As part of the process of developing consensus on means of describing learning resources, UKOLN has created entries for some metadata schemas for educational materials in a metadata registry.

The DESIRE Registry was created as part of a collaborative EU-funded project to enhance resource discovery for researchers. The DESIRE registry has been implemented using a relational database (mySQL) and utilises a data model which supports the registration of data elements from multiple vocabularies or namespaces. Information about valid values for data elements may be recorded by associating elements with schemes.

The model incorporates the concept of application profiles to describe the (re-)use of data elements for a particular purpose or application. An application profile does not define new data elements; it draws together previously defined elements from (specific versions of) multiple vocabularies/namespaces. Further, it may

• introduce constraints on the value of an element by associating the element with schemes specific to the application,
• introduce constraints on the obligation to use an element (e.g. make the use of an element mandatory) or the occurrence of an element (e.g. whether an element is repeatable),
• refine the semantic definition of an element to make it narrower or more specific to the application domain.

The concept of the application profile has proved particularly useful in the context of schemas for educational resources. Many of the schemas use subsets of elements drawn from other metadata element sets (typically IMS, IEEE-LOM, Dublin Core): in terms of the DESIRE model, they are application profiles.

The registry can record correspondences between the elements of these namespace vocabularies and the units of an underlying "semantic layer", which permits the generation of "cross-walks" between namespaces. In practice, this has proved to be slightly problematic. It requires careful interpretation of the semantics of individual elements. Also it requires that either the semantic layer includes units of similar specificity to the namespace elements (the shortcomings of the use of a general-purpose semantic unit set are highlighted when dealing with domain-specific schemas, as in the case of MEG) or that the registry incorporates a mechanism to describe partial or "fuzzy" semantic correspondences.

The purpose of creating entries in a metadata registry is to make information about the structure and semantics of metadata element sets available in both human-readable and machine-readable form, so that they are available both to

• human readers who wish to discover what schemas and profiles are already available, to explore and compare the structure and semantics of those already registered, and perhaps to bring their own schemas and profiles into alignment with them;
• software agents that wish to process metadata instances and need to reference descriptions of the structure and semantics of the schemas which those instances use.

In its present form the DESIRE registry fulfils only the first of these objectives. Another registry project of which UKOLN is a partner, SCHEMAS, is working on the construction of a registry which stores the content in RDF/XML-based forms. The goal of the SCHEMAS registry is to make its content available in forms which are useful both to the human browser and to a software agent.

The DESIRE registry contains several MEG-related entries.