Page History

1) What is the purpose of the TERENA OER portal initiative? (The motivation and the high-level architectural issue)

Fundamental questions must be clarified in a white paper about:

• how we can be different form others;
• what kind of repository we are proposing;
• what objects we are collecting;
• what tools we are planning to use;
• who the target user community is;
• how this initiative can be positioned in the context of global and other national activities
• and so on...

AGREEMENT

A paper must be drafted that summarizes the basic principles behind the TERENA OER portal initiative, gives a high-level view on the problem space in a European/global context, as well as define a common language, system architecture, and framework for recommendations. Potential issues must be addressed by the paper are:

• Dealing with multiple languages
• Understanding the users' requirements
• Supporting national repositories (with knowledge and tools)
• Aggregation of open materials
• Flexible metadata schema
• Focus on easy-to-use tools and procedures
• Understanding on existing metadata and practices already available in the community

The paper can then be agreed and presented at LAK2013 (April 2013, in Leuven Belgium) and TNC2013 (June 2013, in Maastricht, Netherlands)

ACTION 1 on Eli Shmueli (IUCC) to draft a paper (bullet points are sufficient to start with) and circulate it on the mailing list. Others are invited to contribute.

2) TERENA OER pilot implementation (The technical issue)

Proof of Concept metadata harvesting from SWITCHCollection to the MAOR test portal has been done.Look at the URL below:

http://maor.iucc.ac.il/t_materials.php?sort_property=dateCreated&page=1&search_q=keywords%3Dmaor%26anyKeyWords%3Dtrue%26allKeyWords%3Dfalse%26exactPhraseKeyWords%3Dfalse

Lessons learned: it is technically possible, but there are issues with both a) the metadata quality and b) the openness of content.

Other metadata to harvest:

• Vicente Goyannes (Uni Vigo) offered Campus do Mar: http://tv.campusdomar.es/oai.xml

• Gytis Cibulskis (Kaunas University of Technology) offered ViPS records archive

The next steps we should take are:

• Start small with sharing and aggregating of (fully) open materials
• Look around what other projects do (in Europe, Brazil, etc.) not to reinvent the wheel.
• Promote the use of the European portal and involve in global initiatives.

AGREEMENT

A deep technical dialogue must be started among the portal technology developers (MAOR, ARIADNE), the national repository owners (SWITCHCollection, Campus do Mar, Kaunas University, etc.) and the university people representing the end-user community. The discussion should include topics as follows:

• Metadata schema for aggregation, mapping, translation, etc.
• Potential functional integration of ARIADNE (as an aggregator) and MAOR (as a portal)
• Tools, protocols, procedures, implementation options.

ACTION 2 on Peter Szegedi (TERENA) to Doodle for a technical meeting as soon as possible.

3) Mini-survey about metadata (The metadata issue)

Concerns have been shared by the attendees about the finding of good quality metadata, the willingness of content owners to improve metadata, and about the aggregation of metadata in an easy way.

AGREEMENT

University people participating in TF-Media must be our anchors in the community to figure out what's going on in this field at the national/institutional levels. What kind of metadta is generated, how that existing metadata can be exposed, what tools are there for aggregation/harvesting, etc.

ACTION 3 on Giannis Stoitsis (GRNET) to draft some questions for the mini-survey and circulate them on the mailing list. After agreement, the final set of questions can be sent to the national/institutional repository owners as well as to the academic community for gathering answers.

Anchor
4th meeting 4th meeting

4th Technical VC Minutes (16 November 2012)

http://emeeting.campusdomar.es/r_public_list/32e7d3cb8dee8f254f462fb57c1bef0a

Anchor
5th meeting 5th meeting

5th VC Minutes (21 February, 2013)

Date and Time:

Thursday, 21 February 2013 at 16.00 CET

List of participants:

Kostas, Zenon (GRNET), Vicente (UVigo), Antonio (ISEP), Eli (IUCC), Jose-Maria (RedIRIS), Peter (TERENA), and a guest speaker Sorel Reisman (California State University, MERLOT)

Notes:

• Peter (TERENA) gave a brief update on the actual status of the TERENA TF-Media OER pilot concerning the 8 open questions listed and discussed at the previous (4th technical) meeting (see above). Peter asked for a round of updates:
  
