George Oates has a great post on the Open Library Blog which looks at library data from the perspective of mass processing and variety of data sources and argues for a “minimum viable record”. To a certain extent this ties in with what I wrote about last week. How can we simplify metadata and still meet the goals of how a library catalog serves the user? What are the key elements to describe a resource in an accurate, easily usable and processable manner? To me, the concepts of “good enough” and “minimum viable record” seem to be related in that they try to distill the essence out of descriptive library metadata.
Bibliographic data is used by patrons, librarians and machines – for display, discovery, disambiguation, matching, faceting, exchange, to name but a few of the current use cases. For each of these user groups and tasks different metadata elements are more or less important. “Good enough” and “minimum viability” have to be judged according to these aspects; it’s hard to come up with an across-the-board definition, although there are certainly areas where human and machine requirements overlap (what could be called the “essence”). Additionally, we always need to bear in mind that we don’t yet know every detail of what future systems will be able to do with our data – so how much simplification is justifiable? I imagine that for FRBR structures you would again need more details in order to distinguish different expressions or manifestations (whose place of publication may differ, for example).
Sticking to the metadata essence could make matters more efficient for both humans and machines, and we might even achieve greater consistency and interoperability. Enjoying more freedom than libraries, the Open Library could be a testbed for the practicality of “minimum viable records”.