From Dempsey’s Blog Post:
Libraries are very used to managing metadata for information resources – for books, images, journal articles and other resources. Metadata practice is rich and varied. We also work with geospatial data, archives, images, and many other specialist resources. Authority work has focused on people, places and things (subjects). Archivists are concerned about evidential integrity, context and provenance. And so on.
I want to talk about two important directions here.
Entification: strings and things
Google popularized the notion of moving from ‘strings’ to ‘things’ when it introduced the Google knowledge graph. By this we mean that it is difficult to rely on string matching for effective search, management or measurement of resources. Strings are ambiguous. What we are actually interested in are the ‘things’ themselves, actual entities which may be referred to in different ways.
Entification involves establishing a singular identity for ‘things’ so that they can be operationalized in applications, gathering information about those ‘things,’ and relating those ‘things’ to other entities of interest.
Systems of description are inevitably both explicitly and implicitly constructed within particular perspectives. Metadata and subject description have long been criticized for embodying dominant perspectives, and for actively shunning or overlooking the experiences, memories or expectations of parts of the communities they serve. They may also contain superseded, obsolescent or harmful descriptions.
Libraries have spoken about “knowledge organization” but such a phrase has to reckon with two challenges. First, it is acknowledged that there are different knowledges.
Second, knowledge may be contested, where it has been constructed within particular power relations and dominant perspectives.
See Also: Lorcan Dempsey’s Blog
See Also: Lorcan Dempsey’s Bio (via OCLC)