  • Vicente (UVigo) reported that  he is working together with Giannis (GRNET) to implement the 1st "proof of concept" TERENA OER system by using software components of the two corresponding projects: Campus do Mar and ARIADNE. The pilot system will use the ARIADNE aggregation engine and metadata store as well as the PuMuKit web portal front-end. The simple metadata schema will be based on both IEEE LOM (maintained by IEEE institute) and Dublin Core (maintained by Dublin Core Metadata Initiative).
  • Eli (IUCC) reported that he is willing to offer a local installation of the MAOR system to be considered as the 2nd "proof of concept" TERENA OER system for comparison. Eli introduced Prof. Sorel Reisman (California State University) who will update the meeting participants about MELOR initiative in the US.
  • Jose-Maria (RedIRIS) commented that he is happy to participate with the ARCA repository however, ARCA speaks RSS only. It was said that the TERENA OER system's aggregation engine should be flexible enough to speak both RSS and OAI-PMH to the local content repositories.
  • Antonio (ISEP) summarized the sub-task of the pilot group as follows: 1) architecture design, 2) pilot implementation, 3) partnership. It was agreed that the pilot implementation phase needs a common understanding on the service architecture design first. The clarification of the architecture design is the highest priority, before we start selecting software tools for implementation!
• Prof. Sorel Reisman (California State University) talked about the architecture of the MERLOT (Multimedia Educational Resources for Learning and Online Teaching) system initiated by California State University in the US. It's been developing and running for 15 year now. The service community includes both academic and industry partners. MERLOT has a three levels deep layered architecture. The web portal interface is changing as we speak, trying to be much simpler (user-friendly). The metadata aggregation is using a proprietary subset of various standards. It stores and searches metadata in several languages and can turn the search results into an RSS feed. There is a federated search in about 20 libraries connected. Several web services are available (e.g., simple search, key search, etc.), some of them are for free of charge to the academic community. MERLOT is primarily a metadata repository (participants can contribute by exposing their metadata - batch input process) but there is a brand new "content builder" function of MERLOT where content can be managed.
• Peter (TERENA) commented that the TERENA OER architecture should have as modular/layered design as the MERLOT architecture has. There should not be a competition between the UVigo+GRNET vs. IUCC implementations but a potential collaboration to a common service architecture for TERENA OER. To be able to define the details/potentials of such a technical collaboration, better understanding on the actual components and functions of both the ARIADNE and MAOR systems is needed!

Agreements:

1. TERENA OER must be a stand-alone OER system that does not use ARIADNE or MAOR services but builds on the (open-source) software components developed and made available by ARIADNE and MAOR projects.
2. The architecture of the TERENA OER system must be modular/layered (including the aggregation engine, metadata store, and web services) preferably combining the best of the existing tools such as ARIADNE, MAOR, PuMuKit, and others.
3. In order to start implementing the TERENA OER pilot, a better/deeper understanding on the actual software tools and provided functions is needed. Therefore, Peter (TERENA) calls for a small technical meeting between Eli (IUCC), Giannis (GRNET), and Vicente (UVigo) as soon as possible. The pilot group will be informed about the outcome of the technical meeting.

Anchor
6th meeting 6th meeting

6th VC Minutes (1 March, 2013)

Date: 1 March 2013 at 13.00 CET

Invited participants: Giannis (GRNET), Vicente (UVigo), Eli (IUCC), Peter (TERENA)

We've all started to converge to a common concept that we have to eventually implement as a TERENA OER.

An important intermediate milestone would be the TERENA Networking Conference (3-6 June), where three potential things will happen:

• • Vicente will present the MOOC concept to the TERENA Technical Advisory Council (TAC) also mentioning the TERENA OER as one of the key building blocks.
  • There will be a session dedicated to "Teaching and learning online with multimedia". Take a look at the talks accepted at https://tnc2013.terena.org/core/session/3
  • We'll have chance to demonstrate the TERENA OER proof-of-concept portal over the coffee and lunch breaks.

Hopefully, these things together will set the scene and prepare the ground for an official TERENA OER pilot project approved by the TEC and funded by TERENA. This will only happen after the TERENA Networking Conference. So, our first priority would be to work towards the proof-of-concept TERENA OER to be demonstrated at TNC2013. I have identified the following actions today:

1) Peter (TERENA) to come up with the basic functional requirements for the OER web portal that might have implications on the entire aggregation architecture. To be sent to Giannis.

2) Peter (TERENA) to come up with some design principles (i.e. look and feel) for the actual web interface and/or point to an re-usable example. To be sent to Vicente.

• • TERENA visual elements
  • Preferable portal example: https://videos.unileon.es/

  • UVigo: We can prepare a mock-up that way with the TERENA look and feel but, we must wait for the metadata analysis in order to know if it will be doable, I guess.

3) Giannis (GRNET) to propose a metadata schema (potentially LOM) to be used by the proof-of-concept aggregation engine. In principle, we have to use the metadata that is available but we have to put efforts in finding the repositories that can expose 'good quality' metadata to us.

