Directory Data Model

This specification describes the data model used by the Talis Directory defined in terms of the types of entities being modelled, their relationships to one another and the properties they may have. The directory uses RDF as a highly flexible data format capable of being rapidly evolved to encompass new entities and relationships. Each entity in the directory is represented by an RDF class and may have specialised versions for particular data types (for example: Library is a specialisation of Location). In common with all RDF-based data models the properties described are those typically expected to be present but there may be additional ones not documented here. The presence of these additional properties will not affect the behaviour of the system. The RDF model encourages reuse of existing work to promote interoperability between disparate systems. Therefore the data model uses several existing schemas such as those devised by Dublin Core, W3C and the FOAF project. Data exposed in these formats can be expected to have a significant number of tools and applications available that will already be able to consume it.

Documentation Conventions

The following sections define the classes expected to be used in the Directory. A short definition of each is supplied followed by a listing of the expected properties. The numbers in brackets after the property name indicates the number of occurrences of that property expected in the model. We expect the wiki-like nature of the directory to result in many partial descriptions of entities and so there are very few mandatory properties. Properties are either references to other entities identified by URIs (including email addresses and web pages) or they are strings. This should be clearly indicated in the descriptive text for each property. An example of expected content is sometimes provided. For reference types the example will be enclosed in angle brackets ( < and > ), whereas string types will be enclosed in double quotes ( " ). All free text values can be annotated with the language of the text.

RDF uses URIs to identify classes of things and the properties those things posess. Rather than write the full URI for every term discussed in this document a common abbreviation method is used. Each term is written in the form "prefix:name" where prefix is a short mnemonic that stands in for a part of a URI. The URI of the term can be found by concatenating the URI represented by the prefix with the name part. The colon is merely used to separate the prefix from the name and does not form part of the final term URI.