Purpose-Library-world "languages of description" are increasingly being expressed using the resource description framework (RDF) for compatibility with linked data approaches. This article aims to look at how issues around the Dublin Core, a small "metadata element set," exemplify issues that must be resolved in order to ensure that library data meet traditional standards for quality and consistency while remaining broadly interoperable with other data sources in the linked data environment. Design/methodology/approach-The article focuses on how the Dublin Core-originally seen, in traditional terms, as a simple record format-came increasingly to be seen as an RDF vocabulary for use in metadata based on a "statement" model, and how new approaches to metadata evolved to bridge the gap between these models. Findings-The translation of library standards into RDF involves the separation of languages of description, per se, from the specific data formats into which they have for so long been embedded. When defined with "minimal ontological commitment," languages of description lend themselves to the sort of adaptation that is inevitably a part of any human linguistic activity. With description set profiles, the quality and consistency of data traditionally required for sharing records among libraries can be ensured by placing precise constraints on the content of data records-without compromising the interoperability of the underlying vocabularies in the wider linked data context. Practical implications-In today's environment, library data must continue to meet high standards of consistency and quality, yet it must be possible to link or merge the data with sources that follow other standards. Placing constraints on the data created, more than on the underlying vocabularies, allows both requirements to be met. Originality/value-This paper examines how issues around the Dublin Core exemplify issues that must be resolved to ensure library data meet quality and consistency standards while remaining interoperable with other data sources.