view printer-friendly version (opens in new window)
Using SHAME to fill your SCAM
The deadly serious people at the Knowledge Management Research group at Sweden's Royal Institute of Technology just released their open source Standardized Hyper Adaptable Metadata Editor (SHAME). The editor is designed to be used in conjunction wih the equally open source, RDF-based Standardised Content Archive Management (SCAM) system. The pair are meant to kick start eLearning on the semantic web.
Given that SCAM is now in its second iteration, and is built in co-operation with the Swedish Agency for School Improvement and Uppsala Learning Lab as part of a nation-wide effort to create a Public e-Learning Platform (PeLP), neither software package appear to be fly-by-night operations. Should anyone, for some reason, get that impression.
As metadata editors go, three things are remarkable about SHAME. First, it is resolutely standard compliant; the demonstrator can edit both Dublin Core (DC) and IEEE Learning Object Metadata (LOM) RDF records. Given that the draft IEEE LOM RDF binding was largely developed (and implemented) by the same people, the tool could be seen as something of a reference implementation.
The DC and LOM capabilities of the demonstrator are not the end of its capabilities, though. The second major characteristic of SHAME is its "Hyper Adaptability", which means that it can be used to edit any resource-centric RDF based metadata. It therefore has the ability to be configured for whichever datamodel or set of vocabularies suit a particular purpose. It can even do this on the fly: just switch between the DC, LOM or any other form view on the same record and the SHAME demonstrator will simply show the same essential metadata, constrained and presented according to the dictates of the chosen datamodel. The demonstrator even comes with an editor function to edit the metadata forms it presents to the user who'll do the actual metadata creation.
Which points to the third major feature: rather than a monolithic application, SHAME is actually a framework for the building of editors, presentations and query interfaces to RDF metadata. That means that the core is essentially a library that can be integrated into any other Java eLearning application; web based, stand-alone or as part of something much bigger. The demonstrator is just a means of exposing the functionality of the system, and not necessarily the main aspect of the software.
Having said that, SHAME, like most other metadata editors, aims to make the chore of metadata creation as easy as possible by hiding much of the complexity from the user. How much of that complexity is hidden depends on how the forms are defined.
The main impetus behind the development of SHAME was the need for a flexible, easy to use editor to go with the recently revamped SCAM archiving tool. SCAM is based on the same standards, and, intriguingly, an RDF version of the ubiquitous IMS Content Packaging format. Technically, it is a reasonably portable and flexible RDF layer between a conventional relational database (MySQL, PostgreSQL and Hypersonic, at the moment), and a learning object searching or playing tool such as a VLE. The implementation is based on the widely used Enterprise Java Bean software component system.
Functionally, it should be able to archive almost anything you like, including portfolios and the like, but its main function is to expose collections of learning objects to searching and harvesting.
To properly understand both tools, it should be borne in mind that they spring from a strongly held vision of what metadata should be. In the Knowledge Management Research group's view, metadata just cannot be static, singular and objective. That is, a description of a learning resource by one person using a set of true/false type categories, to be valid forever is unlikely to be useful for everyone. Your idea of the Educational Level of a learning resource may be quite different from the person who created it, for instance. Likewise, you may think that the Topic a resource addresses is just one of many more. Or you may like to know things that can't be recorded by the resource creator: the size of groups the resource has actually been used in by other people, for example.
Hence the fundamental choice for RDF by the KMR group. Using this semantic web technology, there is no pressumption that there needs to be just the one metadata record. Anyone can add their own metadata, using whatever datamodel and vocabulary suits them. The ultimate result of such a vision would be constantly shifting layers of metadata around learning objects, with each layer representing a particular perspective. It is this that both SCAM and SHAME were designed to support.
Both SCAM and SHAME have their own pages on Sourceforge, from where the source, the binaries and (javadoc) documentation can be downloaded. You can find out more about other software made by the KMR on their tools page.