I am working on an RDF model for my cataloging rules. So far I have created the following documents:

  1. a list of tables of classes and properties
  2. a draft of the tables in RDF Word icon
  3. a draft of the tables in RDF as a .txt file Word icon
  4. a draft of the examples in RDF Word icon

This RDF model is very much a work in progress. In attempting to build it, I have stumbled over a number of unresolved issues, which I list in:


I would be very grateful to hear from people who are more expert data modellers than I how they would solve these problems.

Each question is listed separately at my blog.

One complaint that has been made about the examples is that entities are still identified with language string identifiers in combination with machine-actionable identifiers; the efficient approach eventually will be to use machine-actionable identifiers everywhere except in the record for the entity itself, and retrieve the language string identifier from that record whenever the entity needs to display. This would be equivalent in our current environment to putting the heading for a work, for example, only in an authority record for the work, and linking all work-related bibliographic records to that authority record with a machine-actionable identifier such that whenever a bibliographic record displays, the software retrieves the human-readable work heading from the authority record. I created one alternative example for the Tess of the storm country film to demonstrate this. However, I have not taken the time to create rdf examples for every heading used in these examples, and the examples would be very hard to read without the human-readable headings, so I hope everyone can just try to imagine that they might not be there ultimately.

A bibliography of the works I consulted before building this model.

Last modified: February 18, 2008, my