Using metadata in catalogs
Metadata is not a new concept in cataloging; classical libraries or video stores already use metadata. For instance, a form describing a book by title, author, edition, the name of the person who last borrowed it, and a storage code is a set of metadata about the resource book. The Global Spatial Data Infrastructure group has studied the role of metadata in the distribution of geographical data . They see three levels of metadata:
• Discovery metadata should provide support for answering the following questions: “What data sets exist? Which data set contains the data I am looking for?” They include, for instance, title, spatial extent, and feature catalog of a data set.
• Exploration metadata should provide support for assessing a data set’s fitness for use. They include, for instance, quality elements.
• Exploitation metadata should provide support for the retrieval and use of a data set. They include, for instance, the projection used in the spatial reference system as well as the application schema.
Former metadata were specific to the system that owned the described resources and distributed them (e.g., a library, a grocery, or a video store). The distribution of geographical digital data is a different context — the cataloging service is not supposed to own the resources it describes. A resource may actually be cataloged by several services, and a service may catalog resources from various owners. It is up to the resource owner to provide metadata about his resource for catalog services to distribute it. This calls for the definition of metadata models so that resources of the same type can be described the same way, which eases the cataloging process. A metadata model should specify the structure of the description as well as how to document it, so that metadata stemming from various metadata producers will be homogeneous enough to be managed together in one catalog service.