This paper presents the outcome of a study performed in the Netherlands on handling multiple metadata streams regarding digital learning material. The paper describes the present metadata architecture in the Netherlands, the present suppliers and users of metadata and digital learning materials. It furthermore describes the roles of suppliers and users that were defined. Based on the needs of the present and new roles, and based on the present architecture the necessary changes to the metadata architecture are described to make it more future-proof and to ensure it can handle multiple metadata streams.
In the last couple of years, the use of metadata to be able to search and find appropriate learning materials has increased drastically. With this increase, more organizations are interested in producing digital learning materials. To be able to find appropriate learning materials, metadata records need to be provided. In the past, most digital learning materials were created by publishers, who also provided metadata records for their materials. Now that not only publishers are creating digital learning materials, but also schools, individuals and many more, the metadata records for one piece of learning material are not provided by one group anymore. These new groups of people and organizations providing metadata create new streams of metadata records. Metadata streams are metadata records flowing from the creator of the record, to the central database where the records are used to find learning materials. These new streams were not foreseen when developing the present metadata architecture1. Instead of handling only one metadata stream, the educational content chain now has to handle many streams, all with different characteristics, and often providing additional metadata information about the same piece of learning material. This leads to difficul1 A definition of how the systems handling metadata interoperate. ties in the educational content chain2, as the present architecture is based on the principle of only one metadata record for each piece of learning material.
Based on these developments in the educational sector, TNO [
Kennisnet is the foremost public knowledge centre regarding information and communication technologies (ICT) and education in the Netherlands. TNO is an applied scientific research company that applies scientific knowledge with the aim of strengthening the innovative power of industry and government. Verdonck, Klooster & Associates (VKA) is an independent consultancy company with extensive experience at the intersection of strategy, implementation and ICT in the public sector. 2
Constant developments in society, and the need of the Netherlands to extend their competitive knowledge industry, lead to a need for changes in the Dutch educational sector. A major development in the educational sector is the need to educate people during their entire life; education does not stop anymore after obtaining a degree. The digital era provides huge possibilities to support lifelong learning, and furthermore the present generation of students has high expectations about the digital possibilities in education.
These developments lead to more flexible education, and therefore ask for new
educational material: digital material. To ensure a good user experience, and to ensure
interoperability and reusability of digital material there is a need for agreements.
Kennisnet therefore developed an educational content chain to make agreements
about the use of digital learning materials and to ensure interoperability, reusability
and usability [
The present content chain has implicitly been developed with only one main stream of
provisioning (authorative [
Recent developments though, show more and more sources providing complete metadata records, or supplying additional metadata to the original record. These new records lead to more than one metadata record in the central database for most of the learning materials. The availability of multiple metadata records for one piece of learning material in the present content chain makes it difficult for users to locate useable learning materials as the search results show one piece of learning material multiple times (once for each metadata record).
Furthermore these multiple records also have another effect: users cannot search in
the richest metadata, resulting in not finding useable learning materials. Table 1
shows two metadata records for the same piece of learning material. If a user searches
for ‘Geography’ as course, he will find the material. If he searches for ‘Layer of the
earth’ he will also find the material. When a user searches for the combination of both
keywords the present content chain will not return the material as the combination of
the key words are not available in one metadata record.
The study showed that it is difficult to the characterize parties (people or
organizations supplying learning materials and metadata). There is no such thing as a general
publisher or a general user. Each party has different requirements and performs
different roles, as has been argumented in [
The needs of the roles are: suppliers, prosumers, and scouts want to provide metadata to make the learning material findable. Enrichers, reviewers and users want to provide additional metadata to make the material more findable for specific target groups. 5
Based on the defined roles and their needs, a new architecture for the educational
content chain was defined. In this architecture every role can provide the information
they define or have present. Figure 2 presents the old and new architecture. The new
architecture includes everything that is in the figure; the old architecture has the same
structure, but only consists of the roles and records that are marked in a grey. On the
left side of the figure the different roles are represented. The different roles are
divided into three groups to give the users of the search engine the possibility to make a
distinction while searching. Because of the distinction in three groups, users have the
possibility to search only in metadata records provided by the source of the learning
material. Each group has its own contact point at the central search platform (Edurep
[
Source: Suppliers
Scouts Prosumers
Expert group: Enrichers Reviewers User: Users
Complete metadata records Additional metadata information
Reviews
Complete metadata records Additional metadata information
Reviews
Edurep
Searching metadata records, distinction between three groups:
Source
Source + Expert Source + Expert + User
Searching reviews, distinction between three groups:
Source
Source + Expert
Source + Expert + User
ommendation that was defined is: “Ensure that each piece of learning material has a
unique identifier, and ensure that the user can search in the richest metadata
information (a combination of all available metadata information) available” [
By ensuring that each piece of learning material has a unique identifier, it is
possible to reference different metadata records that apply to the same piece of learning
material. If these references can be made by the central database, it is possible to
search in the richest metadata possible. It is furthermore possible to combine multiple
pieces of metadata to create the best set of metadata for one piece of learning
material. To ensure unique identifiers for each piece of learning material, agreements have
to be made on how to generate unique identifiers. The suggestion made in the study is
to use a formal URN (Uniform Resource Name) [
Furthermore, two other recommendations have been defined based on the study: “Ensure that the educational content chain is managed” & “Provide a service to add additional metadata”. The first recommendation is important since the management of the content chain is currently not explicitly defined. The second recommendation is important to ensure that all roles defined have the possibility to provide metadata. Most roles already have some kind of tooling available, but f.e. for enhancers and reviewers, no tooling is available.