Insights to Interoperability: Can you describe the scenario of the experiment that you demonstrated in the LIFE-fest in Paris?

Stefaan Ternier: In the LIFE-event we wanted to demonstrate how we can search learning resources that are in one Learning Object Repository (LOR), in this case within the FIRE-federation, through another repository, this being the ARIADNE-knowledge pool. The two repositories in this case were both made searchable through a LOM-based SQ-interface. Through these SQ-interfaces both LORs could search one another's networks.

So we had a very simple experiment: we uploaded a set of slides and indexed them in the ARIADNE repository. Next, we made a search using the FIRE-system, and were able to retrieve the same set of slides. So, even if the demonstration looked very simple for the audience, it was a demonstration that simple standards can work in practice.

Insights to Interoperability: What different e-learning standards were used for this?

Stefaan: The whole work on SQI, Simple Query Interface, is based on the use of existing standards. The SQI is a fruit of a collaboration between different European actors in the field of Learning Object Repositories. The SQI work was carried out in collaboration with EUN, ARIADNE and Educanext, to mention a few. These actors realised that to achieve a higher level of interoperability in propagating the search queries across the different repositories something like this is needed. The CEN/ISSS Learning Technology Workshop seemed a suitable place for this type of work.

The other standard used was LOM, Learning Object Metadata, which has been used to describe the learning resources in the format of metadata.

Insights to Interoperability: Why was this important for ARIADNE? And for the cooperation between digital repositories?

Stefaan: As ARIADNE's mission is to enable better learning through the development of tools that enable "share and reuse", a vast amount of learning object is necessary. By doing the (necessary) "plumbing" with standards behind the scenes, we aim to provide transparent access to a critical mass of learning resources to all ARIADNE users.

This was also one of the reasons for us to join forces with some major players on the global field, namely MERLOT, education.au, LORNET and NIME, through the Global Learning Objects Brokered Exchange (GLOBE). One of first steps of this alliance was to enable searches across all the repositories of the partners involved. In future, other organisations in the world will be able to join and contribute to this global network.

Insights to Interoperability: What will happen next? Vision?

Stefaan: As more resources are now unlocked and available for searching, simple queries yield more and more results. For that reason we started to develop a more advanced query language (plql), that enables the specification of more precise queries, resulting in less but better results. Through this query language, learning tools will be able to take into account a users profile, context and other relevant parameters to provide more meaningful results.

Insights to Interoperability: At the LIFE-fest there was a lot of interest in how SQI relates to the OAI protocol for metadata harvesting (OAI-PMH). Can you elaborate on that?

Stefaan: In my opinion, both standards can serve the same goals, while still being complementary. OAI-PMH is a standard that allows harvesters to extract metadata from a repository. Search clients can then issue queries on their local copy of the metadata.

With SQI, a query is shipped and executed at the repository premises. This has the advantage over OAI-PMH: we always search in the most recent version of the metadata and that search clients can be very light-weight, as there is no need for them to implement huge metadata stores.

The disadvantage is that each repository in a network gets hit for each query that is sent into the network. Depending on the use case and requirements, one could opt for either or both solutions. A repository that wants to enable searching, but has intellectual property rights on its metadata (like MERLOT), is better off with SQI, as it cannot give away its metadata. On the other hand a repository that has limited bandwidth or resources, might be better off with harvesting. Nothing, however, prevents a repository to implement both standards. We do envision future scenarios where we will wrap OAI-PMH harvested metadata into SQI enable metadata caches.

Insights to Interoperability: Thanks, Stefaan, and thanks for participating in the LIFE fest too!

LIFE: http://life.eun.org
SQI: ftp://ftp.cenorm.be/PUBLIC/CWAs/draft/SQI_V1.0beta_2005_04_13.doc
http://www.prolearn-project.org/lori
ARIADNE: http://ariadne.cs.kuleuven.be
About Query Languages: http://ariadne.cs.kuleuven.be/lomi/index.php/QueryLanguages_stable
FIRE: http://fire.eun.org
GLOBE: http://globe.edna.edu.au/globe/go
OAI: http://www.openarchives.org/OAI/openarchivesprotocol.html