The semantic web fused with web 2.0 technologies.
Anthony Michael McDonald (Friday, 4th February, 2011 (12:16 p.m.))
I've been thinking about how to use DLMs more widely in systems we deploy here at Newcastle. There are learning and teaching systems, administration systems and others, so there's a pretty broad base of requirements to consider and get ideas from.
One of our most important systems is our Medically-orientated VLE, called the Learning Support Environment (all the teaching and learning content is behind a firewall), which has the following components;
Making sure all this material is up-to-date, findable, current, accurate and in the right context is somewhat challenging. For example, the study guides are loaded up from highly structured and well-styled Word documents, which are saved as RTF, then translated into docBook and finally decomposed into something we can then add dynamic data such as timetable information and teaching resources, which are presented (in context) in the online environment.
In effect, we are using the structure of the Word document, which is a linear re-presentation of our curriculum, as our substrate upon which to build dynamic content. It works, but it's not very flexible.
Suppose we were to turn around and use DLMs for that substrate, how would it work out, and how would it look?
Well, we would need to deconstruct the study guide into something which is 'node'-sized for the DLMs machinery. This could be at the paragraph level or smaller. That isn't so bad to do, we have a lot of contextual information on the guide itself (where it sites in the curriculum, who is the module leader etc) which would contribute to over-arching metadata on the document. We would then need to add DLM-specific metadata on each node. The metadata is quite varied, from simple one word descriptions (eg simple tags) through to multiple-selections for licence usage of the material itself (we very much believe in OER!). The metadata also helps us to decide how the content should be rendered - eg as simple HTML, as something which is only released in a specific time frame, something that is only seen by particular categories of user, etc This deconstruction is certainly doable, and the DLMs team has already done this for small sections of study guide material.
How the material is presented is vitally important, as simply showing a list of nodes that in some way match a study guide is no better than doing a tagged search. The material needs to be presented in order, and in context. Fortunately the DLMs nodesets have an 'ordering' attribute, which allows us to collect together a series of nodes, arrange them in the order the module leader wants and then re-present them to the end user. This re-presentation can be in standard HTML, or as something that goes through a processor to provide PDF, RTF or something else.
Now where it gets interesting is that the nodesets themselves are simply a collection of pointers to where the real information lies, information which could be a blob of raw text, a formatted HTML chunk, an image, a link to a series of external resources, timetable information, a discussion thread or indeed anything you like. You could imagine a nodeset which represents where in the programme the topic 'kidney' is mentioned, and perhaps slim that down by saying you're only interested in assessment episodes (I think students might be interested in that), and then expand it out by saying you're ok with links 1 jump away from the word 'kidney', which might bring in a lot more nodes that have the word 'renal' and many other concepts associated with them.
You end up with a way of re-presenting all the information in the curriculum, and all the information that the curriculum links to, in ways in which the original creator of the content will not have thought of, and because you can export a nodeset in various formats, including FreeMind format, you make this 'web of nodes' to the content more widely available.
I'll come back to this theme again, but thought it would be good to get it out there and let people know what some of the possibilities of DLMs are.
No comments have been added, why not be the first.
Please feel free to comment on "DLMs as a 'substrate' for academic content".