Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

The OER small project is to bridge the gap between the end of TF-Media (finished on Dec 31, 2013) and the beginning of GN4 (April 2015). The reason why the idea must be tested in a small project before it’s introduced in GÉANT is the fact that the critical mass (in terms of participants, support, interest, etc.) has to be gained before any sustainable service development can be done. OER seems to be a typical “chicken-n-egg” problem at the moment (i.e. without a working prototype it’s hard to gain significant interest and without significant interest it’s hard to convince the development) therefore, the OER small project has to take this initiative. The OER is not the service that the NREN community is desperate to build (e.g., like the Trusted Cloud Drive pilot was in 2012) but it’s something that the lead has to be taken on (e.g., like the NRENum.net service pilot was in 2008) in order to facilitate the development of future value-added services on top (including MOOCs and others).

Image Modified

Anchor
del
del
Delivering the pilot

...

8th meeting - 9 October 2014 @ 9.00-17.00 CEST (Face to face meeting in Amsterdam, Netherlands)

9th meeting - 4 December 2014 @ 14.00 CET

Next coming:

...

10th meeting - 19 January 2014 @11.00 CET

11th meeting - 28 January 2014 @10.00 CET (Pilot has been CLOSED)

 

Anchor
11th
11th

Note
title11th meeting

 

Anchor
10th
10th

Note
title10th meeting

Recording:

Notes:

As we commented during the last meeting (19th Jan) I consider the development of Geant-OER Portal done. We are currently doing tests and QA on it.
Any help testing the portal and trying to find bugs is very welcome. You will find it here:
Please, try it and send any bug reports to this list.
Open Issues:

  1. No duration and direct URL to media on OAI-Repos:
    • OAI harvested Repos are not providing duration and direct URL to the media due to DC schema limitations or other issues( iTunesU harvested Repos are working great)
  2. No Licensing info:
    •  Only copyright owner info is coming from the agregator, no licensing info for any object
  3. Deleting strategy:
    • A deleting strategy should be defined. When should be an object deleted from the portal´s DDBB ? The first day it's not included in the agregator feed ? (all paradata and link to the object will be lost). We recommend a "soft-delete" strategy (not included in the current development).
  4. VIDEOTORIUM: No Subject info
    • No subject information available from VIDEOTORIUM, no iTunesU or UNESCO Codes available.
    • A Requirements document is needed for new repos willing to join (specially for non iTunesU ones) 
  5. Static Info pages needed
    • Info, FAQ, About and Contact pages are just mockups. Real info is needed
Comments from Giannis:
  • Home page should stress the fact that this service provide access to OERs from diverse sources. We need a marketing moto at the top of the home page "Discover multimedia open educational resources from XXXX organisations" or "XXXXXX Open Educational Resources from XXXX organisations at your hands"
  • the portal has two many menus. I would propose to have a more clean design with a single main menu 
  • the main menu contains links that are more suitable for footer. The main menu should have only links related to content discovery i.e. browse, search and advanced search and the About
  • we could add some nice logos of the data providers and some basic info http://terenatv.media.uvigo.es/en/library/abc.html
  • it is confusing that some of the resources open in the GEANT OER and other directly on the original site
  • mainly Spanish content promoted in the home page. This is not very good for a federation service. 
  • the menu on the left side menu is confusing. I would propose to have a top menu with main options like Browse, Search, About and transfer the top menu contents in footer
  • Recorded date field seems not to be correct. Is this really recorded date or ingested into the portal date?
  • Add friendly names for the Language filter and not ISO codes e.g. en, es
  • Could we have facets with statistics for each value e.g. English (10)?
  • using term agriculture in http://terenatv.media.uvigo.es/en/buscador/index?institution=all&lang=all&subj=all&year=all&duration=all&material=all&search=agriculture I get 8 results while using the search box powered by google I get 2. This is confusing. I would strongly recommend to use the Pumykit search mechanism in both cases
  • footer content seems not related to the scope of the project
  • add how to contribute link with guidelines for the providers on how they can provide their content and link this section to the wiki page that Kostas has shared
  • while we are showing audio as an option in the format filter there are no results and the term "no videos" is used. Maybe we could use the term "no resources" instead

Anchor
8th
8th

Note
title9th meeting

Participating:

Peter, Kostas,Vicente, Eduardo, Giannis, Eli

 

Agenda:

1. Summary/Update - Current Status

2. Focus on the Content, Functionality

3. Update on the back-end engine

4. Front and back end interaction

5. Timeline

 

