OER and the aggregation question
Over the years, web thinking about bringing content together has shifted its focus between portals, repositories and registries. Along with others, recently I’ve been using the term “aggregations” to try to avoid the definition potholes in that road. The Resource Discovery Task Force is exploring the role and mechanisms for effective aggregations to support research. The Learning Registry Project in the US is looking at the best “web-scale” ways to surface content useful to learning to the user. The UK Academy/JISC OER Programme’s Thematic Collections projects are building different types of aggregations. I am working with the Jorum service on refining its purpose in an environment where web teaching resources to the web is easy. There are a range of aggregator services in this space, such as Xpert. This post is not about any particular such service but about modelling them as a type of service.
I’m interested in the pragmatic approaches to digital infrastructure: who needs to does what, who pays the people developing the infrastructure, and what incentives are there for people to join in. I think these can usefully be asked about whether/how to aggregate “OERs”.
OER and aggregation
The landscape for considering the aggregation of OER looks something like this to me:
- The concept of OER is fuzzy, the line between “open” and “not open” is fuzzy, and what defines content as educational is also fuzzy-edged
- In modelling the impact/benefits of OER we need to include individual, institutional and social good: approaches to digital infrastructure investment need to reflect this
- Effective publishing and appropriate use of digital resources are part of digital professionalism, and additionally, citation and attribution are part of digital scholarship: academics need to aspire on both fronts, whether they are mainly teachers or researchers
- Rights, attribution and citation are of growing importance, technology is moving fast, licensing models will continue to develop, digital infrastructure need to have some flexibility
- There are some interesting design patterns and infrastructure learning curves to be described in this space: driving consistency of metadata vs lowest common denominator, driving standards vs using native web standards, extent of integration with social media etc
- Some say “the web is the aggregator”, and are wary of creating glossy showcases, however, showcases can have a role in increasing uptake as well as giving senior stakeholders somewhere to point to to show the results of investment
- There are a range of reasons to aggregate, and they need making explicit so as to situate the effort in the right place in the value chain, and to understand the use case so that we can know if we’re doing the right thing
Leaving aside the question of whether to aggregate the item/file/object itself, I’ve been thinking about why we want to create aggregations of metadata anyway. With help from Andy McGregor, here’s a list of reasons why we might want to create aggregations. This is not systematic or carefully modelled. If someone has done this more thoroughly already for the OER space I’d be very grateful for references. This is just a starting point.
- Curate: Select sources for your audience: take the role of curator
- Promote: Push a range of content to an audience
- Present: Enable views across a set of resources, such as: present by geography (map), present by latest, present by most used
- Visualise: Specifcally present content visually such as graphs, maps and charts
- Showcase: Bring together content to show the value of the collection
- Speed: cache for faster searching
- Scale for SEO: create points of scale for effective search engine optimisation
- Normalise: vehicle/driver for normalising metadata, driving consistency
- Broker: It’s a many:many service, a broker between multiple content sources and multiple potential users.
I suspect there are reasons around access control too, but as I’m thinking about OERs here I won’t explore that, suffice to say even with open web content it is possible that the users environment might be locked down (schools, NHS) : a quality assured aggregation might be allowed through the security layers.
Implications for OER aggregator services
- Content in: Some of these use cases can be met by automated aggregation on the fly, others use handpicked sources but automated presentation, others need heavy editorialising. Some will need a relationship with content owners providers, they will need registration and quality assurance. Others will not.
- Content out: Some aggregations might primarily serve other services: widgets or mobile apps or search services. Other will be modelled primarily for human interfaces.
- This all has implications for how aggregation services should be measured: what are their performance metrics? Volume of content in? Volume of content out? Levels of human use? Levels of machine use?
- And how might people’s use of the aggregation change over time? Does your careful search engine optimisation start to develop user recognition of you as a trusted aggregator, leading them to be likely to follow search results surfaced by you? Or might your careful curation of the best of the web gradually be replicated in the users’ personal bookmarks to the extent they no longer come back? Is that ok? Do people use you because they are new to the topic you cover, and use you to orient themselves, or do they use you to keep them updated in their specialist are? Do your users stick with you or do they only use you as part of their learning curve? Thinking of users as cohorts situated in time might help you grow with them?
- A lot of this is known by services but perhaps not always clearly articulated. As the landscape, or ecosystem, becomes more cluttered with aggregations, big and small, perhaps we need to work towards greater clarity. Which services can collaborate, which are competing, which will consume each other?
Aggregation services need to pitch their value to the content providers and the content users: as an aggregator what value do you add? And who should pay for the work?
Feedback welcome via comments and/or OER Discuss Mailing List