That was the topic discussed recently by OCLC Research Library Partners metadata managers, initiated by Jennifer Baxmeyer of Princeton, MJ Han of University of Illinois at Urbana-Champaign, and Stephen Hearn of the University of Minnesota. With the increasing availability of online metadata, we are seeing metadata added to discovery environments representing objects of widely varying granularity. For example, an article in Physical Review Letters—Precision Measurement of the Top Quark Mass in Lepton + Jets Final State—has approximately 300 author names for a five page article (some pictured here).
This seems disproportionate, especially when other objects with many contributors such as feature films and orchestral recordings are represented by only a relative handful of the associated names. If all the names associated with a film or a symphony recording were readily available as metadata, would it be appropriate to include them in library cataloging? Ensuring optimal search results in an environment in which metadata from varying sources with differing models of granularity and extensiveness poses challenges for catalogers and metadata managers.
Abbreviated forms of author names on journal articles make it difficult, and often impossible, to match them to the correct authority form, if it exists. Some discovery systems show only the first two or three lines of author names. Research Information Management systems make it easier to apply some identity management for local researchers so that they are correctly associated with the articles they have written, which are displayed as part of their profiles. (See for example, Scholars@Duke, Experts@Minnesota or University of Illinois at Urbana-Champaign’s Experts research profiles.) A number noted that attempts to encourage their researchers to include an ORCID (Open Researcher and Contributor Identifier) in their profiles have met with limited success. Linked data was posited as a potential way to access information across many environments, including those in Research Information Systems, but would require wider use of identifiers.
Much of the discussion was about receiving not enough good metadata from vendors rather than too much. A number of the metadata managers viewed quality at least as important as granularity. Some libraries reject records that are do not meet their minimum standards, while others apply batch updates before loading the records. One criteria for “good enough” metadata is whether it is sufficient to generate an accurate citation. Metadata quality has become a key concern, as evidenced by the Digital Library Federation’s Metadata Assessment Working Group, formed to “measure, evaluate and assess the metadata” in a variety of digital library systems. Record-by-record enhancement was widely considered impractical.
Information surplus will only increase, with accompanying varying levels of metadata granularity. It remains to be seen how the community can bridge if not integrate the various silos of information.
Karen Smith-Yoshimura, senior program officer, works on topics related to creating and managing metadata with a focus on large research libraries and multilingual requirements.