Earlier in the week we at Talis hosted a Research Day the theme for which was around what we refer to as Project Xiphos. Through Project Xiphos, we are exploring the impact of applying the latest Web scale technologies, including Web 2.0 and the Semantic Web, to the challenges of education and research within Higher Education institutions.
The program for the day was quite varied, with speakers talking about a number of different issues related to higher education We opened with Peter Murray-Rust from the University of Cambridge, who spoke passionately about the need to open up research and research data. Following him was Andy Powell from Eduserv who talked about Web 2.0 and Repositories, I found his talk to be rather enlightening since I’m fairly new to this domain and still learning. Following Andy was Carsten Ullrich from Shanghai Jiao Tong University. Carsten had flown in from China to give a presentation on “Why Web 2.0 is Good for Learning and for Research: Principles and Prototypes“. I had already seen a variation of this presentation at WWW2008 last month, in fact those of us who were there felt that his ideas were hugely relevant to some of the issues we are trying to understand around Higher Education so we invited him to speak at our event. Carsten is based in the e-learning lab – and is researching into how learning can be made easier and more interactive using technology. Most recently his team have been looking at Web 2.0 technologies/approaches and how they can applied to learning. What they found is that these approaches were transformational, in other words that you you have to change the way you teach to use these approaches and benefit from them. The final speaker before lunch was Alan Massen from University of Ulster, who presented a project he has been involved in using a Hybrid Learning Model to Describe Learning Practices. The model he presented addresses a issue in higher education around the fact that teachers don’t posses vocabulary to describe their teaching practice/pedagogy, and actually this information is quite important to students as well since it can be used to help them understand their learning goals. After lunch my colleague Chris presented one of our research prototypes, Xiphos Network, which creates a social network from a Web Scholarly Data. Following him another of my colleagues, Ian Corns, presented Project Zephyr, a resource (reading) list application that we will soon be trialling with several institutes here in the UK. Following this, the final session of the day was me, giving a presentation on Open World Thinking. The remainder of this post will focus on what I said during that presentation, with some links at the end to my slides.
Open World Thinking
I started off by introducing myself and explained that whilst my colleagues had presented some examples of applications we had developed that addressed some of the issues raised during the day, my talk was going to be slightly more esoteric … and understanding that we might need to change the way we think about problems.
In order to set the tone for some of what I wanted to address later in the presentation and in attempt to get the audience to start to appreciate the complexity of some of the problems we want to be able to solve but currently can’t because of the way we think about things I posed a couple of questions for the audience:
1. What is the most referred to text in first year undergraduate computer science courses in the UK? and 2. Based on pedagogical approaches, what are the recommended resources required to teach an emerging subject in a new department a University somewhere?
The questions are largely rhetorical, or just plain impossible to answer. So they next thing I asked was whether the audience felt that the reason we couldn’t answer the question was because the data didn’t actually exist? I stated that I believed that the data probably did exist but if it did then it existed within individual institutions, across a myriad of internal systems, and sometimes even across departments. This led me to make the observation that Higher Education Institutes are not just silos they are in fact “silos within silos”. I also pointed out that as long as they remained silos we wouldn’t ever be able to answer the kinds of questions I had posed earlier. Which is the fundamental reason why we need to start finding ways of linking data together across these silos.
I diverted slightly to ask the question “Why have we ended up with silos?”, and offered my own answer that really these silos were a product of close world thinking, historically the systems implemented with institutions where designed to solve a set of problems for that institution, they were never designed with interoperability in mind, and were really about controlling data. However the problems we are trying to solve today are different to the problems of old, the world has changed and so we need to change too. One way of addressing these problems of interoperability, problems of sharing and reuse is to be more Open. Tim Berners-Lee said that “Openness tends to be an inexorable movement through time” and that’s something that I believe is true. I mentioned that I wanted to talk about two aspects of this what I describe as an Openness of Description and an Openness of Access.
However before going any further I wanted to make an important point, that what absolutely not talking about is a technology change. What I’m talking about is a paradigm shift, a very different way of thinking about the problems we are trying to solve. I then displayed a picture of the Linked Data Graph with another of Tim’s quotes that “Linked Data is the Semantic Web done right, and the Web done right”. I explained that this graph represented data sets published by communities in an agreed form in order to facilitate re-use and linking disparate data sets together. Through this level of openness others can come along and build new applications and services that use this data. I talked about how this facilitates the notion of “Designing for Appropriation”, where the creators of an artifact might intend it for a purpose but others can appropriate it for a completely different use. This is also the promise of the semantic web the fact that we no longer need to rely on structures to be defined up front, we can slice and dice this graph of data in order to create structures on the fly. However in order to achieve this we need to design data at the right level of granularity.
So openness of description is about agreeing on ways of describing certain kinds of things. When you have shared, open ways of describing things it makes it easier to Share, to relate things together and to integrate across. I also anecdotally pointed out that as Rob had pointed out to me “through openness of description and dereferencible URI’s you get interoperability for free”, which is true … kind of 😉
One way agreeing on shared descriptions is through the use of Ontologies, I explained what they were and cited a few examples. In order to illustrate the point further I used Alan Massen’s Hybrid Learning model as an example and suggested that since it was a “controlled vocabulary” that defined a set of “concepts” and the “relationships between those concepts” what he actually had was the basis of any ontology. I also suggested that if every institution in the UK/World used this ontology to describe their courses you could present an enormous amount of information to students in a standard way which might make the selection of courses or indeed the choice of which institution to go to more transparent in terms of the learner understanding what the institute provides and how, but also what is expected of them as students.
I pointed out that Talis has worked on developing a number of ontologies that are all being used to underpin the applications we are producing as part of Xiphos, but others outside of Talis have also started to adopt. I pointed the audience to www.vocab.org for more information.
I then went on to talk about Openness of Access ( which is not Open Access ), it’s the idea that we need to provide users with ability to consume content and information anywhere, anytime, anyhow. I also pointed out that unless we do this we can’t create the kind of truly personalised learning experiences we envisage. Part of this is recognisng what I’ve referred to a lot recently as the need to develop applications as Contextualised Perspectives onto this amorphous web of scholarly data.
This is a large part of the paradigm shift, recognizing that if this Web of Scholarly Data exists then we don’t own it, it exists independently of the applications and services that are built upon it. However it becomes the job of application vendors or developers to create value by developing Contextualised Perspectives onto that graph of data that allow their users to perform a set of activities, or achieve some goals that he or she sets out to. In other words this Web of Scholarly Data allows us to create contexts on the fly that are relevant for a task your doing at the time your doing it … in some ways thats the grand vision. A perspective could be a facebook application, or an iPhone app, or some functionality embedded in an institutions VLE, an enterprise application … anything … but the point is that we need to create these perspectives since it’s only through them that ordinary users can make sense of it all.
… *phew* … I think i’ve pretty much covered most of what I said … however for a slightly more coherent view of it all, I’ve written a piece on Open World Thinking in issue 2 of Nodalities.
You can access the slides here.