Lisa Barrier, Asset Cataloger (right) and Kathryn Gronsbell, Digital Collections Manager (left) work together in the Carnegie Hall Archives. Kathryn is also an active member of the DLF Metadata Assessment Working Group, and co-leads its Website subgroup.
In celebration of Preservation Week 2018, Carnegie Hall Archives released the initial version of its Digital Collections Metadata Application Profile. The Metadata Application Profile (MAP), co-authored by Lisa Barrier (Asset Cataloger) and Kathryn Gronsbell (Digital Collections Manager), describes metadata elements for item-level asset records within the Carnegie Hall Digital Collections. Our goal for developing the initial MAP was to begin to assess our metadata maturity. We recognized the opportunity to document and share the Carnegie Hall (CH) metadata standards, cataloging procedures, and controlled vocabularies. We want to share the relatively streamlined process for generating and publishing a MAP to encourage others to consider this path for self-assessment of collections metadata.
We expected this level of metadata wrangling and organization to be a daunting task. However, we realized the benefits of utilizing the resources in the DLF AIG Metadata Application Profile Clearinghouse Project, which is part of the larger Assessment Toolkit. We opted for a simple MAP profile format of a “Quick Look” summary and a separate, detailed elements table. We created the Quick Look, a list of metadata elements and their obligations (required, mandatory if applicable, and optional). We then compiled element descriptions and pulled sample data from the CH Digital Collections to create entries in the Elements table. The samples helped us identify controlled vocabularies, free-text fields, and different sources and structures. We described input guidelines to clarify how to populate each element field. We referenced the Sample Metrics for Common Metadata Quality Criteria and followed instructions for building “Your Application Profile” in the Framework section of the Toolkit.
This metadata gathering and documentation process took approximately eight hours, over three days. Internal documents and implementations were referenced to create the elements table and Quick Look:
- staff training wiki entry for uploading and tagging material;
- cataloging requirements;
- Carnegie Hall/Empire State Digital Network (ESDN) Mapping exercise supported by the Metropolitan Library Council (METRO) to prepare for contributing to the Digital Public Library of America (DPLA);
- local taxonomy managed in the CH digital asset management system; and
- integrated performance history data, which has its own authority control (version available at data.carnegiehall.org).
We published the MAP via our institutional GitHub account (github.com/CarnegieHall), using the GitHub Pages option with a default Jekyll template. GitHub Pages is a feature that semi-automatically turns repository content into streamlined websites, useful for presentation and ongoing maintenance. We translated the elements table from a draft spreadsheet to a text document using the lightweight markup language Markdown. Markdown text displays in GitHub documents and GitHub Pages as formatted text. Because of the lengthy elements table, we chose to create a MAP overview homepage to link to the elements and to the Quick Look. We borrowed language and structure from Clearinghouse examples and other open documentation projects to draft a brief overview, feedback options, and acknowledgements. The publishing, formatting, summation, and copy-editing process took roughly four hours, over two days.
Our initial MAP release demonstrates significant progress in metadata documentation through a small investment of time. Two staff members organized, drafted, and implemented the CH Digital Collections MAP in about 12 hours, over 2 weeks. Through the process, we prioritized metadata fields for evaluation, revision, or removal. We used the MAP to update our internal Digital Collections wiki, which guides CH staff members how to add and edit system metadata.
While some elements map to Dublin Core and to DPLA properties, others are CH specific. These elements require further research for normalization and interoperability. Future additions to the MAP will include mapping to DPLA properties, Dublin Core, and other appropriate metadata schema.
We recently contributed the CH MAP to the DLF AIG Metadata Application Profile Clearinghouse Project. We hope that others can borrow formatting or content from our profile, as well as provide constructive feedback so we can continue to correct, clarify, and improve the site.