Gordon Dunsire, Corey Harper, Diane Hillmann, and Jon Phipps have published "Linked Data Vocabulary Management: Infrastructure Support, Data Integration, and Interoperability" in a special issue of Information Standards Quarterly devoted to linked data issues.
Here's an excerpt:
Recently there has been a shift in popular approaches to large-scale metadata management and interoperability. Approaches rooted in semantic Web technologies, particularly in the resource description Framework (rdF) and related data modeling efforts, are gaining favor and popularity. In the library community, this trend has accelerated since the World Wide Web Consortium (W3C) re-framed many of the Semantic Web's enabling technologies in terms of Linked Open Data (LOD)—a lightweight practice of using web-friendly identifiers, explicit domain models, and related ontologies to design graph-based metadata. As more and more RDF-based metadata become available, a lack of established best practices for vocabulary development and management in a Semantic Web world is leading to a certain level of vocabulary chaos. Strategies for vocabulary publishing, discovery, evaluation, and mapping have the potential to change the conversation significantly.