D26 - Metadata Watch Report #5

Home

Project

Partners

Metadata Watch Report #5

APPENDIX B: Domain report: educational sector

Correspondent: Erik Duval, Univ. Leuven

1. Introduction

This report focuses on implementation activities, with a particular emphasis on multilingualism and controlled vocabularies.

2. Standardization work

2.1 IEEE LTSC LOM

Within the IEEE LTSC Learning Object Metadata (LOM) specification, controlled vocabularies are defined for 19 elements.

Central to support for multilingualism in LOM is the notion of LangString: when element values are defined as free text, lingual variants of the values can be provided in a number of human languages.

2.2 CEN/CENELEC LTWS

The CEN/CENELEC ISSS Learning Technologies Workshop includes both multilingualism and controlled vocabularies as major work items. With respect to multilingualism, three project teams are working:

to ensure that the IEEE LOM, as the globally accepted solution, is capable of addressing specific European cultural requirements (such as multilingualism).
on standardization actions to permit the identification of alternative versions of resources, in different languages, as well as the origin of the translation. This is taking place within a LOM context.
to ensure that LOM is localized and translated in the languages of the EU and EFTA countries. Translations of earlier versions of LOM already are available from the Workshop's web pages. These will be replaced in due time by updated and widely accepted revised versions.

On controlled vocabularies, a project team is collecting and organizing a register of taxonomies and repositories relevant to European learning, via an on-line repository. This will benefit interoperability between European learning technology systems and

services as metadata implementations will be able to rely on standardized taxonomies and vocabularies. It is expected that many will be developed and implemented at national level. Focus is on the identification of existing taxonomies, their applicability and interrelationships. Where possible, mappings or translations will be made between various taxonomies and vocabularies used in multilingual and multicultural learning domains.

3. Consortia based work

The consortia mentioned below, some of which contribute to the development of standards, adapt these standards to the needs of their constituencies, a process referred to as 'profiling' in the standards world.

3.1 ARIADNE

The ARIADNE Foundation develops and exploits the Knowledge Pool System, a distributed database of reusable learning components, with associated metadata that describe them.

ARIADNE has developed controlled vocabularies of science types, disciplines and sub-disciplines: this is a hierarchical structure of 2*10*10 entries, used to indicate the semantics of a learning object. The ARIADNE tools draw heavily on the LTSC LOM vocabularies. In some cases, the ARIADNE application profile restricts the value space to a subset of the LOM value space.

The latest generation of the indexation and query tools (based on the recently balloted LOM v6.1) supports multilingualism in a number of ways.

The language of the user interface can be selected from seven languages.
Controlled vocabularies have also been translated in all these languages, so that, for the elements with controlled vocabularies, the selected or retrieved values are also translated.
For all elements with free text values, the user can insert or search for a number of lingual variants, with an indication of the language(s) being used, as per the LangString construct in LOM.

3.2 IMS

The IMS consortium re-synchronized its metadata specification in May with the recently balloted LOM v6.1: the IMS information model v1.2 includes the LOM specification.

The 'Best Practice and Implementation Guide' includes suggestions of 37 schemes and controlled vocabularies for 13 elements. When no single vocabulary is judged to be dominant, multiple suggestions, with their origin and suggested applicability, are listed. Schemes are listed for elements such as language (RFC1766, ABS1267, ISO639, ISO3166 and Z39.53). Vocabularies are listed for elements such as classification (LOC, LCSH, DDC, UDC, CIP, SCIS, GEM, etc.)

No mention is made of multilingualism in the guide.

3.3 ADL & SCORM

The ADL Sharable Course Object Reference Model (SCORM) profiles LOM for raw media, content and courses.

A 'curricular taxonomy' is defined as the controlled vocabulary for the content structure hierarchy, mapping the U.S. Army, Navy and Marine Corps, as well as the Canadian designs to raw media, content and courses. For 8 CMI (Computer Managed Instruction) API data elements, vocabularies are defined. These are lists of 2 to 8 values.

SCORM does not address multilingualism explicitly.

3.4 EDNA

EDNA relies on DC with 8 additional elements (Audience, Approver, CategoryCode, Entered, Indexing, Review, Reviewer and Version).

