oral histories. Based on the statistics from ROAR (Registry of Open Access Repositories), deposits in Knowledge Bank increased steadily from 2004 to 2007. In 2007, 5 days of the daily deposits were over 100 items, and 47 days of the daily deposits were between 10 and 99. The Knowledge Bank has been harvested by Google, Google Scholar, OpenDOAR, CIC Metadata Portal and OAIster. Walsh discussed the KB metadata application profile, Metadata Registry and preservation metadata. She explained the effort they took to make better metadata and better display interfaces, including to make metadata more shareable, perform authority control of author names and subjects, create community metadata application profile, customize input forms to allow users to choose from predefined field value (such as department, award statement, interviewee, interviewer and subject), to customize item metadata display, and to repurpose MARC records. As Walsh explains, it is to “add a measure of data control in [their] institutional repository in the interests of both quality metadata and shareable metadata.” Walsh’s valuable and comprehensive experience as well as presentation will be a good reference for other librarians who also work with institutional repositories.

Amy Jackson & Myung-Ja Han, Project Coordinator, IMLS Digital Collections and Content, Metadata Librarian, University of Illinois Library at Urbana-Champaign
Topic: Changes in Interoperability of Dublin Core Metadata Records Over Time

(Presentation slides available at: )

The IMLS Digital Collections and Content project (, a collection registry and item-level metadata repository, is an IMLS (the Institute of Museum and Library Services) National Leadership Grant Program (NLG). It began in December 2002, and recently extended to 2010. Currently the collection registry has 180 NLG projects and 15 LSTA (the Library Services and Technology Act) projects, and overall harvested about 310,448 records.

To study sharable metadata quality change over time, the researchers at UIUC did qualitative and quantitative analysis of records harvested from Jan. 1, 2001 to Dec. 31, 2006. Quantitative analysis looked at the use of core fields and the length and repetition of the fields. Qualitative analysis examined the misuse of Dublin Core (DC) fields and the mapping errors. The results of the former analysis shows a decline in the use of all eight core DC fields, in more detail, the most often missing elements are creator and rights, and format and description fields have shown the most significant decline in use since 2003… It is not surprising to find out that “users can only search across all records by searching on the title field”, and in fact, “metadata creators are becoming more discriminating in their use of DC fields in the local context”. The results of the latter analysis displays that the misuse of DC elements is not uncommon (such as the misuse of Date and Coverage fields, Source and Relation fields, Format and Description fields, and Type and Format fields), confusion in descriptive metadata and administrative metadata, and information lost in mapping from local scheme such as MARC to DC.

Based on the analysis, it is concluded that positive changes in metadata practices have not been observed. The researchers recommend that the data providers publicly document crosswalking practices and communities publish local metadata practices. It is also suggested to expose native metadata in addition to DC for the mapping process and to ensure that creators receive appropriate training in creating sharable quality metadata.

Kristin Martin, Electronic Resources Cataloger, Catalog Department, UNC Chapel Hill
Topic: Building a Collection of Electronic Theses and Dissertations: Metadata Issues and Lessons Learned

(Presentation slides are available at: )

The UNC (University of North Carolina at Chapel Hill) Libraries provide access to the electronic theses and dissertations (ETDs) through both the traditional catalog and CONTENTdm. Starting from May 2008, electronic submission becomes mandatory for the UNC graduate students. The ETD website is at

Martin gave us an overview of the ETD (Electronic Theses and Dissertations) project at UNC. She discussed metadata used for ETDs, which tried to follow the standards of NDLTD (Networkded Digital Library of Theses and Dissertations), UNC’s local ContentDM data dictionarly and AACR2 and LCRI. As these standards will not always agree, their two major cataloging agencies also showed inconsistency in terms of metadata creation practices. Despite these reality dilemmas, additional challenges exist in the ContentDM software, the workflow, PDF format and the DC-MARC crosswalk implementation. Martin described the transformation process of their expanded version of Dublin Core (in ContentDM) to MARC (for OCLC and local catalog). The main reason they start from ContentDM is that “selected fields can be prepopulated from information provided by the Graduate School,” and “difficulty maintaining consistent URL without beginning in ContentDM”. It is a very sincere sharing of a reality show.

This entry was posted in ALA Midwinter 2008. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *