Difference between revisions of "Registry Meeting Notes, DC-2010"

Latest revision as of 08:17, 20 November 2010

Registry Community Meeting at DC-2010

Wednesday, Oct. 20, 2010, Pittsburgh, PA

[link?] As part of those results, Corey described some institutions that were running metadata registries, including Oxford University (vocabs and taxonomies), Space Telescope and Science registry, OCLC, DCMI, etc.

The survey also asked what was missing from many of the available options, with answers including:

• Export and options for viewing data
• Ability to harvest content from elsewhere vs. manual input
• Content negotiation (human vs. machine-readable content)
• A variety of encodings, such as JSON and RDF
• Version control
• Publishing of change sets as feeds

Registries?.

Registry providers in the audience discussed their work and took questions from the audience.

Diane Hillmann provided information on the Open Metadata Registry, including the change in name from the NSDL Registry and anticipated improvements in development currently. The OMR sustainability plan includes working with JES and Co. (an educational non-profit and co-sponsor of DC-2010) with a focus on integrating their current projects with the OMR. The Registry Consortium is still in the plan, though the start timing has yet to be determined, and will provide a broader base for decision making and help with continuing funding.

The strengths of the OMR include: history and versioning functionality (part of the crucial vocabulary and management functions), open operation and open source software, activity notification available via RSS feeds. The emphasis on URI management, using any domain, has meant that supporting the RDA and IFLA vocabularies (FRBR, FRAD, FRSAR, ISBD, etc.) has been possible.

The Registry remodel is already in progress, focused on the challenge of Application Profiles (and Description Set Profiles) as well as cross-vocabulary mapping. Planning includes an RDF instance editor driven by DSP/OWL to eventually replace the current user interface. Corey Harper suggested that he would consider proposing a pre-conference for Code4Lib to install and work with the OMR code and will work with Jon Phipps on the proposal.

Mitsuharu Nagamori discussed the DC Registry and the use being made of its software to support the Japanese Diet Library (see his paper from DC-2009 on the project). The DC Registry is searchable, includes DC terms in 25 languages, REST, SOAP, and SPARQL interface. All metadata terms are written in RDF and maintained by Tom Baker andXu Bo. The DC Registry is open source and has been in place since 1998, now being maintained by the University of Tsukuba. The registry running this software at the Univ. of Tsukuba contains over 75 vocabularies beyond the DC Terms and their translations.

From these presentations, the group progressed to discussion lead by Emma Tonkin of design patterns, examples, and requirements that the group should include in its work plan for the coming year. Part of the design pattern conversation discussed the range of implementation options, from very simple and lightweight solutions such as web pages or spreadsheets documenting metadata vocabularies, to more robust but still light-weight registries such as the software described by Mitsuharu, to the much more robust services offered by the OMR. Depending on hosting models and institutional needs, these represent a wide spectrum of implementation options.

The To Do list:

• Business cases
  • good arguments and examples
  • Diane's text in google docs (see above)
• Functional requirements for management tools and services
• Comparison/inventory of tools (in cooperation with DC tools)
• DC provenance - present use case for registry provenance
• Registry Standards: interoperability
• Sharing and storing vocabularies and ontologies - requirements
• Webinars, activities, regional meetings
• Registry: do we need a new name for what this is?
• Preservation and curation - Work with DC Preservation Community?

Emma suggested that the group take a look at the wiki containing the earlier work on distributed registries and consider how we can build on what is there (we need a link from the Registry Community site on http://dublincore.org).

• Terminology, registries and services?
• Terminologies
• Registries
• Schemas
• Structures
• Metadata services
• Ontologies and semantic web
• Encoding schemes
  • classes
  • properties
• Value Domains

It was finally determined to leave this discussion alone for now and focus on a scope description (a couple of paragraphs at most) about what this community is and is not.

Brainstorming continued on a list of functional requirements, including:

• Version control
• Import/export
• User management (owners and end users, including trust issues)
• Provenance metadata
• Notification (end users and owners)
• Vocabulary mapping (identifying external events such as discovery, including notification of mapping assertions and bilateral endorsements)
• Authentication and authority (to allow mapping external to the vocabulary)
• Tracking and exposing usage data
• Language support
• Social networking supporting and vocabulary development
• Storage and display
• Discovery across registries
• Web services and /APIs

[Thanks to Leigh Bain for her excellent notes!]

Notes edited by Diane Hillmann and Corey Harper