May 17, 2001
Amsterdam, The Netherlands
Present: Micah Altman, Atle Alvheim, Karsten Boye Rasmussen, Ernie Boyko, Bill Bradley, Cavan Capps, Pat Doyle, Ann Green, Reto Hadorn, Peter Joftis, Emiel Kaper, Jean-Pierre Kent, Ken Miller, Jostein Ryssevik, Thorny Staples, Wendy Thomas, Rolf Uher, Mary Vardigan, Patrick Yott
Report on evaluation
In response to a request by NSF, the DDI conducted an evaluation of its objectives early in 2001. Four independent evaluators met with the DDI Working Group and then submitted reports. The reports were positive and offered constructive suggestions for further development of the DDI specification.
Report on status of August proposal
Work on an August 15 submission to NSF is progressing at the Roper Center and ICPSR. Committee members were asked to comment on the proposed intellectual agenda for the submission and to identify any significant omissions or extraneous tasks.
Hierarchical data specification. It was agreed that, while the aggregate specification is nearly completed and suggests a new approach to dealing with hierarchical and other complex file types, work still remains in this area, so this should definitely be a task for the proposal.
Comparative, multi-country, and time-series data specifications. The Committee also saw these tasks as essential. It was suggested that we should start the process of defining these models by developing criteria lists as guides to more comprehensive descriptions.
Tools. In terms of requesting funding to build tools, the Committee was divided. Some members saw value in having tools to encourage potential users to employ the standard. Others felt that tools should emerge from the user community and then be made available for sharing.
Interoperability test-bed. Having a forum for implementers to exchange marked-up resources and to test comparability was suggested.
Other ideas put forward for the proposal focused on developing a management strategy for stewardship of the DDI and a continuing emphasis on harmonization with other metadata standards.
Report of CESSDA Expert Seminar on the DDI
Reto Hadorn, of the Swiss Data Archive, reported on the meeting of representatives from the European archives that was held on Monday, May 14. This group discussed the implementation of DDI at each local site and described the procedures that were being used. The archives employ different solutions, but for many the DDI work presents an opportunity to clean up existing documentation and add to it. All the archives reported that the process of markup involves a lot of manual "handwork." Another issue many archives are facing is how best to group variables in rational groupings so that concept terms may be applied at the variable group level, rather than at the level of a single variable. Reto suggested that the DDI Committee consider developing a metadata management instrument in the form of a relational database to be shared among the archives.
Also on the seminar agenda was whether the Integrated Data Catalog would migrate to NESSTAR, how to deal with languages other than English in DDI mark-up, and controlled vocabularies. On the last point, Ken Miller, who is working on the LIMBER project to develop a multilingual thesaurus, suggested that the DDI incorporate controlled vocabularies by pointing to thesauri external to the DTD to avoid problems that would arise if subject terms were part of the DTD itself. This would have the added advantage of permitting the use of other specialized thesauri for fields such as medicine for which thesauri already exist.
At the Fall 2001 meeting of this group, more work on controlled vocabularies will take place. The group will also discuss recommended fields for the DDI, the language issue, and other topics and will submit recommendations to CESSDA, NESSTAR, and the DDI Committee after the meeting.
Procedures for revising the DTD
It was suggested that ICPSR set up a forum so that each proposed change to the DTD can be vetted and commented on. This should happen in advance of the June 29 meeting.
Outcome of tabular/aggregate data discussions in Voorburg
The Working Group meeting in April involving Wendy Thomas, Bill Block, and Emiel Kaper was successful and resulted in a proposal for extending the DTD to cover aggregate data.
Emiel showed a PowerPoint presentation to describe the "Voorburg Compromise." Revisions to the physical file description (Section 3) and the variables description (Section 4) are suggested. The proposal separates the logical description of data from its physical instantiation. Emiel set forth three criteria for acceptance of the proposal: it should provide for a logical description of the "ncube" or data matrix; it should provide for the description of physical storage; and it should provide the necessary functionality in terms of hinging, searching, and inheritance.
The proposal, some examples of markup using it, and Emiel's presentation will be added to the DDI Web site for review before the June meeting.
Remodeling the DDI specification
Emiel presented a draft of how the DDI might be reconceptualized as an object-oriented model. There was some discussion of whether we should follow this path or simply press forward with the model we have. The Committee was urged to think further about this in preparation for the June meeting. This presentation will be posted to the Web site as well.