Retrieved from the TEL opinion blog, October the 15th, 2006 

GRID technology is a more and more important field of investigation in computer-science. It could be seen as the digital implementation of the old proverb “l’union fait la force” (unity makes strength), by its way of using distributed computing resources to make a kind of super computing device. Indeed, there are many problems to solve to ensure the reliable and efficient cooperation between machines having different pieces of a computing task, their solution requires the convergence of theoretical frameworks from parallel, distributed and high performance computing. Taking up the challenge of GRID technology is strategic. A proactive policy does now support the progress in this domain.

 As is often the case, the emergence of a new advance on the technology front stimulates a new exploration on the front of educational technology. This is the old technology push story, which in my opinion is not wrong as long as it is not driven by a digital mirage. The recently edited book “Towards the learning GRID ” is a good occasion to learn where we are with this domain and what is the vision of its potential for TEL. The authors, among which several key participants of the Kaleidoscope Learning GRID SIG , set the scene in a very relevant way: “previous projects that have set out to improve learning through novel technologies have often failed to leave any significant mark because they did not give priority to the social, economic and technical perspectives of the key human factors” (p.63). So the learning GRID venture for which the ultimate goal is to ensure “the possibility to personalise the learning processes according to the learners’ preference and style” (p.74), will not be replicating errors we made in the past.

 Following the leaders of the ELeGI European project our colleagues from CRMPA and DIIMA , the added value of the GRID approach is to give a framework to realise the effective sharing of heterogeneous resources based on the concepts of service, distributed collaboration, and virtual organisation (p.66). A characteristic of GRID technology, they remark, is that it is not based on a mere protocol communication among the resources, but on message communication among services; This remark suggests that the resources could engage conversational interactions about what is requested and what is responded. This means a kind of intelligence that, if it were to exist, must be based on an understanding of the learning/cognitive and pedagogical/didactical requirements. But at this point we may be disappointed by reading the authors suggestion that “an innovative aspect is that our general model is the presence of three models: Knowledge Model, Student Model and Didactic Model, which substantially interact among them to define the specific and personalised formative path” (p.73). Of course, I think that they are not coming back to the seminal ITS trilogy. So, where does that leave us?

In a fully distributed world (wide web) these models should be distributed as possible features of the resources. Moreover, they will be explicitly or implicitly tangled (e.g. a knowledge model for a learning environment de facto include an epistemic and a didactical model—the latter being at least that of a didactical transposition). Then, I would suggest that:

The didactical nature of a GRID for human learning is an emergent property of the resources it gives access to and the interaction it supports, and not the property of one of the components

This resonates well with the “learning ambient intelligent vision of the learning domain” on which these authors conclude their contribution (p.74). Then we may choose to consider teaching as an emergent property of a system instead of one of its dedicated functions. This view may invite us not to consider GRID technology as an end, but as “[the] technology for building the next generation of learning environments” (p.182):

“Through the adoption of [GRID technologies], we can have a wide-scale learning resource sharing in heterogeneous and geographically distributed environments, the implementation of learning organizations in which different actors (universities, teachers, learners), sharing a common target, are able to cooperate to achieve a result” (p.183)

Let’s add that the actors could be artificial as well as human, and we get the picture of a hybrid world driven by the learners’ needs. Eventually, this is a picture that we could have drawn with the modern multi-agents technology. So what is the specific added value of the GRID approach? It is claimed that it is a solution to fulfil educational specifications that we all share nowadays, but still this solution is presented in rather general terms. We are almost sent back to the initial question but from a slightly different perspective: what are the problems relevant for GRID research when raised from a learning perspective? What are they, not in general educational terms, but in computer-science terms?

These comments are based on the following chapters taken from Ritrovato P., Allison C., Cerri S.A., Dimitrakos T., Gaeta M.  and Salerno S. (eds. (2005) Towards the Learning Grid Advances in Human Learning Services (Frontiers in artificial intelligence and application series, volume 127). Amsterdam: IOS Press.

Gaeta M., Ritrovato P. , Salerno S. (2005) Making e-Learning a service oriented utility: The European learning GRID infrastructure Project. (pp.63-78) Capuano N. , Gaeta A., Laria G., Orciulli F., Ritrovato P. (2005) How to use GRID technology for building the next generation learning environments. (pp.182-191) Lemoisson P., Cerri S. , Sallantin J. (2005) Conversational interactions among rational agents. (pp.214-230)