Metadata and Digital Repositories

SIG Meeting Report: Higher Education Academy, York, 01 Mar 2006

Report by Neil Fegen, based in part on notes by Julie Allinson. We tried to record all the presentations, but the quality of the recordings was variable, with some completely inaudable. The others are provided as MP3 files, which we hope you find a useful supplement to the PowerPoint slides (even though they might be hard to follow at times).

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

A summary of the evaluation forms is also available.

Introduction

Phil Barker gave a brief introduction to the day and to the aims and activities of CETIS.

For more details, see Phil's PowerPoint presentation [63kB].

Introduction to Reference Models

Julie Allinson (of UKOLN) presented an Introduction to Reference Models, describing how the Digital Repositories Programme has some twenty-five projects - organised into clusters - and covers a cross-domain and range of issues. Julie also gave a summary of the issues raised and progress made at the Repository strand of the November JISC/CETIS conference, including the usefulness of reference models in clarifying 'known and unknown unknowns'.

For more details, see Julie's PowerPoint presentation [586kB].

A Reference Model for the ASK Project

Howard Noble presented work on the Accessing and Storing Knowledge (ASK) Project, which is funded under the JISC Digital Repositories Programme, and aims to build a repository or repository services. Key services from the e-Learning Framework referenced include: Authentication, Authorisation and Content Management. Howard concluded that the process of creating a reference model requires: the improvement of communication channels and facilitation of the process around development of interoperability specifications; and a vision for the interaction between a VLE and repository services.

For more details, see Howard's PowerPoint presentation [126kB].

Assessment item banks as repositories

Mhairi McAlpine from SQA presented on assessment item banks, describing item banking and the associated elements. An item consists of a question plus associated information, including: metadata, additional QTI metadata, resources (e.g. images, base QTI files, culturally appropriate stylesheets), feedback, etc. Two versions of an item bank infrastructure model were shown. Mhairi also talked about the importance of standards compliance and robust testing, and the implications for digital repositories, which included security and dynamic collaboration issues.

For more details, see Mhairi's PowerPoint presentation [354kB] and download Mhairi's presentation as an mp3 [6.95Mb].

An overview of OAIS

Michael Day from UKOLN talked about the OAIS reference model, with reference to: the background; definitions, high level concepts, and mandatory responsibilities; functional model; and information model. Michael mentioned other perspectives, such as preservation and archive interoperability. The implementation of OAIS was shown, along with conclusions from UKDA and TNA assessment.

For more details, see Michael's PowerPoint presentation [174kB] and download Michael's presentation as an mp3 [9.41Mb].

Discussion: Is OAIS suitable as a reference model for repositories?

OAIS as a communication tool?

Michael Day noted that the OAIS vocabulary is not uncontested so may not be the most neutral set of terminology. Expecting different communities to agree on a single vocabulary is not perhaps the best option. Mapping the OAIS terminology to that used in different areas would be a useful exercise and developing some crosswalk-type documentation.

OAIS as a reflection of practice?

There was some discussion of the preservation elements of OAIS and whether these would be necessary for repositories. Although true that not all repositories have a responsibility for preservation, it is something that repositories should be aware of. OAIS provides a model for communicating policy and practice relating to preservation and offers the flexibility for defining "long-term preservation". Many repositories are already fulfilling the mandatory responsibilities but that OAIS helps communicate these activities.

OAIS: reference model of framework?

Howard Noble pointed to Figure 6 in the paper: 'Service-Oriented Frameworks: Modelling the infrastructure for the next generation of e-Learning Systems', and asked whether OAIS would be classed as a Framework. There was some disagreement with this diagram and it was generally agreed that OAIS is a high-level reference model, rather than a framework, with lower level reference models falling within or beneath it. This lower level of detail is where connection with the services in the e-Framework would be identified. This relationship with the e-Framework was seen as an important one, but there was some feeling that the e-Framework is not dynamic enough at present. Current developments - such as the appointment of an e-Framework editor - are working towards a more dynamic e-Framework. It was noted that the e-Framework itself is not mandating any compliance and that 'bricks' might be coupled with non e-Framework services ('stones'). Other Frameworks also exist, e.g. BECTA, eGIF.

Repository: towards a definition

There was consensus on the first 3 points of the repository definition:

• content is deposited in a repository, whether by the content creator, owner or third party
• the repository architecture manages content as well as metadata
• the repository offers a minimum set of basic services e.g. put, get, search, access control.

• the repository must be sustainable and trusted, well-supported and well-managed.

It was felt that although important, this point is a value judgement rather than part of a core definition. This notion of trust, sustainability and management is central to OAIS, although OAIS does not mandate any real compliance. It is the role of policy to specify the extent to which a repository fulfils this point. If repository developers are being guided by a reference model, then are they already considering issues of trust and sustainability? If this point remains as part of the definition, does this exclude many of the growing body of popular repositories such as community wikis, or should these not be accepted as true repositories because they do not fulfil this aspect of the definition?

The reference model, by implication, defines what a repository is, but does not enforce a strict definition or any mandatory elements. It is a map of the space, of its services and data, etc. rather than a compliance checklist. It was agreed that it is not the intention of a reference model to exclude repositories. The definition is generic for this reason and changes in usage and landscape might necessarily alter this idea of sustainability and trust.

Adapting OAIS to support community needs?

PREMIS started with a framework explicitly based around the OAIS information model, but has since redefined its metadata model. OAIS concepts are implicit, but PREMIS created own event-aware model.

InterPARES modelling of specific ingest functions for records management

This returns to the idea that OAIS model is communication tool -- it asks questions about a repository's functions, information and processes. Production of a working repository is outside OAIS and investigating an OAIS-lite approach for a different domain is potentially an interesting application of the model.

Next steps and possible further work

Give feedback on the results of the discussion to the Digital Repositories Programme support team.

OAIS

• More detailed analysis and evaluation of OAIS, including its terminology, functions, information model and mandatory responsibilities
• Involve OAIS experts and representatives from the various stakeholder communities (e-learning, repositories, etc) in this analysis work
• Further explore issues raised through future SIG meetings?
• Mapping of OAIS elements to the repositories landscape, as identified through the developing typology/ecology
• Documenting and discussing this via collaborative mechanisms (wiki, jiscmail, etc).

Reference Models

• Evaluate any other candidate reference models, e.g. CORDRA
• CETIS reference models projects and repositories meeting for discussion of different implementations and approaches, identifying exemplars and best practices and opportunities for sharing lower-level detail of actual service implementations.

E-Framework

• Identifying the relevant e-Framework 'bricks' to fulfil functions identified in the OAIS reference model
• Contributing to the e-Framework, keeping the e-Framework dynamic and possibly extracting some definitions of concepts from the OAIS
• Identifying gaps in the e-Framework and existing/potential problem areas

Attendees

The following registered for the meeting

Many thanks to all who attended, especially to those who gave presentations and to those at the HE Academy who helped with the organisation.