Towards a cooperative ecosystem of data vocabularies

Until the mid-1990s, when the Web arrived, metadata vocabularies and concept schemes were controlled sets of words printed in big books or hard-wired into local applications. The global infrastructure of the Web provided a managed space of globally unique identifiers for the concepts underlying those words. A Web-based language for data arose, with RDF as its grammar, URIs as its words, and Web browsers as its dictionary. After the publication of SKOS in 2009, legacy thesauri and vocabularies were increasingly translated into this data language and published on the Web with URIs. There may now be twenty different URIs for a frequently used concept like "Rice", but the shared data language at least makes it easier to express globally usable mappings among them. At a time of stagnant budgets, information overload, and rising requirements for interoperability, this capability suggests new ways for institutions to cooperate and divide maintenance responsibility for data and metadata vocabularies. Complex challenges such as increasing crop yields under conditions of global warming must be addressed by mashing up data from a diversity of spreadsheets, databases, and instrument readings. Cooperatively maintained concept URIs can provide reference targets to which their data elements can be mapped.

 Invited Speaker

Dr. Thomas Baker

Chief Information Officer, Dublin Core Metadata Initiative