DPIP Priorities: Needs, Challenges, Issues

Metadata production and standards

Matthew Beacom, Rebecca Irwin (site visit text by Katie Bauer)

December 19, 2005

 

Issues

 

Serving the metadata production, development and re-use needs at Yale University Library requires an overarching strategy to define the mission for metadata production and development services, structure the mix of activities that make up metadata production and development services, and relate them to other aspects of the library’s larger integrated access mission.

 

Metadata is a critical part of the infrastructure necessary to support teaching, research, and publishing at Yale. Yale University Library needs to create and sustain a coherent, agile, and scalable metadata production and development program for digital resources, collections and services available to the Yale Community.  The mission of such a program would be to provide administrative, descriptive and structural metadata that allows members of the Yale Community to manage, discover, and use the digital resources, collections, and services available at Yale.   The metadata produced by such a program would support the library’s larger integrated access mission by creating conditions that help to optimize the re-use potential of digital materials at Yale.

 

The components of such a metadata production and development program are:

           

Metadata Integration for Access

YUL needs to produce metadata for a variety of uses, services, tools, materials, and collections and do this in ways that support re-purposing the metadata and associated digital resources in an integrated environment. The current status at Yale is a mix of digital service silos (the Beinecke DL, MADID , Insight for visual resources, the Medical Library’s digital resources, electronic materials in Orbis, etc.).  These and other silos reflect both the idiosyncrasies of YUL organization and distinctive patterns of use that are grounded both in distinctive types of content—image, language, data—and distinctive uses—classroom use, publication,  and research. From the point-of-view of integrated access, work is needed on tools for federated searching to combine like databases and metadata harvesting to create consolidations of metadata for digital resources that can support new services.

 

At Cornell the assignment of metadata is part of the workflow of every DCAPS project.  Metadata is both structural and descriptive.  The two types of metadata happen in separate parts of the workflow.  Structural metadata is assigned in the digitization process, and descriptive metadata is applied separately.  Some projects come to them with legacy metadata which is then mapped to their standards.  Key to their metadata workflow is the presence of a full time programmer in the Metadata Group.  This person functions to map metadata and to load to the appropriate system. 

 

 

Virtual Metadata Team

YUL needs to share management of metadata production and development across units and administrative structures by developing a virtual team with responsibility for metadata production services and metadata development activities. Coordination of metadata production, development, and re-use at YUL is essential to use human resources effectively and create coherent, broad services for users.

 

At Cornell, the provision of metadata is a major component of all digitization projects done through DCAPS.  Structurally and bureaucratically the 8 person metadata group at Cornell is part of the Cataloging Department.  Functionally, the metadata group is part of the DCAPS team, and the head of the metadata group serves on the DCAPS management team.  This approach grounds the metadata group in a traditional library unit, while providing support for DCAPS activities.

 

Metadata Staffing

YUL needs to assign staff—C&Ts, M&Ps, catalogers, systems, public services, etc.—to metadata production services and metadata development activities. Designated staff are necessary to understand our capacity and costs in ways that allow us to take advantage of opportunities, plan for production needs, and develop needed skill sets.

 

At Cornell the Metadata Group consists of catalogers, support staff, and one programmer.  The programmer is key to the work of the group.  The creation of this group with a steady level of digitization work has meant that they have had both the need and the ability to develop expertise in a relatively short period of time.  Their work is not so fragmented that they can't concentrate on digital projects and metadata for digital projects.  They have a core group for sharing ideas, and their work receives the necessary programming support. 

 

To begin a DPIP program at Yale, the two librarians who currently support metadata might be sufficient for metadata staff, but they would also need to have access to a full time programmer and at least one support staff person.

 

Metadata Research and Development

YUL needs to use the metadata production and development services virtual team for research and development focused on the role of metadata in user-oriented services.  Metadata production at YUL must have a research and development aspect to adopt new technologies and meet user expectations for metadata-based services.

 

The most important part of metadata activity at Cornell is a practical application of emerging standards and theory to actual practice.  This is important in a world where standards are not yet set.  It did not seem that a lot of time is spent there in discussing standards.  Rather the metadata group makes policy decisions and is confident in their ability to migrate their data to any new standard which may emerge.

 

Tools

YUL needs to use a shared suite of supported metadata production tools to create, edit, exchange, re-format, and re-purpose metadata for resources in Yale University Library (both batch and individual processes). 

 

The DCAPS Metadata Group at Cornell has adapted several tools for filling this need.  They have not found the perfect solution.  They use XML editors, and unfortunately the need to understand XML adds a layer of complexity for staff who only need to enter data.  Ideally they wish to invest in tools that will hide some complexity, but they have made no decision yet. 

 

Element Sets

YUL needs to use shared metadata element sets like the Yale Element Set to support discovery, management, and use of resources across collections (through cross collection search tools or harvested metadata). Extensions to the Yale Element Set for various types of formats such as audio and moving image materials and preservation activities are needed.

           

Content standards

YUL needs to use shared metadata content standards to provide consistent semantics in metadata across collections. YUL will need to use a variety of existing, emerging, and novel content standards—AACR, DACS, CCO, etc.—to serve its diverse collections and uses.

 

Cornell has a decentralized model as Yale does.  DCAPS is not decentralized but it is serving the needs of a diverse set of projects from many different collections.  In such a decentralized model it is clear that the adoption of standards will be key so that the various collections will all be consistent in their use of metadata, so that collections can be searched together even if they "live" on different platforms. 

 

Best Practices

YUL needs to use shared best practices across all metadata production units to maximize quantity, quality, and consistency. Best practices include use of element sets, content standards, work flow designs, and benchmarking production costs and productivity.

 

Documentation

YUL needs to create or adapt and use shared documentation to support best practices—policies, procedures, tool use—and support training staff who are engaged with metadata production and development.

 

Training

YUL needs to develop and implement cross unit training programs for staff engaged with metadata production and development.

 

The metadata group at Cornell plays a key role in educating and supporting metadata activities in other library units (Cornell has a very decentralized structure where cataloging may take place at all of the 20 unit libraries.)  The virtual nature of the Metadata Group within DCAPS helps in this activity.  This gives Cornell a core group of metadata staff who are committed to the DCAPS agenda, but at the same time are spreading metadata practices out to an ever-widening circle of other staff.  Eventually, this will mean that metadata practices are embedded in all cataloging activity, and the metadata group may cease to exist.  Currently, it is clear that a separate well-staffed group has a key role to play.