Over deze norm
Since the release of CWA 14855, significant work has been carried out defining a DCMI Abstract Model which has been made available, at the time of endorsement of this CWA (January 2005), as a DCMI Working Draft [POWELL]. The DCMI Abstract Model provides a reference model against which guidelines for encoding DC metadata can be compared, clarifying many aspects of the nature of a Dublin Core description. Alignment with this work suggests the use of more precise terminology when describing DCAPs. CWA 14855 defines a DCAP as a declaration specifying which metadata terms an organisation, information provider, or user community uses in its metadata. A DCAP identifies the source of metadata terms used. Optionally, CWA 14855 states, a DCAP may provide additional documentation on how the terms are constrained, encoded, or interpreted for application-specific purposes. However, taking the DCMI Abstract Model into account we can be more precise in our language within the definition of a DCAP. A DCAP is primarily concerned with declaring 'property usages', and the DCAP specifies and optionally constrains or restricts the use of properties. These changes in terminology are outlined in the Definitions section below. Given the limited timescale and effort, these DCAP Guidelines are concerned with modelling DCAPs that are used as the basis for descriptions of a single resource, and do not address the more complex case relating to 'description sets' relating to multiple resources as outlined in the DCMI Abstract Model. Further work is required to consider how a DCAP might be structured to declare property usages about multiple resources. In summary, this CWA will focus on the expression of stable, well structured, non-complex DCAPs. Both human readable DCAPs as discussed in CWA 14855, and machine-processable representations of DCAPs outlined in this CWA are intended to relate to the same DCAP model. However, due to the difference in the underlying representations, there are some differences in the information provided by each. CWA 14855 permits some flexibility in the presentation of human readable DCAPs by introducing a Principle of Readability that allows for additional prose to be included in the representation of the DCAP to provide context. The Principle of Readability allows for redundant inclusion of information within a DCAP in order to enable the reader to gather comprehensive information from one single document. Following this principle, the human-readable representation of a "property usage" can include information about the property that is "used". Therefore, for a usage of dc:title that makes the Title property mandatory but keeps the DCMI definition of this property, it is helpful for a human-readable representation to include the DCMI definition of Title, rather than just a terse URI Reference [URI], whereas in a machine-processable DCAP, such information regarding the definition of dc:title would be redundant. Similarly, a human-readable DCAP might include details of the relationships between properties, whereas this information is redundant in a machine-processable DCAP. However, in both cases these relationships should also be expressed in the metadata vocabulary from which the property usages are derived. The presence of contextual information within a human readable DCAP means there can be no assumption that such DCAPs will be convertible into machine-understandable forms without the use of ad-hoc heuristics or manual intervention.
||Guidelines for machine-processable representation of Dublin Core Application Profiles