One process-level point I want to highlight is that a “consensus” metadata standard developed without some level of concurrent adoption or testing in institutional data workflows and projects is likely to need major revisions. The development of the Ecological Metadata Language (EML) in conjunction with staff at the U.S.’s Long-Term Ecological Research sites is a good example of the challenges of separating development and testing. Reflecting on the lengthy, conjoined revision-and-implementation process that followed the announcement of the EML 1.0, Millerand et al. 2007 conclude that “we slice the ontological pie the wrong way if we see software over here and organizational arrangements over there.” Changes to information standards are often inseparable from changes in organizational roles, for example as the local managers responsible for implementing new standards must become (and be recognized as) developers in their own right in order to articulate how local circumstances match onto abstract expectations. This suggests a high value on starting to test metadata standards in the core activities of key stakeholders before attempting to finalize agreement on a general consensus standard for describing collections.
Millerand, Florence, and Geoffrey C Bowker. 2009. “Metadata Standards. Trajectories and Enactment in the Life of an Ontology.” In Standards and Their Stories, edited by M Lampland and Susan Leigh Star, 149–65. Ithaca, NY: Cornell University Press.