The DCMI Metadata Registry Working Group Meeting
DC-2003 Seattle, Washington
30 September 2003

The working group meeting was well attended, and covered the following areas:

1. A working group overview presented by Harry gave a brief review of the history and background to the WG. The current charter is 
now fulfilled as regards developing the DCMI Registry, and the chairs will be recommending an updated charter.

2. Report on activities over the last year
Harry summarized work carried out over the last year on the DCMI Registry. The application interface has been enhanced, an application
interface has been developed, and  an administrative module implemented.  A number of new translations were added to the Registry over 
the year and the Registry now supports translations in 24 languages. 

The Registry architecture is evolving from a centralized to a distributed architecture, with installations of the Registry at 
four locations (Germany: University of Goettingen; Japan: ULIS, Tsukuba; UK: UKOLN, University of Bath) as well as at OCLC. Installation 
of the Registry in different regions or locales enables support for specific communities of practice. For example, locale or domain 
specific extensions might be included in local registries. Distributed registries will be administered by the communities they serve, 
supporting translations appropriate to that community, and including metadata important to that community.

The expected benefits of deploying a local registry were summarized as 
- Provides enhanced and authoritative access to information that is important to diverse communities of practice
- Enables access to locale and domain-specific terms, extensions, application profiles, documents, classification schemes, etc.
- Cooperation with broader DCMI community is provided with application API that supports discovery and exchange of metadata

3. The Registry was demonstrated, although the network connection was not of sufficient quality to give a detailed demonstration.

4. Open issues
The following open issues were identified:
- How to most effectively provide access to information not readily available in RDF (e.g. application profiles)
- If and how to manage access to prior (historical) versions of terms,
- How to manage different versions of translation
- How distributed registries would collaborate, and relation to central registry
- How to manage 'canonical' version of a term
- How to manage provenance information in a distributed environment
- Enhancements needed to make application easier to install and extend

4. A proposed charter was outlined, this has been presented to the Advisory Board and is currently being reviewed.