For certain elements, schemes are specified: for DC.Subject for instance, reference is made to LCC, LCSH, APSDEP, SCIS, ASCED, DDC, UDC, MeSH and edna-KLA. For other elements, controlled vocabularies are listed, such as for DC.Type, where vocabularies are

defined for document types, curriculum types and event types. Similar vocabularies are defined for audience, sector, user level and coverage.

With respect to multilingualism, the only explicit reference is to use the Alternative refinement for the Title element to include translations.

3.5 EUN

The EUN specifies two metadata element sets. For learning objects, it relies on Dublin Core, with additional elements for rights, approver, release, user level and version. For collections, the specification is based on RSLP and Renardus.

In the 'European Treasury Browser', an extensive controlled vocabulary has been developed that refers to individual development, learning and research, school activities, leisure activities, teaching and training and evaluation and guidance, the educational system, subject, facilities, communication, culture, political and social aspects, health, environment, society, international organisations, geopolitical areas and languages. Other vocabularies are provided for the user level, type. For pedagogy, the GEM vocabulary is used.

In order to support multilingualism, the EUN specifies for which elements language specification is necessary. Moreover, the subject vocabulary is available in English, French, German, Italian and Spanish. Swedish, Danish and Greek versions are under development.

3.6 Gateway to Educational Materials

The Gateway to Educational Materials (GEM) extends the Dublin Core element set, with 8 additional elements.

Controlled vocabularies are defined for format (17 values, based on Internet MIME types), grade (8 values) and resource type (29 values, incl. Course, event and tool). More elaborate vocabularies are specified for

Audience: 18 values to indicate for whom the tool is intended and 46 for the ultimate beneficiary;
Pedagogy: This is clearly the most elaborate GEM vocabulary: three categories of pedagogy are distinguished:
teaching methods: 37 values detail the type of educational activities, including for example cooperative learning and role playing;
grouping: 7 values, ranging from individualized to large-group instruction;
assessment: 10 values, including peer evaluation and portfolio evaluation.
Relation: 20 relation types (incl. IsParentOf and isCriticalReviewOf);
Subject: a hierarchical structure with circa 15 top-level elements and around 15 elements for each top-level element at the second level.

There is no explicit reference to multilingualism in the GEM specifications. Interesting to note is that the controlled vocabulary for the language elements includes 5 values (only).

4. Projects and Research Activities

Many projects include metadata tools and infrastructures in their R&D efforts. The first results of this research are beginning to appear in publications. As an example, the recent ED-MEDIA conference included about 30 papers on applications of LOM and 10 that refer to Dublin Core [Vitelli and Montgomerie, 2001]. Publications that focus exclusively on educational metadata applications (both LOM and DC-Education based) include [Greenberg, 2000] and [Duval & Robson, 2001].

Often, the conceptual remarks based on the experiences of the authors are quite relevant. Sometimes however, the comments point to 'features rather than bugs', in the sense that the original developers of the metadata specification had anticipated some of the problems. In those cases, the typical consensus in the standardization bodies was that it was better to have a 'good enough' solution now, rather than a 'perfect' solution in an all too distant future.

Many of the metadata tools seem somewhat immature: they typically start from the LOM specification, rather than from the user characteristics and tasks. Hence, the usability of many of these tools is rather poor. It can be expected that his situation will improve: whereas the current generation of tools is heavily focused on demonstrating how the technology can work, (we hope that) future tools will concentrate more on helping the user to get a job done.

5. References

[Ariadne, 2001] http://www.ariadne-eu.org/

[Duval, 2001] Erik Duval. Standardized Metadata for Education: a Status Report, Proceedings of ED-Media 2001: World Conference on Educational Multimedia, Hypermedia and Telecommunications, Tampere, Finland, pp. 458-463, June 25-30, 2001.

[Duval & Robson, 2001] E. Duval and R. Robson (eds.). Special edition on Metadata, Journal on Interactive Learning Environments, 2001.

[Greenberg, 2000] J. Greenberg (ed.). Metadata and Organizing Educational Resources on the Internet. Journal of Internet Cataloging, Vol. 3, No. 2/3, 2000.

[Vitelli and Montgomerie, 2001] J. Vitelli and C. Montgomerie (eds). Proceedings of ED-Media 2001: World Conference on Educational Multimedia, Hypermedia and Telecommunications, Tampere, Finland, June 25-30, 2001.

Maintained by: UK Office for Library and Information Networking (UKOLN)
Last updated: 01 October 2001