DC Registry WG meeting report
DC2004, Shanghai, 2004-10-14

The meeting was attended by 31 people.

Harry Wagner presented some slides to help guide the discussion over a
number of topics (the slides are available
here: http://dublincore.org/groups/registry/dc2004_registry.pdf ).

We started with a brief overview of the registry, including:

- an overview of the project's mission statement and goals
- a recap of our current status
- an overview of the application architecture and the technology it is built
upon

The remainder of the meeting was spent on an open discussion of our
prior-year accomplishments, the activity
at the distributed sites and a review of the survey results.

Prior-year accomplishments:

- published a new working group charter.  That charter is available here:
http://dublincore.org/groups/registry/

- added support for a REST-style application interface.  REST is a style of
Web services based on HTTP.  It is
popular due to it's lightweight implementation and ease-of-use.  The
registry application interface is a collection of
Web services.  This includes both REST-style and SOAP-style services.

- installed a distributed registry at the Library of the Chinese Academy of
Science.  This registry will play a small
role in a much larger project to link all of the Chinese digital libraries.

- published an installation guide and an administration guide.  The
installation guide is available here:
http://dublincore.org/dcregistry/pageDisplayServlet?page=install.xsl and the
administration guide (still in draft)
is available from the link provided in the administration section.  There
were a lot of comments regarding the
installation guide and installation in general.  A suggestion was made to
add 2 new sections to this document; an efforts
guideline that provides detail regarding how much effort and expertise is
required for various types of installation
(i.e., an out-of-the-box install vs. an installation with various local
extensions).  It was also suggested that a
section be added to address security.  There are no security issues with a
registry install, but it was felt the document
should state this.  An additional request was made to review the use of
packaging to further simplify installation.

- metadata and user-interface translations for Czech, Ukrainian and Welsh
were added.

- the user interface was completely re-designed to simply use.
Functionality that enables registries to add usage
examples for terms was added.  Additionally, quick-links to different
encodings (including RDF and N3) was added.

- there were numerous enhancements to the administrative component to
simplify installation and support.

- the registry was migrated from Jena 1.6 to Jena 2.0, providing additional
support for the ontology layer.

There was some discussion about the current activity at the distributed
sites.  Registries are currently deployed at
OCLC (Dublin Ohio), The University of Tsukuba (Tsukuba Japan), The
University of Goettingen (Goettingen Germany) and the
Library of the Chinese Academy of Sciences (Beijing China).  Representatives
from each of the distributed sites, with the
exception of Goettingen, were present and were able to comment on their
current activity and future plans. One item worth
noting is that each of the distributed sites are using their registry
implementations for research, with the exception of
registry in Beijing.

The remainder of the meeting was spent reviewing the results of the recent
survey
(http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0409&L=dc-registry&T=0&F=&
S=&P=313).  While the number of responses
generated by the survey was smaller than we had hoped, the information
provided proved quite valuable.  Survey results
indated that the most important feature of a metadata registry was to
provide an authoritative and trusted source of
information, followed closely by an application interface and inter-registry
cooperation.

The greatest barrier to installation was determined to be cost.  This came
as a surprise since the registry is open source.
Some follow-up will be needed to determine exactly where cost becomes a
factor, or if the question was simply poorly-worded.

Application profile generation and metadata crosswalks were rated the most
inidspensible new features for a metadata
registry and is something that will be considered in the coming year.  Where
this functionality fits within the
primary role for the registry (essentially a resolution service) is not
clear.  One other item important to note was the
number of write-in features for functionality already supported by the
registry.  This indicates that there is
still work to be done in promoting the registry and raising awareness of the
benefits provided.

Best Regards,
Harry Wagner
http://oclc.org/research/staff/wagner.htm