View the presentation slides  (2.1 Mb)

Abstract

Metadata is all-pervasive in eResearch and high quality metadata is critical to the discovery, understanding and re-use of research results. Examples of metadata within eResearch span the spectrum: from simple Dublin Core for discovering publications; through highly individualistic, discipline-specific schemas that aid the classification and retrieval of physical artefacts; to metadata describing the settings of an instrument used during a particular run of an experiment. The problem of displaying this metadata in a form that facilitates fast, easy visualization, editing and input is also pervasive across disciplines and has resulted in a multitude of quick and dirty solutions to what is generally perceived as a simple but tedious problem.

Technologies employed have included: HTML[1], XForms[2], and various 4GL application builders. Documenting, maintaining and supporting this wide variety of applications, user interfaces and technical platforms is problematic and expensive. More importantly, ensuring that the metadata is accurate, complete and conforms to a given schema is an expensive, complex but critical task. Poor quality metadata due to human or machine error may result in lost, misleading or un-usable data or the need for expensive post-capture quality control, cleaning and enhancement. The aim is to streamline the capture of high quality metadata at the time of research data capture – making it as seamless and pain-free to researchers as possible.

Our collaborations with a number of research communities at the University of Queensland within the DART[3] and ARCHER[4] projects, have led to the development of a metadata editor service (MDE)[5] and associated Metadata Schema Registry (MSR)[6]. These services have been designed so that they may be easily incorporated into any web service-based process which needs to display metadata records for review, enhancement or editing. The MDE is dynamically generated from a given schema which is specified at run-time and retrieved from the Metadata Schema Registry (MSR). The dynamically generated input form, pull-down menus (based on controlled vocabularies) and validation mechanisms ensure that the loaded or input metadata conforms to a community agreed standard. By validating user input against the given schema (whenever the user enters or saves metadata), we ensure that the metadata is of the highest quality and conformity with regard to: structure, elements, value types, controlled vocabularies and constraint rules.

About the speakers

Ron Chernich is a Principal Research Fellow working in the field of metadata management and annotation services for eResearch within the School of ITEE at the University of Queensland. Ron has over 30 years of experience in IT working in Australia and overseas in commercial software development. On returning to Australia from the USA in 1998, Ron joined the Co-operative Research Centre for Distributed Systems (DSTC) at the University of Queensland as the Director of Engineering. He remained with them until the Centre closed in June 2006, driving the development of commercial software products from DSTC research outcomes. These included dCON, a Java implementation of the Object Management Group (OMG) cosNotification specification for asynchronous messaging, and MetaSuite, a complete solution for schema independent metadata incorporating editing, validation, searching, and retrieval. Both systems attracted government, academic, and commercial users. Currently, Ron is leading the development of a metadata annotation service for mutli-media digital collections that will become part of the Atlas of Living Australia.