Real-world metadata registries; sharing concepts, schemas and semantics.

Our current metadata schema registry initiative, the IEMSR, has recently celebrated its third birthday. The IEMSR is based on the principle that semantic annotation of resources is precisely as accurate as the user performing the annotation, meaning that the best way to improve the quality and use of metadata is to provide a more supportive environment for metadata producers and consumers. A metadata registry is an environment within which users may discover existing schemas, or collaboratively develop an appropriate set of concepts.

In this paper, we take a look at relevant research, describe the results of user testing, and fill out some of the gaps in our understanding in the practical use of schema registries. As a result of this review process, we provide a series of guidelines for those supporting schema or application profile development, and sketch out what the next generation of registries might resemble.

Some of the arguments underlying schema development are modern manifestations of very classical issues and relate to deeply-held assumptions in various fields. There’s a general multidisciplinary effort going on today that takes inspiration from all sorts of areas, from computer-supported collaborative work to philosophy to theories of construction of knowledge and understanding, in an attempt to comprehend and predict the result of user engagement with the semantic web and its lower-case or informal relatives. With the increasing popularity of self-consciously informal metadata creation and community-based Web 2.0 approaches to knowledge management, formalised processes are perceived by many as excessively long and complex. A well designed schema registry could support these processes by addressing some of the difficulties faced by users.

Terms are created according to their author’s understanding of the system, the resource and their own aims. But if the result is a socially situated artifact, grounded in the user’s unique context, what does this mean for sustainability? How do we support schema evolution and development? How can we encourage visitors to take the time to populate the registry?

This presentation provides an overview of issues and research trends; we hope to spark some interesting discussion and feedback.