Use Case: Radio Station Archive Digitization Owners: Steven Morris David Smith William Waites Presenter: Ray Denenberg
Current Practice
Radio stations archive audio programs
Often not digitized
Radio stations archive audio programs Often not digitized Little or inconsistent metadata
Radio stations archive audio programs Often not digitized Little or inconsistent metadata Occasional efforts to digitize these programs and create metadata
Radio stations archive audio programs Often not digitized Little or inconsistent metadata Occasional efforts to digitize these programs and create metadata Metadata creation usually ad-hoc
Goal
indexed and searchable
cross references to other events particularly news broadcasts
indexed and searchable cross references to other events particularly news broadcasts enable federated searching Across programs and cross referenced events
Scenario
Scenario …… ……. at the Radio Mogadishu archives
An expert works with an archivist to create and annotate digital archives
The archivist has archives stored on old tapes, has a catalog system, but much of the information is simply in his memory.
An expert works with an archivist to create and annotate digital archives The archivist has archives stored on old tapes, has a catalog system, but much of the information is simply in his memory. As the expert creates the digital versions, he annotates them with information from the catalog system, the archivist, and a native (Somali) speaker listening to the tapes.
Application of linked data
Exploit the extensibility of RDF
New predicates can be defined as needed. Examples:
Exploit the extensibility of RDF New predicates can be defined as needed. Examples: Define a subproperty of dc:identifier for the Radio Mogadishu Archives tape labelling scheme.
Exploit the extensibility of RDF New predicates can be defined as needed. Examples: Define a subproperty of dc:identifier for the Radio Mogadishu Archives tape labelling scheme. Annotate a recording with information about the participants: is it an interview? Who is the interviewer? The interviewee? The station director at the time?
Exploit the extensibility of RDF New predicates can be defined as needed. Examples: Define a subproperty of dc:identifier for the Radio Mogadishu Archives tape labelling scheme. Annotate a recording with information about the participants: is it an interview? Who is the interviewer? The interviewee? The station director at the time? Create URIs for the people involved.
Problems and Limitations
Little guidance for creating metadata about audio recordings
Where to annotate who did the digitization? Who transcribed the metadata?
Little guidance for creating metadata about audio recordings Where to annotate who did the digitization? Who transcribed the metadata? When to create Works, when Manifestations.
Little guidance for creating metadata about audio recordings Where to annotate who did the digitization? Who transcribed the metadata? When to create Works, when Manifestations. No vocabulary to describe the state of source material, e.g. "readable", "partially- readable", "unrecoverable"
Problems and Limitations (Continued) Confidence/Uncertainly How to preserve information about the transcriber's uncertainty. “Looks like an X but might be Y”. "Is that an 8 or a 3?"