4) Giannis (GRNET) to distribute the analysis of the metadata survey. Based on that, we have to select the actual providers for the proof-of-concept (the candidates are MAOR, SURF, PSNC, UVigo, ARCA).

• • TERENA harvesting report

  • Following the analysis of the survey we are now working on the analysis of the existing metadata provided by the candidate providers. This is connected with the paper that we are preparing for the LACRO workshop and will allow to define an element set for the LOM based metadata schema for TERENA OER. This will be ready by 1st of April. Kostas will then transform all the harvested metadata to this LOM based metadata schema that will be used by the first version of the TERENA OER portal for searching and browsing services.

5) Giannis (GRENT) to distribute the ARIADNE RESTfull API description to the group.

• • ARIADNE Repository Services REST Interface
  • ARIADNE federated search

6) Vicente (UVigo) and Eli (IUCC) to clarify the solution how the PuMuKIT and/or the MAOR web portals can communicate with the ARIADNE RESTful API.

• • We have done a proof of concept working with the ARIANE API from PuMuKIT and it runs very well. We are just waiting for the final API definition. It can be also great to have a "dummy" aggregator, with the same configuration of the final one, to test with. Maybe Giannis can help us on this.
  • DRAFT service architecture (by Vicente):

Anchor
7th meeting 7th meeting

7th VC Minutes (27 March, 2013)

Invited participants: Giannis, Kostas (GRNET), Vicente (UVigo), Eli (IUCC), Peter (TERENA)

Some notes and actions:

1) Eli, to come up with an Excel sheet proposing mandatory, recommended and optional metadata fields for the OER portal pilot. After several iteration rounds among us, we have to agree on a basic metadata field set that we can recommend.

2) Giannis and Kostas, to distribute the preliminary analysis of the actually harvested metadata fields from 9+ different repositories. The result of the analysis will then be checked against our recommendation in order to identify the "friendly repositories" with good enough metadata to work with.

3) Vicente, to clarify the web portal functional design based on the recommended metadata fields (1) and taking into account the result of the metadata analysis (2). Also, to send the first TERENA-branded design draft of the portal to Peter.

4) Vicente, Giannis and Kostas, to set up a virtual meeting about the technical details of the ARIADNE and PuMuKit interaction.

5) Eli, to send the target URL of the MAOR metadata exposure mechanism (the web service) to Giannis in order to harvest high quality metadata from the MAOR repository.

Further agreements:

• TERENA OER should focus on audiovisual content in the first place (having something catchy to show off). Later on, other type of materials such as data sets, papers, etc. can also be collected.
• We should focus on (work with) only 2-3 "friendly repositories" in the first place.
• We have to have information on the actual categories of the content in order to represent them into the PuMuKit-based web portal. We have to understand how different taxonomies applied by different repositories can match for instance with the UNESCO taxonomy of categories.
• Translating key words to categories or mapping different taxonomies are not trivial issues therefore out of the scope of the pilot (for now). We have to use that categories that we are getting from "friendly repositories".
• Eventually, the TERENA OER pilot can result in best practices and implementation options for those repositories that has no adequate quality metadata.

8th VC Minutes (14 May, 2013)

Participants: Giannis, Eli, Vicente, Kostas, Ruben, Peter

Referring to our discussion today, I would kindly suggest to forget about the overall system/service design and ideal architecture for a moment and let's just focus on a potential tangible demonstration that we can give at TNC2013.

The aim of the demo would be to show the potentials and benefits of a pan-European OER metadata aggregation portal via catchy multimedia contents openly available for teachers and students.

• We want potential end-users to say: Ah, I can easily find/re-use open educational multimedia resources linked to each other in a thematic structure throughout several national repositories. I can easily improve my teaching/learning experience with these.
• We also want potential content owners to say: Ah, this initiative is great, my content repository is small but joining this federation we can make our content visible as well as we can access others' relevant content easily. How can I join?

To achieve this objective we need:

1) a web portal prototype (It's already done by UVigo)

2) a metadata aggregation engine working (to show the harvesting via standard protocols)

3) at least three different national/institutional repositories with metadata fields matching to the portal needs.

The harvesting engine part is not that critical as it has no impact on the user experience during the demo. All metadat snapshots can be pre-loaded into a local database, doesn't matter. We can explain how the API-based harvesting should work.

The main question is which repositories contain the metadat fields required by the portal? The web portal needs at least:

• Title to show
• Keywords to search
• Valid URL to navigate to
• Thumbnails' URL to show (optional)

ACTIONS:

1) Eli to provide a metadata snapshot of the MAOR repository off-line (XML). This can be pre-loaded into the local portal database for the demo.

