My article, The Use of Non-MARC Metadata in AALL Libraries: A Baseline Study, has been published in Law Library Journal, 103: 631-658 (2011). Here is the abstract:
This article reports results of a 2009 survey of AALL libraries respecting non-MARC metadata practices, with a focus on interoperability. Results cover types of collections described with non-MARC metadata, as well as metadata standards, platforms, and tools. Results suggest substantial, though incomplete, awareness among respondents of metadata interoperability and the factors that enable it. This study is intended as a preliminary inquiry that affords some sense of law libraries’ recent non-MARC metadata activity and offers findings that may furnish a baseline for future studies of such activity.
Here are the major findings:
Non-MARC metadata are being used most often to describe manuscript or archival collections, resources in portals or subject-specific collections, and collections of primary legal resources, and are being used to describe unpublished and published materials in approximately equal proportions. […]
The platforms, systems, or components thereof utilized most commonly in connection with non-MARC metadata activity are generic web servers, followed by PHP, Microsoft Access, Linux plus Apache, MySQL, and locally created systems. […]
Use of open, standard descriptive metadata schemas, or of systems that can output metadata in such schemas, appears to be widespread. However, nearly thirty-five percent of respondents reported using no or minimal non-MARC descriptive metadata for at least one digital collection or project, and almost one-fifth of respondents reported utilizing locally created schemas in systems that could not output descriptive metadata in open, standard schemas. […]
Dublin Core is the most frequently employed descriptive metadata schema, followed by locally created schemas, HTML or XHTML “Title” or “Meta” tags, and EAD. […]
More than half of respondents reported utilizing mapping between descriptive metadata schemas, most commonly from qualified to unqualified Dublin Core, a mapping that is consistent with the OAI-PMH specification. […]
The most frequently identified structural metadata format was PDF, followed by HTML/XHTML and EAD. [….]
Nearly two-thirds of respondents report using systems that do not comply with OAI-PMH, in connection with at least one current collection or project. […]
Taken together, these findings seem to reflect substantial, though incomplete, awareness among respondents of metadata interoperability and the factors that enable it. In addition, these findings point to two factors as possible obstacles to wider and swifter adoption of practices that foster metadata interoperability: the legacy of early digitization activities—which involved the creation of large collections of “flat” HTML or XHTML documents containing little or no descriptive metadata—and the maintenance of databases of current primary legal resources, which require fast and inexpensive metadata processing techniques.
I’m most grateful to the survey respondents — who requested confidentiality — to Kathleen Cowan of bepress for furnishing information about the Legal Repository and DigitalCommons platforms; to Leslie Carr of eprints for answering questions about that platform; and to Janet Sinder of Law Library Journal for comments on earlier drafts of the article.