The title of this post is a question, and I’m actually interested to hear a few different things from people out there:
- not only what you think the difference between ‘digital repository’ software (e.g. DSpace) and ‘learning object repository’ software (e.g. CAREO) actually is, but also
- are any differences you think exist there because these types of systems really are trying to accomplish different functions, or are dealing with heterogeneous resource types, or are they different simply because they originated in different camps within educational institutions (libraries and archivists versus educators or educational technologists)? Other reasons?
- should institutions adopt a single institutional repository system for both archiving research publications and their learning objects? Do you work at, or do you know of, an institution that is adopting a single repository software solution for all of there needs?
more…
There are a few reasons I’m asking these questions. One is that I’m currently working on a project to compare various software (and hosted services) that educational institutions can use to store, manage and deliver ‘learning objects’ (leaving aside for a second what those actually are) – systems like CAREO, Intrallect intraLibrary, Connexions and the Basic E-Learning Tool Set to name a few on what is not that huge list of things currently known as ‘learning object repositories.’
Someone suggested we include DSpace on the list, and there have been other mentions of Fedora as another possibility. I was at first pretty sceptical – while on the surface both types of systems ostensibly provide a searchable archive for digital materials from an institution, not only are the materials typically quite different (e.g. scholarly research papers versus teaching-focused content and applications; PDF files versus zipped SCORM objects, Java applets, etc.), but ‘digital repositories,’ with their heritage in the library and archives world, typically support different metadata schemas, and are focused on interoperating with different kinds of systems (library catalogues instead of, say, course management systems), and, I think, embody a different business logic. Sure, people work on metadata crosswalks, and things like the IMS DRI spec seem like maybe they bring the two types of systems closer together, but still, they seem to be different tool for different needs for at least a few good reasons.
And yet… go to DSpace’s very own FAQ page, and there you will find listed ‘Learning Objects” as one of the digital content types DSpace can support. Is this just overreaching on the part of DSpace, or are the business logic and needs that are often addressed by two different applications similar enough that they could be served by an individual application?
My sense is that, no, these are two different (if related) types of applications for good reason, and that the folks who work on crosswalks and interoperability have the right idea – there needs to be different schema to reflect different domain concerns, and different applications to embody the different business logics needed to address distinct problems, but that working towards standards is a good thing in enabling cross-application/cross-domain functionality for end users who may not particularly share your domain’s particular context, constraints or concerns. But I’m interested to hear what other’s think – is there value in these two functions converging in one application (albeit one that would still likely need to support different schemas, and different skins for the different user bases); are you pursuing a strategy that seeks to bring these two repository approaches together?
My take is this: CAREO is intended to be used to support teaching and learning (learning objects, teaching resources, etc…) while DSpace is more of a document storage tool.
The next version of CAREO (which has been rumoured for some time now, but is finally starting to see the light of day) will do stuff like allow you to create presentations from sets of learning resources, and other cool things.
I have been working with ADLIB (http://adlib.athabascau.ca/) for some time and I think that we are still in the early days of what could be a very interesting shift in the storage and access of learning objects whether they are documents or “learning objects” whatever those may be when all is said and done. We have come a long way in establishing the infrastructure and pushing standards (IEEE LOM, CanCore, DublinCore, etc.) but I think we still have a long way to go in the areas of digital rights and working out incentives for the creation and metatagging of quality learning objects.
Another concern I have is the possible proliferation of JLO’s (Junk Learning Objects). While we now have some very explicit metadata schemes now to aid discoverability, they are still open to abuse.
Scott, in my experience discussions of repository technologies often take place with no distinction drawn between learning object repositories, and learning object metadata repositories; or, if you will, ‘LOB repositories’ and ‘LOM repositories’. IntraLibrary is an example of the former. From what I know of it, it will store IMS-LDs and full-blown EML, in addition to digital files, and various flavours of IEEE-LOM/IMS metadata. It certainly seems much closer to a ‘digital repository’ than the current version of CAREO, which is what I’d term a LOM repository.
A theoretical LOB repository, built on a native-XML or object-oriented database, and housing both metadata and learning objects, provides synergies in terms of authoring, configuration, programming, and search/retrieval that a LOM repository, with pointers to external/distributed objects, cannot.
On the other hand, LOM repositories such as CAREO provide a much faster way for institutions to catalog, search, and retrieve digital learning content which already exists, and there’s no requirement for a change in the institution’s authoring practices. I’ve heard that CAREO installs on an OS-X box in an hour or two, so it seems like a great starting point for any institution that can’t commit a lot of resources to implementing a unified solution like IntraLibrary.
Regarding your last question, I would say “It depends on what you mean by ‘single solution’.” Implementations of Oracle or Bluestream XStream, (two radically different products),configured for IMS-LD authoring via Web-DAV in one interface, and storage of file objects with metadata, in another, might be considered single solutions, and there are probably turn-key products lurking just around the corner that allow for both.
So, I think a more important distinction lies in the difference between products which centralize metadata and learning objects, vs. products which only centralize metadata….
In Doug’s terms, intraLibrary can be used either as a Learning Object Repository or an Learning Object Metadata Repository. (It’s also as fast to install as CAREO apparently is 😉 ) Unfortunately intraLibrary doesn’t yet have built-in support for EML or IMS Learning Design yet (See the LAMS tool at Macquarie for that) – perhaps confusion has arisen because Intrallect has been doing other work in that area.
As far as the main issue goes. I think the key features that (currently) define a learning object repository system include:
import/export/searching/harvesting of IEEE LOM/IMS/SCORM metadata, and related application profiles
import/export/preview of IMS and/or SCORM Content Packages
Digital Rights Management, remote querying, agregation & disagregation and globaly identifiers are all going to be key issues in the future.
The JISC-funded JORUM+ project in the UK has completed an excellent study scoping a repository for UK Higher Education, which can be downloaded from their website.