This session is really why I’ve come to the CETIS Conference (apart from the general opportunity to meet and chat to people which is also great) – it’s about “The Learning Registry” (@learningreg and http://learningregistry.org). The Learning Registry is not a destination – it’s about building infrastructure – and in some ways has both parallels and relevance to the work the Discovery programme is undertaking (which I’m involved in).
A simple use case for the learning registry is:
- Nasa publishes a physics video
- PBS posts a link to the video
- NSDL posts a link to the video
- A school uses the video in a course in their Moodle VLE
Each place/portals where the link is used or published only knows about use of their link or copy of the resource. So Learning Registry aims to support way of sharing this type of activity ‘in the open’ – so that this can be captured and reflected – the ‘social metadata timeline’ – Learning Registry is to provide infrastructure to support this. Learning Registry describes this type of activity/usage data associated with a resource ‘paradata’ – although learning registry learning registry doesn’t care what type of data it stores (as long as it can be expressed as JSON)
The learning registry is “an idea, a research project, an open source community project, a public social metadata distribution network’…
The guiding principles: be enabling, capability not solutions, no barrier entry, no single point of failure – everything distributed…
Not going to try to blog the technical architecture of.. but summary of APIs
- Distribute API – uses http POST. About copying data from one node to another – i.e. achieving the distributed part of the architecture
- Publish API – how you get stuff into a Learning Registry node (that is, you, the producer of information, publish it *to* the learning registry node) – uses http POST. Learning reg also supports SWORD for publishing data into a node
- Obtain API – getting data out of a learning registry node – uses http GET
- Harvest and OAI-PMH APIs – another way of getting stuff out of the node. Harvest returns JSON but supports OAI-PMH type actions. OAI-PMH also supported.
It is stressed that this is really a project at the start of its work – the way to engage and to find out how to do this stuff is to join the community – join the developer list etc. and raise issues, ask questions – this is part of the experiment and will inform the development.
JLern is the project to setup an experimental node in the UK – being run by Mimas.
2 kinds of nodes in the Learning Registry:
- Common node
- Gateway node
JLern have setup a ‘common node’ – this can support:
- Publish services
- Access services
- Distribution services (JLern now have a 2nd common node up and running to try these)
Common nodes can be part of ‘networks’. Networks can (only) be connected via ‘gateway nodes’
When networks are connected, this is called a ‘community’. A ‘network community’ is a collection of interconnected resource distribution networks. A resource network can only be a member of one community.
Now have published the JORUM metadata (via OAI-PMH) – so about 15k resources. Open University now looking at similar activity. Now Jorum exploring framework for capturing paradata about resources.
Gathering ideas and use cases now – e.g. see JLern challenge from dev8D http://dev8d.org/challenges/
The JLern ‘Alpha’ node is at alpha.mimas.ac.uk – you can authenticate using details given in this blog post http://jlernexperiment.wordpress.com/2012/02/02/alpha-node/.
As already mentioned they’ve harvested JORUM OAI-PMH data and published on JLern alpha node.
They now have ‘Beta’ node (this doesn’t represent a level of development – just naming convention I think) – this is running on Windows (Alpha is on Linux). Also planning a ‘Gamma’ node running on Amazon EC2.
JLern hackday held in January – write up at http://jlernexperiment.wordpress.com/2012/02/21/the-hackday-report-and-reflections/, and also a Java Library for interacting with Learning Registry nodes at https://github.com/navnorth/LRJavaLib .