Metadata and Digital Repositories

SIG Meeting Report: University of Liverpool, 30 June 2004

Report by Phil Barker

A list of attendees is given at the end of this report.

Introduction

Phil Barker gave a brief introduction to the day and to the aims and activities of CETIS. [Introduction, powerpoint presentation, 62kB.]

Specification Update

Phil also gave a brief update on concerning some spec developments, which touched on:

• the release as a public draft of IMS Learning Resource Meta-data version 1.3
• the circulation of the second draft of the UK LOM Core
• work on creating a metadata FAQ
• the release as a public draft of IMS Resource List Interoperability
• the release as a public draft of IMS AccessForAll Meta-data

[Recent Spec Developments, powerpoint presentation, 146kB]

Metadata Workflows in Learning Object Repositories

Jane Barton, of the Centre for Digital Library Research at Strathclyde University gave an overview of a new JISC-funded project at CDLR that will investigate the practicality and sustainability of different approaches to the creation of metadata for institutional learning object repositories and ePrint archives. The project will attempt to unpack what is meant by "metadata quality" and will investigate how this is supported by the UK LOM Core and Dublin Core, covering not just the suitability of metadata for discovery and retrieval but also for the management and preservation of resources. The results will support JISC activities such as the X4L and FAIR programmes and the JORUM. The timescale for all of this work is tight, with the final report being due in September, so consultations, reviews and case studies will take place over the summer.

[For more details see Jane's powerpoint presentation, 64kB]

Metadata Collection the HLSI Way

Kieran O'Farrell of the High Level Skills for Industry project gave a description of how that project had addressed problems with the quality of the metadata being entered into their repository by content producers. A review of the metadata in the repository by metadata specialists (librarians) had identified problems such as metadata descriptions that related to the format of the object rather than its content (e.g. "a flash object"), over zealous use of cut-and-paste, and template defaults that were never changed. Kieran reported how a working group of librarians, a subject tutor and a software developer had suggested changes to the metadata creation tools, work practices and systems in order to improve the quality of the metadata. This included:

• Simpler entry forms and wizards for resource authors to add basic metadata
• Using library-trained metadata specialists to complete the metadata authoring.
• Using language appropriate for UK Further Education
• Sparing use of templates
• Creating a good practice guide.

[For a fuller explanation see Kieran's slides, powerpoint, 250kB]

JISC Information Environment Metadata Schema Registry

Pete Johnston of UKOLN presented an overview of a new Jisc funded project to create a Metadata Schema Registry for the JISC IE. The  project will provide a pilot metadata schema registry as a shared service in the JISC IE that will provide machine (and human) readable information about metadata vocabularies used by other services in the IE. It is hoped that the disclosure of this information will aid interoperability by enabling the discovery and reuse of metadata terms and vocabularies in common use and providing pointers to supporting materials such as use guidelines, bindings, transformations etc.  The Schema registry would thus store information on metadata application profiles being used and would support the development of interoperable application profiles. Pete then described some of his work on modeling Dublin Core and LOM application profiles, contrasting the selection of single terms from the "flat" Dublin Core to the "pruning" of branches from the LOM hierarchy, the latter of which had not been well represented in previous metadata schema registry work.

[Pete's slides are available on the UKOLN website in Powerpoint and HTML format]

Harvesting LOM Records Over OAI

First, here's something which wasn't presented at the meeting due to lack of time. I had prepared some background to the RDN-LTSN interoperability project, covering who the RDN and LTSN are, what they do and where they overlap in what they do, and a general overview of the OAI-PMH based approach taken to making sure that they approach the overlap in their remits without duplicating effort. [This summary as powerpoint slides, 107kB]

Paul Hollands, of LTSN-01 gave a presentation based on his work in developing software and guidelines to support the RDN-LTSN partnership projects. Paul described the approach which has been taken: creating a profile of the LOM for this work (the RLLOMAP, which though it was developed independently from  the UK LOM Core has been made compatible with it); adapting OAI-PMH software which supports to Dublin Core so that it also support the RLLOMAP; and some of the problems encountered on the way (with his solutions). I don't think Paul will mind being described as a "constructive skeptic" when it comes to the LOM and the UK LOM Core and his work has been useful in highlighting issues such as the difficulty of applying these specifications to the description of a broad range of resources, the lack of OAI tools for harvesting LOM records, and confusion caused by difficulty in forming a collective understanding of the LOM and how to use it.

[Paul's presentation is available on the LTSN-01 website at http://www.ltsn-01.ac.uk/static/slides/metadata_sig1.html

Accessibility Properties for Learning Resources

Finally Andy Heath gave an account of work aiming to provide resource metadata that would work in synergy with the information about a user's resource preferences given in formats such as IMS ACCLIP (the Accessibility Learner Information Package specification). The IMS AccessForAll Meta-data Specification is one example of such resource metadata, and Andy focused on his work with CEN/ISSS APLR which is looking at how to create an application profile of the LOM for this type of metadata. Andy showed how there were two approaches possible: one based on re-using what was already in the LOM and thereby benefiting from the current adoption of these elements; and the other highlighting the accessibility metadata as separate, highly constrained and therefor more interoperable, category extending the LOM. Andy showed how the debate between proponents of the two approaches had progressed, but at this time was not settled.

[Andy's presentation is available as in Powerpoint, 63kB, people interested in the discussion can join the APLR discussion list and view the archive.]

Attendees

Many thanks to all who attended, especially to those who gave presentations and to those at the UK Centre for Materials Education who helped with the organisation.