- - -

1. Current Status

Vicente: working mainly on data structure, adapting the PUMUKIT structure to the application profile

working on the import scripts - not as straightforward as expected; modifying the scripts all the time - related to the structure and the vocabulary

[demonstrates JSON file example: issues with specific fields, supposed values, file formats do not match - need a translation at an aggregation level, media format doesn’t appear, strange language codes - or appear to be English, but isn’t - language codes should be translated, lack of numeric codes, etc.] - issued needed to be solved in order to get good data

Peter: problems related to the metadata translation - should be translated into one format or schema

Vicente: some issued should be fixed at the repository level

Peter: proper translation needed, metadata not exposed by local repositories - a) ask them to change that, b) try to translate it automatically

will have to tackle that later on

Vicente: our own repositories - improving the data exposed in order to make it easy to harvest, but data coming from iTunesu - cannot count on cooperation with the data provider; translation should be done at the repository level

we have all the tools, just need to focus on improving the harvesting process

Peter: can compare the local repository and the one harvested - enough for us to prove that it will work

Kostas: the phase of transformation is problematic, we don’t know what are the elements used from others’ perspective, might have problems with mapping the local elements with the elements of TERENA OER

Peter: thumbnail gets lost for example; need to look for the basic fields we agreed on

Kostas: last time we were harvesting there were no thumbnails

Vicente: part of the collection, not of the item, but we are not working with collections now

Kostas: could harvest information where thumbnail is a part of the item

Vicente: need to interact more to solve minor issues

Kostas: send me the list of issues

 

Vicente: regarding the portal - some objects have been harvested, currently the URL exposed is not the direct URL to the media, but the player page (can be changed)

categories, filters

search - first implementation

information available at provider level, possible to harvest it

 

2. Focus on Functionality

Peter: functions need to work, else the end-users will leave, if something doesn’t work, should not be on the pilot version; ok to put new features in as they become functional

Vicente: this is work in progress, users wont be exposed to anything that won’t be working, functionalities there for demonstration purposes only, will be removed

 

- - -

 

Peter: Kostas, once the translation is done, others will work or are there differences?

Kostas: main issue is translation, has to be done together with Vicente and his team

Peter: scalability of this process in the future - tedious manual approach for now for every repository; next repository will have to be treated the same? doesn’t seem scalable

Kostas: difficult to create a transformation file for every provider, will take time

Giannis: the first version is for the internal consumption, but later it will be important to communicate to the data providers willing to join

Vicente: maybe institutions willing to do what is needed, but might not; or we can learn to harvest the apple feeds properly

Peter: would not fixate too much on the apple one, we want to connect different sources

Giannis: need to have some educational metadata, else we will just replicate the iTunes harvesting mechanism; need to understand the semantics of the metadata - main problem, else can be corrected

 

ACTION: Vicente to send a detailed list of issues to Kostas

 

Eli: need to see how many hours Kostas will need for every repository 5-10 hours - worth it, else need to consider

 

Giannis: either remove the subject from the front end, or not use metadata that has no subject element - will apply to many elements

 

Eli: if we will see that we need more go arounds to fix something, maybe we can discuss the technology aspects in the next phase

Peter: we will have development time in the next phase (GN4, April next year) - Kostas will be spending a lot of time working on it

 

- - -

 

Timeline

 

Peter: one more meeting before holidays or better in January? Can also follow up individually  - (will be back on 5 January)

 

Giannis: specific collections from the iTunes site - who is responsible for those collections?

Peter: iTunesU was identified as an escape for now, but need to focus on the local repositories - everyone in their own countries should look for it, happy to do that in the Netherlands (also need to address the language question later)

Giannis: i will look for Greek collections

Peter: next meeting in January

 

Software fixes by Kostas

  1. Improved the mappings from native RSS Itunes U metadata and DC to the LOM based OER Application Profile.
  2. Extended the RSS Harvester to gather information from RSS Channel element and inject it to each and every element where the specific info is missing for example the itunes:image.
  3. Extended the seach API to offer duration, format and contributors name.
  4. The elements' possible detected languages are limited to en and es.
  5. The dates issue is not addressed(and it will be very difficult to be addressed since there is a great diversity in the way dates are defined)
  6. A new repository(Standford) is added.Peter sent it to me before sometime ago.It is RSS Itunes U based.Check the dashboards for more info.

...