2) Vicente to make sure that the UVigo repository makes the thumbnails available for the portal.

3) Peter to figure out if and how SWITCHcollection repository makes the valid URLs available.

4) Kostas to double-check the latest metadata harvesting results against the portal's needs (4 metadata fields described above) and let us know.

Let's see if we can progress by Tuesday next week and have a VC again. Otherwise we have to postpone the demo, I'm afraid.

Latest RESULTS:

• Here you will find the portal mock-up http://terenatv.media.uvigo.es/
• And here a rough explorer of the GRNET-Ariadne database: http://terenatv.media.uvigo.es/explore.php

3)

Kostas provided the followings:

• Latest repository/metadata analysis (set of XLSs) and presentation.
• Presentation at LAK2013/LARCO conference.
• List of selected repositories for the portal prototype.

4)

Michael Stücheli (SWITCH) provided the architectural design of the Swiss national repository portal. SWITCH is happy to share knowledge with the TERENA OER pilot participants. 

9th VC Minutes (9 July, 2013)

Participants: Giannis, Eli, Vicente, Kostas, Peter

Vicente said that the proof-of-concept portal works but the usability is poor due to missing metadata information such as direct URL to open content or thumbnails of multimedia resources.

Kostas also acknowledged the fact that it's hard to find the location of the matadata fields by the Ariadne harvesting engine due to the various schemas used by the repositories.

There are two ways to solve this:

1. We come up with the recommendations/agreements on where to put the metadata.
2. Repositories define where they put the metadata so we know where to harvest from.

Option 2 was preferred that needs no modification on the repositories' side just the creation of separate (repo schema <=> harvesting schema) mapping files per connected repositories. Note that the actual metadata field location depends on the schema used as well as on best practices (e.g., in LOM there is no trivial place for thumbnails).

It was also proposed to filer out those objects that has no correct metadata available (for instance, drop all object without a thumbnail URL available). Such a filtering can be done by the Ariadne harvesting engine.

Giannis also proposed two potential solutions for the specific thumbnail issue:

1. Ask the repositories to create them and make them available (i.e. let us know where they put them).
2. We can create the thumbnails at the Ariadne engine side and store them in a separate file system attached to the content URL.

Option 2 requires some development, of course. BTW, the issue of thumbnails is just one specific problem, there are other, more important issues like the availability of direct URLs to multimedia content, for instance.

A gradual approach for connected repositories can be applied as follows:

• a) Select the "friendly" repositories giving us the right metadata in a known location and format; Campusdomar, MAOR and SWITCHcollection are the candidates.
• b) Apply filtering on the Ariadne engine side to filter out objects without the metadata needed.
• c) Extend this method to other "non-friendly" repositories or maybe create thumbnails - where missing - by the engine.

Once we have agreed on the final list of friendly repositories and tested them out, we might want to re-do the metadata analysis exercise.

List of "friendly" repositories:

1) Campusdomar

- Issue: Where to put the thumbnails URL in LOM.

2) MAOR

- Issue: OAI-PMH target is needed for the Ariadne engine to harvest.

- Issue: Static XML snapshot is needed for the portal to display.

- Issue: MAOR will not provide thumbnails (needs to be created by the engine, maybe)

3) SWITCHcollection

- Issue: Content is available only after federated AAI. Ask them to make some content openly available.

4) SURFmedia

- Issue: SURFnet is not maintaining the collection any more. Maybe TERENA OER can take that over...

ACTIONS

* Get Campusdomar and MAOR working with Ariadne harvester and the Proof-of-Concept portal.

* Organise a meeting with SWITCHcollection people on making content open.

Next meeting will be on Tuesday, 23 July, 2013 at 2pm CEST.

Comments:

• perform an analysis of the metadata for MAOR and Switch
• create provider specific transformations from the provider's metadata schema to TERENA OER IEEE LOM schema
• add a filtering process at harvesting engine to filter out the metadata records that will not provide url to the actual digital resource
  • Kostas commented that the actual filtering has already be implemented in the Ariadne engine.

Next steps proposed by Vicente (UVigo):

• 1) Contact SwitchCollections and/or SurfMEDIA in order to get feeds meeting this minimum requirements 
• 2) Adapt the Ariadne engine to collect from this new content providers 
• 3) Ingest MAOR snapshot in the portal database or in the Ariadne engine
• 4) Analyze the metadata available in CampusdoMar, MAOR and SwitchCollections
• 5) Make a decision about facets that can be easily developed with the metadata available
• 6) Add the prototype the ability to extract thumbnails from the media if not provided
• 7) Improve the prototype and release the 1.1 version of this Proof-of-concept